Complex Strategic Choices PDF

Decision Engineering
Series Editor
Rajkumar Roy
For further volumes:

http://www.springer.com/series/5112
Steen Leleur
Complex Strategic Choices
Applying Systemic Planning for Strategic

Decision Making
123
Steen Leleur
Department of Transport
Technical University of Denmark
Building 115, Bygningstorvet
Kgs. Lyngby 2800
Denmark
ISSN 1619-5736
ISBN 978-1-4471-2490-0
DOI 10.1007/978-1-4471-2491-7
e-ISBN 978-1-4471-2491-7
Springer London Heidelberg New York Dordrecht

British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Control Number: 2011944751
Springer-Verlag London 2012
Apart from any fair dealing for the purposes of research or private study, or criticism or review, as
permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced,
stored or transmitted, in any form or by any means, with the prior permission in writing of the
publishers, or in the case of reprographic reproduction in accordance with the terms of licenses issued
by the Copyright Licensing Agency. Enquiries concerning reproduction outside those terms should be
sent to the publishers.
The use of registered names, trademarks, etc., in this publication does not imply, even in the absence of
a specific statement, that such names are exempt from the relevant laws and regulations and therefore
free for general use.
The publisher makes no representation, express or implied, with regard to the accuracy of the
information contained in this book and cannot accept any legal responsibility or liability for any errors
or omissions that may be made.
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)
Preface
Complex Strategic Choices sets focus upon providing deliberate and methodical
support for decision-makers engaged in strategic decision making. The book aims
especially at shedding light on decision support under conditions that are uncertain
and complex. In this respect systemic planning (SP) will be introduced as a new
approachthe SP approachto deal with future-oriented, often long-term, actions
which can improve our capability of meeting present-day strategic challenges.
The book is aimed at people with an interest in new possibilities for supporting
decision making. This group of people includes managers and corporate employees
and also students and researchers of various kinds and for that matter all people
seeking to come to grips with the challenge of what we may term deliberate, informed
change. Known in some contexts as decision engineering (DE), the book will show
that DE is embedded into and dependent on a plethora of knowledge types which
range from applied mathematics known as operations research via economics and
organisation theory to social science issues. However, throughout the book focus will
be maintained on formulating a framework of practical advice. Stated quite simply,
this practice framework will aim at supporting analysts and decision-makers in
achieving better complex strategic decisions.
Normally a book is read in a linear way, which in this case means from Chap. 1
to Chap. 8. Such a reading of the book is based on the perception that it can be an
advantage that the theory is in place when the new methodology is presented as a
coherent process-and-methodology framework in Chaps. 7 and 8. However, it is
my experience from teaching planning and appraisal courses at the Technical
University of Denmark that theoretical issues appear to be more relevant and
interesting if the students early on are at least partially convinced about their
applicability and usefulness. Therefore it may be relevant for the reader of this
book to reverseor at least consider doing sothe order in which the chapters are
read or to split up the reading into appropriate parts. Therefore if you as reader is
curious to find out whether the book offers some interesting news, an option may
be to read Chap. 7 on the case example first and then Chap. 8 which summarises
the main findings and recommendations. Then you may continue with Chaps. 5
and 6 with their presentations of tools and methodology. Chapters 24 about the
v
vi
Preface
grounding of the later practice are probably the most heavy-going and may
therefore as suggested above be postponed and skipped in the first place. Chapter 1
is an introduction that, among other things, sets out the main themes treated in the
book and gives an overview of the chapters. To facilitate the reading of the book
and tie the individual chapters together, each chapter ends with a listing of the
main points and findings of that chapter. Additional technical material is presented
as two appendices.
The SP approach treated in the book has so far been tested on a number of cases
which has shown its potential to those involved. I sincerely hope that Complex
Strategic Choices will inspire and support readers in their dealing with upcoming
complex planning problems and strategic decision making.
Virum, Denmark, October 2011
Steen Leleur
Acknowledgments
In connection with the completion of the book, I would like to thank the following
for collaboration about systemic planning (SP) cases and the implementation of
methodology: Assistant professor Kim Bang Salling, research assistant,
Ph.D. student Anders Vestergaard Jensen and research assistant, Ph.D. student
Michael Bruhn Barfod. I also thank the many students that have taken my planning
and appraisal courses at DTU Transport in the last five years for their useful
feedback on the SP theory and methodology. My thanks also go to the many
participants in Denmark and Sweden that over the last five years have attended SP
decision conferences and contributed in a very constructive and positive way. This
has been of great value to the formulation and adjustment of the practice framework described later in the book. A successful decision conference is in the heart
of systemic decision support in accordance with the SP principles.
Thanks also to professor Michael C. Jackson, professor Frank Stowell and
professor Cathal Brugha who all provided creative suggestions to an early outline
of this book. Furthermore, I am grateful for constructive comments to a draft
version of the book provided by professor Jotin Khisty and professor Alan
Pearman and to both of them also for enjoyable research collaboration over many
years. My writing in the final stage gained from my attendance as a teacher at the
Helsinki Summer School in Transportation 2011. I owe thanks to professor Antti
Talvitie for inviting me and letting me present some of the ideas contained in this
book and for his valuable comments. I also want to thank DTU Transport director
Niels Buus Kristensen for offering me a month free of duties to make it possible
for me to concentrate on the manuscript. Furthermore, I appreciate research grants
from the Danish Strategic Research Council and the EU Regional Development
Fund, Interreg IV-A, for my work in the research projects UNITE (20092012)
and EcoMobility (20102012), which have benefited the development of the ideas
presented here.
Senior editor Anthony Doyle at Springer deserves thanks for having originally
invited me to write a book in Springers Decision Engineering Series. He and his
assistants Claire Protherough and Grace Quinn at Springer furthermore deserve
my thanks for their kind help and support in the publishing process. Also thanks
vii
viii
Acknowledgments
to office clerk at DTU Transport, translator Ulla Salado-Jimena for valuable

language assistance, and last but not least to my wife Susanne Leleur for constructive suggestions and comments on a preliminary version of the book. To my
wife also for understanding and patience when writing work has absorbed my
off-duty time.
Contents
Introduction: Complexity as a Challenge . . . . . . . . . .

1.1 Complex Strategic Choice and Systemic Planning . .
1.1.1 The Idea of Systemic Planning . . . . . . . . .
1.1.2 The Framework of Systemic Planning . . . . .
1.2 An Overview of the Contents . . . . . . . . . . . . . . . .
1.3 A First Look at Complexity and Foresight Problems
1.3.1 Detail Complexity . . . . . . . . . . . . . . . . . . .
1.3.2 Dynamic Complexity. . . . . . . . . . . . . . . . .
1.3.3 Preference Complexity. . . . . . . . . . . . . . . .
1.3.4 Categories of Problems . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
1
2
2
3
5
5
7
9
12
13
The Condition of Complexity . . . . . . . . . . . . . .

2.1 The Systems World of Luhmann. . . . . . . . .
2.1.1 Systems According to Luhmann.. . . .
2.1.2 Systems and Complexity . . . . . . . . .
2.2 System and Environment . . . . . . . . . . . . . .
2.2.1 The Midas Touch of Contingency . . .
2.2.2 Contingency and Decisions . . . . . . .
2.3 The Socio-Technical System and Its Futures.
References . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
15
15
16
17
18
19
20
21
25
Linking Complexity and Simplicity. . . . . . . . . . . . . . . . . . . . . .

3.1 Ways of Seeing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2 Two Epistemic Lenses Concerning Simplicity and Complexity.
3.2.1 The Complementarity of Simplicity and Complexity . .
3.2.2 Systemic Thinking Further Defined . . . . . . . . . . . . . .
3.3 Combining Simplicity and Complexity Thinking. . . . . . . . . .
3.3.1 Towards Interaction of the Whole and the Parts . . .
3.3.2 Reflection-in-Action . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
27
27
28
29
30
30
31
31
33
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
ix
Contents
The Systemic Process . . . . . . . . . . . . . . . . . . . . . . . . . .

4.1 From Novice to Expert. . . . . . . . . . . . . . . . . . . . . .
4.2 Towards Intelligence Beyond Calculative Rationality.
4.3 Arationality and Subworld . . . . . . . . . . . . . . . . . . .
4.4 Unfolding the Systemic Process . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
35
35
37
38
39
41
The Systemic Toolbox . . . . . . . . . . . . . . . . . . . . . . . . . . .

5.1 Systems Science as Three Waves . . . . . . . . . . . . . . . .
5.2 Formulating Five Modes of Enquiry . . . . . . . . . . . . . .
5.2.1 The Two Promises of Applying Systems Science
5.3 Selecting Methods and Techniques for the Toolbox. . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
43
43
48
49
50
53
Setting up the Decision Support . . . . . . . . . . .

6.1 Scoping the Strategic Choices . . . . . . . . . .
6.1.1 The Soft Methods in the Toolbox . .
6.1.2 Critical Systems Heuristics. . . . . . .
6.1.3 Soft Systems Methodology. . . . . . .
6.2 Assessing Consequences and Risks . . . . . .
6.2.1 The Hard Methods in the Toolbox .
6.2.2 Cost-Benefit Analysis . . . . . . . . . .
6.2.3 Multi-Criteria Analysis . . . . . . . . .
6.2.4 Uncertainty and Risk . . . . . . . . . . .
6.3 Creating Choice Intelligence . . . . . . . . . . .
6.3.1 SP as Multi-Methodology Approach
6.3.2 SP as Teamwork . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
55
56
57
58
59
62
62
63
64
65
66
67
68
69
Company Relocation as Demo-Case . . . . . . . . . . . . . . . . . . . . .

7.1 A Complex Decision Task: Relocation
of TRANS-IT Consult . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.1.1 Application of Systemic Principles . . . . . . . . . . . . . .
7.1.2 Consultant Report and Preparation for the
Decision Conference . . . . . . . . . . . . . . . . . . . . . . . .
7.2 Principles and Steps of the Decision Conference . . . . . . . . . .
7.2.1 The Principles of Decision Conferences. . . . . . . . . . .
7.2.2 The Relocation Decision Conference: Decision Tree
and Pairwise Comparisons . . . . . . . . . . . . . . . . . . . .
7.2.3 The Relocation Decision Conference: Scores, Weights
and Trade-Off. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.2.4 The Relocation Decision Conference:
An Intermediary Note . . . . . . . . . . . . . . . . . . . . . . .
7.2.5 The Relocation Decision Conference: MCA Results
and CBA Versus MCA Trade-Off Analysis . . . . . . . .
..
71
..
..
71
72
..
..
..
73
76
76
..
78
..
82
..
85
..
85
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Contents
xi
7.2.6
The Relocation Decision Conference: Scenario

and Risk Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.3 A Summing up of Case Findings. . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
95
95
95
97
108
109
112
112
115
116
119
Appendix A: COSIMA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
123
Appendix B: SIMDEC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
149
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
165
A Summing up: The Challenge of Strategic Decision Making

8.1 Systemic Planning: Practice and Cases. . . . . . . . . . . . . . .
8.1.1 From Detached to Involved Understanding. . . . . . .
8.1.2 A Review of SP Cases. . . . . . . . . . . . . . . . . . . . .
8.2 The Validity and Potential of the SP Framework. . . . . . . .
8.2.1 The SP Framework . . . . . . . . . . . . . . . . . . . . . . .
8.3 Complex Strategic Choices in a Wider Perspective . . . . . .
8.3.1 Known and Unknown . . . . . . . . . . . . . . . . . . . . .
8.3.2 Towards Robustness in Strategic Complex Choices.
8.3.3 SP from Here? . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
90
92
93
Abbreviations
AHP
AI
BCR
BS
CAD
CBA
CDM
CEA
CEO
CF
CG
CGE
CLG
COSIMA
CPM
CSH
CSR
CST
CV
DA
DE
DM
DSS
DTU
EIA
FA
FRA
FW
FYB
GIS
Analytic hierarchy process

Artificial intelligence
Benefit-cost rate
Brainstorming
Computer-aided design
Cost-benefit analysis
Customised decision model
Cost-effectiveness analysis
Chief executive officer
Conference facilitator
Certainty graph
Computable general equilibrium
Centre for Logistics and Goods Transport
Composite model for assessment
Critical path method
Critical systems heuristics
Corporate social responsibility
Critical systems thinking
Certainty value
Decision analysis
Decision engineering
Decision-maker(s)
Decision support system
Environmental impact analysis
Financial analysis
Feasibility risk assessment
Futures workshop
First year benefits
Geographic information systems
xiii
xiv
IP
IRR
ITS
MCA
MCS
MM
MOE
MP
NPV
OR
PA
RA
RCF
REMBRANDT
ROC
ROD
RSF
SA
SC
SD
SIMDEC
SMART
SMARTER
SP
SROI
SSM
STA
SW
SWOT
TGB
TRR
TSI
TV
UU
VF
VTS
Abbreviations
Interactive planning
Internal rate of return
Intelligent traffic service
Multi-criteria analysis
Monte Carlo simulation
Mind mapping
Mode of enquiry
Multiple perspectives
Net present value
Operations research
Preference analysis
Risk analysis
Reference class forecasting
Ratio estimations in magnitudes
or deci-bells to rate alternatives
which are non-dominated
Rank order centroid
Rank order distribution
Reference scenario forecasting
Scenario analysis
Strategic cognition
Systems dynamics
Simulation and multi-criteria analysis
for decision making
Simple multi-attribute rating technique
Simple multi-attribute rating technique
exploiting ranks
Systemic planning
Social return on investment
Soft systems methodology
Stakeholder analysis
Swing weights
Strengths, weaknesses, opportunities
and threats
Traffic Plan GreenlandDecision Tool
(transl. from Danish)
Total rate of return
Total systems intervention
Total value
Unknown unknowns
Value function
Vessel traffic service
Chapter 1
Introduction: Complexity as a Challenge
1.1 Complex Strategic Choice and Systemic Planning

Is robust strategic decision making possible in todays world? Can procedures be
provided that can support and add value to the efforts made in this respect? Such
questions can rightly be asked after a first decade of the twenty-first century where
unimaginable events ranging from the terrorist attack on the Twin Towers in New
York in 2001 to the financial crisis not yet overcome in 2011 have given rise to
cascades of uncertainty in a still more globalised world. In various ways such
grand-scale events are linked to local changes and thereby they contribute to an
increasing uncertainty which impacts on decision making. Such a globallocal
interlinkage that complexifies the decision environment may be captured in one
term and referred to as glocalisation.
Not many companies and organisations remain unaffected in this respect but are
challenged by glocalisation uncertainty and turbulence as concerns their strategic
endeavours. It therefore makes sense to see a complex world as an enduring and
growing challenge for planning and strategic decision making. More than a decade
ago the Danish communications researcher Lars Qvortrup interpreted rising
societal complexity in the following way:
We are dealing with an increasingly complex world which is todays big challenge. We are
as individuals and as communities exposed to an immense complexity. We are, as the
German social critic Hans Magnus Enzensberger wrote some years ago, bystanders to
countless civil wars via the media. We are exposed to legislation which before it is
problematic in its implications, is problematic due to its impossibility to penetrate. We live
in a societywhether we like it or notwhich is an integral part of the world community
and which cannot be reduced to a distinction between those belonging to the immediate
local area versus the others. We are part of a world where currency fluctuation in
Singapore is affecting employment in [the Danish region] Zealand (Qvortrup 2001, p. 50
in transl.).
S. Leleur, Complex Strategic Choices, Decision Engineering,

DOI: 10.1007/978-1-4471-2491-7_1, Springer-Verlag London 2012
1 Introduction: Complexity as a Challenge
1.1.1 The Idea of Systemic Planning

The purpose of this book is to present systemic planning (SP) as an idea and method
that can be applied with a view to making better strategic decisions. The idea of
using systemic planning for strategic decision support is developed initially from
generic ways of seeing and understanding into processes and methodology that
can assist decision-makers in dealing with complex planning and decision making
tasks. In short we can perceive the entire set of proposals from ways of seeing to
ways of doing as a framework that will be referred to as systemic planning. In this
way the SP framework comprises several levels that are interconnected. The
methods and techniques that are finally suggested and exemplified thus link
backwards to the more general approach suggested. The naming of this approach as
systemic derives from the way soft and hard methodologies are found suitable
to be used in combination for what becomes a kind of holistic handling of complex
planning problems. The term holistic implies some kind of completeness so that in
principle everything is taken into consideration. Evidently this isand will
remain soan ideal. Where known unknowns can be grasped by different ways of
exploration, the unknown unknowns are less easily dealt with. Despite a declared
holistic orientation aimed at with systemic planning we must bear this in mind.
The suggested holistic organisation of planning and decision making activities
can be perceived as a necessary countermeasure in a world appearing as increasingly uncertain and complex. With hard methods included, systemic planning also
comprises the use of systematic method-elements as part of the comprehensive
planning process. The term systematic is here used for planning and decision making
carried out in a pre-fixed, schematic way. Such planning typically applies to tasks
concerning scheduling, allocation, etc., which in organisations theory is indicated
as planning at the tactical and operational levels. It is typical for planning at the
strategic levelbeing the focus of this bookthat the change processes are of an
open-ended nature. When addressing such open-ended change, systemic planning
makes it relevant, as shown later, to combine hard, calculation-oriented methods
with soft, more interpretative and creative methods. The interaction of hard and soft
methods is one of the important features of the systemic approach to planning.
1.1.2 The Framework of Systemic Planning

The overall framework of systemic planning (SP) for strategic decision making
consists of the following interlinked levels:
Level 1: Coming to grips with complexity by combining different ways of seeing
Level 2: Designing the SP learning process
Level 3: Specifying the SP toolbox
Level 4: Demonstrating SP on a strategic decision making case
1.1 Complex Strategic Choice and Systemic Planning
The four levels are treated as follows: Level 1 of the SP framework is treated
in Chaps. 2 and 3, level 2 in Chap. 4, level 3 in Chaps. 5 and 6 and level 4 in
Chap. 7. This chapter and Chap. 8 respectively serve to introduce the topic of
complex strategic choices and to summarise and validate its findings together
with some advice on practice for readers who may wish to make use of SP for
addressing complex strategic choices. The contents of the chapters are overviewed below.
1.2 An Overview of the Contents

This chapter, Complexity as a challenge, primarily aims at presenting the
purpose of systemic planning and the challenge of strategic decision making in a
world that presents itself as complex. Initially systemic planning is set out as the
holistic handling of complex planning seeking to include both hard and soft
methodology. The chapter ends with taking a first look at complexity by treating
what is termed detail, dynamic and preference complexity, which leads to a discussion of different problem types. This provides the background for encircling the
possibilities and limitations of systemic planning as a foresight problem handling
approach. These issues will be returned to in the final Chap. 8 when the complete
SP framework has been set out and is validated and perspectivated.
Against this background Chap. 2 goes into depth with The condition of
complexity. The focus is on the German sociologist Niklas Luhmann and his
voluminous work on social theory which often makes a surprising and overwhelming impression on a newcomer to his theory. The chapter focuses on various
concepts that can help establish a theoretical basis for a closer understanding of
planning as related to complexity and thereby reflected in his universal theory of
social systems. This theory offers a number of theoretical findings that systemic
planning can make use of for the grounding of its principles and methodology.
The following Chap. 3 on Linking complexity and simplicity seeks basic
ways of seeing and understanding by introducing and interpreting two basic epistemic lenses we can reflect upon and apply. To judge the relevance of the SP
approach it is necessary to be explicit about how knowledge and insight can be
gained and this is what epistemology is about. In the theory of science the term
paradigm is used for a specific type of cognition and related research designs that
have established themselves as being a sound approachor more technically:
to be valid and constitute a relevant approach in the specific context. For illuminating the basic approach behind systemic planning the French science theorist and
sociologist Edgar Morin is called upon. Attention is first given to his Simplicity
paradigm which may well be said to represent the type of education and training
that economists and engineers face in their university years. Parallel to this, with
the Simplicity paradigm seen collectively as a particular type of epistemic lens,
Morin has also formulated a Complexity paradigm concerning an alternative way
of knowledge gathering and creation; also this paradigm can be seen to represent a
particular epistemic lens. In systemic planning this Complexity paradigm functions as a complement to the Simplicity paradigm.
Chapter 4 on The systemic process introduces the American brothers Hubert
and Stuart Dreyfus by treating their theory of learning. With a background in
philosophy and in operations research (OR) respectively, they have formulated a
learning process into five steps, which deals with what they themselves
acknowledge as representing a development from novice to expert. Although
apparently not familiar with Morins Simplicity and Complexity paradigms, they
describe a learning cycle which in its progression can almost be seen to build on
the paradigm of simplification in combination with the paradigm of complexity. In
this context it leads to the idea of designing a kind of systemic, self-organising
learning process to be made use of in the set-up of systemic planning.
With the theory outline of systemic planning presented in the above chapter,
Chap. 5 on The systemic toolbox addresses the more methodical and practical
aspects of SP. Specifically a number of useful OR methods and techniques are
introduced as a kind of long list for methods of possible relevance for SP. These
are categorised and presented in the form of three waves, with each new wave
we are still in the third wave since the early 1990srepresenting an important new
approach direction. From the long list of OR methods seven hard and seven soft
methods are identified, which are seen to provide a suitable method arsenal in the
subsequent practical adaptation of systemic planning.
In the following Chap. 6 on Setting up the decision support focus is set upon
applying different modes of enquiry based on the findings in the previous chapter.
Next the scoping of strategic choices is addressed and afterwards the assessment of
consequences and risks is treated. Emphasis is placed on describing the interrelatedness of scoping and assessment in the process of establishing adequate
decision support for complex strategic choices. Based on the specific purposes of
scoping and assessment a number of both hard and soft methods are treated. This
chapter ends with an outline of what can be seen to characterise a suggested
concept of choice intelligence.
The following Chap. 7 on Company relocation as demo-case serves to
exemplify systemic planning in the form of a description of a complex planning
task concerning the relocation of TRANS-IT Consult. The case concerns the
application of SP for selecting a new company headquarters location in the
resund region with many different factors influencing the final choice among a
set of eight pre-screened possibilities. The case description includes the various SP
steps and considerations leading towards a final strategic decision about the most
attractive new location.
The concluding Chap. 8 A summing up: The challenge of strategic decision
making first reiterates some of the main concepts of SP and afterwards ten cases
where SP has been applied are reviewed. This forms the background for a subsequent assessment of the validity and potential of the SP framework. Afterwards
complex strategic choices are put into a wider context, where issues about known
and unknown and risk-related Black Swan theory are made use of to indicate what
types of challenges organisations and companies may face with regard to long-term
1.2 An Overview of the Contents
planning and complex strategic decision making. Finally some conclusions are
presented together with a developmental perspective on SP.
The book also contains two appendices giving a more technical description on
how cost-benefit analysis (CBA) and multi-criteria analysis (MCA) can be combined in the COSIMA methodology in Appendix A and multi-criteria analysis and
risk analysis (RA) in the SIMDEC methodology in Appendix B. The CBA, MCA
and RA are all included in the systemic toolbox described earlier in Chap. 5.

In the presentation of the ideas behind systemic thinking and planning in this book,
it is maintained that at least three types of complexity need to be taken into account
in planning and managerial strategic decision making. These are:
Detail complexity
Dynamic complexity
Preference complexity
These three types play a major role in the way we proceed to formulate the
principles of systemic planning concerning how we seek to come to grips with
foresight problems. Each type of complexity will be treated below to assess its
meaning.
1.3.1 Detail Complexity

Looking at a specific planning problem from the viewpoint of a company for
example, we need to demarcate the socio-technical system that the company
represents. Basically, we need a certain level of precision about the system we are
dealing with, i.e. we need to lay down various issues determining our planning
problem with some certainty. To introduce the concept of detail complexity in a
somewhat formal way, we may assume just for the sake of symbolic illustration
that one of the attributes we need is the length of some system element. As
engineering students learn about measurement theory in a physics course, we can
determine the length of a rod by making use of still better measurement equipment.
Similarly, seeking to determine its weight (or any other attribute) we can refine our
approach by adopting a better technique. We try to remove uncertainty simply by
getting more precise data. However, if we refer to the work of Benoit Mandelbrot
and Helge von Koch, see Gleick (1987), and take a fractal view, we can never
determine the length of a coastline, for example. In Fig. 1.1 this has been exemplified by the graph known as the Koch curve, which, in Mandelbrots words, can
be seen as a rough but vigorous model of a coastline (Gleick 1987, p. 99).
Fig. 1.1 The Koch curve as

a fractal resulting from an
endless process of iteration.
The Koch curve is made by
starting with a side of length
1 and then adding a triangle
with sides equal to one-third
and so on. The length goes
towards infinity. Adapted
from Buchanan (2001, p. 48)
Detail complexity helps us focus on the influences from the system demarcations and the system components as they enter at an early stage in our examinations and/or models. Seemingly, the system is something that is given at the
beginning of a study. This view, however, is much too simple because beneath its
mere representation the system is also the result of a history that has frozen into
the concrete system elements and their interrelations. To demarcate the system
properly, we need to become aware of the details and their possible meaning and
influence. This kind of awareness is made explicit to us through the work of
physicists on complex systems. One important finding is that so-called critical
states are ubiquitous. In socio-technical systems the occurrence of critical states is
often what makes problems wicked.
The need to pay attention to the details is well argued in the quotation below
from Mark Buchanan, a theoretical physicist now working as a science writer:
By studying the natural kinds of patterns that evolve in networks of interacting things
under non-equilibrium conditions, we may be able to understand an immense range of
natural phenomena, from our turbulent atmosphere to the human brain. The study of
complex systems is all about things that are out of equilibrium, and on this task, of course,
scientists are really just starting out. So the relationship between the critical state and
complexity is really quite simple: the ubiquity of the critical state may well be considered
the first really solid discovery of complexity theory.
And yet there is another useful way to look at all this. In coming to consider complex
systems, physicists seem to have gained a new appreciation of a simple fact: in the
immediate world around us, history is important. For living things, which ultimately
develop from a single cell, this is obvious. But one cannot even understand the hardness of
a steel pipe, or the irregular surface of a fractured brick, without referring to the full
history of its making (Buchanan 2001, p. 16).
There is no doubt that system demarcation or boundary setting becomes

problematic when the presumably deep information contained in the various
system elements may or should impact on it, but at the same time it is of utmost
importance. Clearly the ubiquity of critical states makes it even more important.
Thus there is no right way of doing the system demarcation, e.g. of fixing the
dimension of detail complexity; the boundary-setting is a matter of choice and
boundaries are partial as described by the well-known systems researcher Robert
Flood:
Defining an action area from the problem context through sweep-in and unfolding, centres
on drawing boundaries around possible clients, and consequently surfacing issues and
dilemmas relating to those clients for discussion. Boundary setting is an issue of great
importance to systemic thinking. Put succinctly, the questions are, Who is embraced by
the action area and thus benefits? Who is out and does not benefit? What are the possible
consequences of this? And, how might we feel about that? Boundary setting thus raises
questions of ethics, efficiency and effectiveness, in a search for improvement and shows
them to be inextricably linked. Boundaries are always open to further debate through
sweep-in and are thus temporary. Boundaries are the result of choice. For each choice
located by unfolding, there are always other possible options that will arise by sweeping
in. Boundaries are therefore partial. The temporary and partial nature of boundary setting
is suggestive of improvements to make, for now, but raises the question of how
improvement is to be secured (Flood 1999, pp. 6465, underlining added).
For these reasons boundary setting becomes a major influence when defining an
adequate action area to identify, for example, the possible means to secure
improvement. At this stage when presenting the theory behind systemic planning
to be applied for providing decision support, it suffices to say that the demarcation
of the socio-technical system that we are examining is by its nature less given
than first impressions might suggest.
The medium in which detail complexity operates is typically space (covering
resources such as persons and their skills, physical facilities, financial resources,
etc., which make up the variables in this space). Clearly with many variablesand
with each variable possibly having many attributes of relevance and with interdependence between variablesthe detail complexity becomes full-fledged.
Briefly stated, detail complexity relates to concerns about means.
1.3.2 Dynamic Complexity

If we look at the theories in Peter Senges The Fifth Discipline from 1990, we find
that Senge operates with a complexity notion that involves both temporal aspects
(complexity associated with dynamics) and detail complexity consisting of a
large number of variables being relevant but difficult if not impossible to combine
and process at the same time. To the surprise of some, Senge and his collaborators
Fig. 1.2 Weather sequences in a computer model: the Butterfly Effect by Lorenz. In 1961 the
meteorologist Edward Lorenz found that small differences in starting conditions could mean a
considerable change in end result. Thus a storm at one location may be seen as initiated by a
butterfly flapping its wings and thereby causing a small disturbance up-stream of the weather
pattern propagation that resulted in the storm. Adapted from Gleick (1987, p. 17)
give less attention to detail complexity than to dynamic complexity (Senge 1990;
Flood 1999, pp. 1314).
In Fig. 1.2 the importance of dynamic complexity is illustrated by comparing
the development of two weather patterns.
With nearly the same starting point, the two patterns diverge over time and end
up with no resemblance at all.
The work of Edward Lorenz in the 1960s was very important in initiating
research on chaos in dynamical systems, although deterministic chaos as a
phenomenon had been known for many years due to the work of, among others,
the French mathematician Henri Poincar around 1900. The use of computers
has come to play a major role in the research that started with the findings of
Lorenz.
With the focus on planning and decision making, we have to interpret the
importance of dynamic complexity by the way it makes long-term forecasting a
highly doubtful undertaking. But many further insights are implied when we
examine complex dynamic interrelationsnot least if our focus is more on human
organisations and their development than on weather pattern propagation. Perhaps
in this organisations context we should see the Butterfly Effect as a storm
started, for example, by the whispering of a rumour one afternoon at the coffee
machine.
The organisations and chaos researcher, Ralph Stacey, has given the following
interesting interpretation of organisational time dynamics (speaking in the context
of the phenomenon of change) by making reference to the so-called leverage
points introduced into systems vocabulary by Peter Senge. With a focus on
studying business units as complex, dynamic systems Stacey says:
The study of complex, dynamic systems provides the insight that the behaviour of a system
cannot be understood simply by examining the systems parts. The system in effect has a
life of its own. The system itself has a major impact on behaviour and therefore on
outcomes. Thinking therefore has to proceed in terms of whole systems, their interconnections, and the patterns of behaviour they may generate. Changes accumulate slowly out
of the interconnections between a systems parts. Focusing on snapshots of the parts,
looking for cause-and-effect links that are close together in time and space, means missing
the slow accumulation of change. Instead of trying to understand quantitative detail of
parts, therefore it is far more fruitful to try to understand the qualitative nature of
interconnections and patterns of behaviour. It is especially helpful to try to find the points
in the system that are most sensitive and amplifyingthe points of greatest leverage. By
operating at these points rather than trying to control details everywhere, managers can
bring about the greatest changes in the system with the least effort (Stacey 1993, p. 110).
Peter Senge and his collaborators have managed to identify a number of what
he calls archetypes of change, which are dynamic organisational patterns. One
of these is the Tragedy of the Commons, which occurs when two systems
operate in the same environment and are rewarded initially by exploiting the
environment (Jackson 2000). The tragedy of the commons was originally coined
by Hardin (1968) in an article in the journal Science, in which he examined
individual actions and their cumulative consequences which, in an unwitting way,
could be systematically destructive for the socio-economic unit made up of the
individual actors. His picture was the medieval English village where each
householder made the apparently reasonable decision to graze as many cattle on
the commons as possible with the result that the commons would suffer overgrazing, leaving each and every householder in a poorer condition.
There is no doubt that a number of archetypesPeter Senge operates with
around a dozencommunicate what we would like to see as collectively gained
lessons that are of importance with regard to interpreting possible development
patterns. Clearly, they play a major role for the manager who does not want to
embark on some kind of course that may later turn out to be less desirable for some
reason. However, downplaying the checking of details cannot generally be
recommended, see the quotation above, nor can the belief that a relatively limited
number of archetypes is capable of unfolding a larger part of the dynamic complexity relating to change in business units or to change in socio-technical systems
in general. As with detail complexity and its means-uncertainty when defining the
action area around complex strategic choices, dynamic complexity indicates
another major type of uncertainty as a basic condition.
The medium in which dynamic complexity operates is time and stated briefly
dynamic complexity relates to concerns about path.
1.3.3 Preference Complexity

The ideas about systemic planning consider at least three types of complexities
which are important for our understanding of the conditions for future-oriented
decision making in organisations. One major influence on these ideas is Herbert
10
Simon via his writings about organisational decision making (1968) and the
sciences of the artificial (1969), the first major exposition of the meaning and
consequences of applying the view of organised complexity on organisations as
systems. Another major influence is Jrgen Habermas (1979, 1986, 1989) with his
elaborate theory on communication. The following quotation by McCarthy from
his magnificent book The Critical Theory of Jrgen Habermas brings the third
type of complexity to the fore, namely what in this presentation is termed the
complexity of interests (following the German term) or simply preference
complexity. Thus McCarthy with reference to Habermas states that:
a precondition of rational consensus is the thematisation of available need interpretations themselves; interests are neither empirically found nor simply positedthey are
shaped and discovered in processes of communication (McCarthy 1981, p. 328).
What the theory of Habermas states here is that preferences (interests) are
tied up with processes of communication and are therefore quite dependent on the
issues raised and debated. To deal with the complexity involved and get to grips
with the interests that might be associated with the various stakeholders, we need
to understand the processes of communication. Normally when referring to theory
of communication we have the work of the mathematician Claude Shannon in
mind. Contrasting, however, Shannons theory of communication with the version
stemming from Habermas we find that the two theories of communication are
completely different. Making use of the complexity notions I have introduced, we
might say that Shannons theory deals with a measurement of message transmission (with the complexity issues involved then relating more to the notion of
detail complexity), whereas Habermass theory examines the basic components of
human language and interaction (based on what he calls validity claims). For our
purpose the theory of Habermas gives the valuable insight that preferences are not
ready-made and accessible for strategic decision making but have to be shaped
and discovered. As can be seen later on this insight has a very practical imprint
on the way that systemic planning is carried out, namely as a kind of search-learndebate process.
The notion of preference complexity can be illustrated as shown in Fig. 1.3
which indicates the shaping of an interest in a symbolic way.
The figure illustratesin a symbolic way onlythe complexity involved when
shaping and defining an interest. In part (a) we have all possible fragments and
influences which in part (b) have obtained a certain degree of common orientation.
What the figure really shows is atomic magnets in a piece of iron: at high temperatures (a) they cannot line themselves up due to thermal jostling, but at lower
temperatures (b) they are able to align with the result that the iron becomes
magnetic. For the purpose here as mentioned the figure is only symbolic and
illustrative. Thus we can perceive the figure as showing a heated debate that may
(or may not) be cooling off and lead to clarification and explication of a certain
interest, depicted as a change from (a) to (b).
Preference complexityand the set of related issues to be addressedhas not
had its proper role in the development of concepts and tools in systems science
11
Fig. 1.3 The shaping of an

interest, symbolically
illustrated by magnets in a
piece of iron. At a high
temperature (a) atomic
magnets in a piece of iron
cannot organise in a coordinated way, whereas this
becomes possible in (b) at a
lower temperature, when the
iron functions as a magnet.
The arrows are here used to
illustrate discordance and
concordance in a certain type
of preference. Adapted from
Buchanan (2001, p. 73)
(Leleur 2008). This may be due to the main professions involved in its development over the five-to-six decades since the Second World War, with scientists,
engineers and economists dominating with regard to theory and practice on the
basis of the terms and premises of their educational background. However, similar
to detail and dynamic complexity when defining an adequate action area around
complex strategic choices, preference complexity is an important issue.
The medium of preference complexity is mind and briefly stated, preference
complexity relates to concerns about ends.
In this way we have obtained complexities that operate in space (detail complexity), in time (dynamic complexity), and in mind (preference complexity).
Later onafter dealing with the basic theories behind the complexitieswe will be
able to see that exploring complex strategic choices relates clearly to all three concerns: means, path and ends. A main theme of the book is the attention decisionmakers in general will have to pay to complexity issues. I agree with Senge in
believing in the importance of dynamic complexity, but pay at least as much attention
to detail complexity. I pay attention to preference complexity, because I have come to
believe that insights into this type of complexity have a special role to play when
addressing strategic decision making. The theories and methods that are presented in
the following chapters all relate to the impact that complexity has on the way decision
support can be provided.
In the following section I end this introductory chapter by addressing a general
classification of problem types of relevance for planners and decision-makers. In
the final chapter of the book this classification will be reconsidered to take account
of the ways in which the systemic planning approach set out can be seen to add
value to current knowledge about ways of qualifying the process and methodology
relating to making complex strategic choices.
12
Table 1.1 Problem types relating to the configuration of means and ends
Problem types
Four different configurations and related approaches
Means/ends
Certain
Uncertain
Certain
A: Computation
B: Judgement
Uncertain
C: Compromise
D: Chaos or Inspiration
Adapted from Khisty and Mohammadi (2001, p. 22)
1.3.4 Categories of Problems

So far, planning problems of managerial and professional concern have been
addressed as one common category. However, we can take a closer look at them
by considering the means-ends configuration shown in Table 1.1:
In the A situation, where we have certainty about both means and ends, our
problem type is one of computation. Input can be stated and by using a proper
algorithm we are able to obtain a solution to our problem. A very simple example
here is a journey from one location to another: we can go by car, bus or train or
some combination, and we know the time when we want to arrive. By consulting a
travel schedule website, for example, we can obtain a selection of the best travel
schedules, maybe including modal shifts, and we can decide which possibility is
the most attractive from the calculated number of minutes for each alternative and
its cost. We can even take comfort issues, etc. into account and obtain a best
solution in accordance with our trade-offs between time, costs and other issues we
handle in an implicit way.
As soon as uncertainty characterises either means or ends, things start to get
complicated: our choice may incline towards the car because we know that the
public transport means, bus and train, operate only with some certainty in the peak
hours for example. Or we may be in a situation where we are a little bit uncertain
about our end point because, for example, we may want our recreational trip to
take us to a place where fishing is goodand if not, we want to be able to continue
to another location and so on.
On a scale we may see A as a conventional planning problem, whereas B and C
represent stages towards the D situation, which I will characterise as a full-fledged
complex planning problem. In this book we will be concerned with all four situations, but the applicability of a systemic approach to planning and strategic
decision making is first and foremost relevant with situations on the way to being a
complex planning problem or one which already is a complex planning problem,
characterised by situations B, C and D, respectively.
Table 1.1 gives quite important information about the challenges of strategic
decision making in a complex world, where at least detail, dynamic and preference
complexity are of influence. At this stage of presenting systemic planning as an
approach to handling foresight problems we may state that type A problems in
many cases are handled well by using hard operations research methodologies for
13
optimising, scheduling etc., whereas type D are the utmost challenges as it

includes also black holes of the future representing the unknown unknown.
With this in mind there might be something to add to our planning and decision
making capabilities for the foresight problems categorised by B and C. Awareness
of what characterises A and D will be of utmost importance in this respect.
In summary of this introductory chapter we may state that the scope of this
book is to provide decision analysts and planners with concepts and tools to
support decision makers facing complex strategic choices that are dependent on
judgement, compromises and inspiration.
Main points and findings of this chapter
Complexity is a real concern for understanding todays challenge of strategic
decision making.
Three types of complexity are described: Detail complexity, dynamic complexity and preference complexity. Each type is illuminated by a symbolic
example to highlight the radical uncertainty that theyseparately or collectivelycan inflict on strategic decision making.
Ordinary concepts in planning such as means, path and ends are under the
perspective of strategic decision support influenced respectively by detail
complexity, dynamic complexity and preference complexity.
Looking at different problem types a finding is that conventional planning
generally in the book referred to as systematic planningshould be reserved for
some well-defined problems, whereas other problems should be approached on
the premises of their complex nature. This makes it necessary to come to grips
more generally with the condition of complexity, which is the purpose of the
following chapter.
References
Buchanan M (2001) Ubiquitythe science of history or why the world is simpler than we think.
Phoenix
Flood RL (1999) Rethinking the fifth disciplinelearning within the unknowable. Routledge,
London
Gleick J (1987) Chaosmaking a new science. Viking, New York
Habermas J (1979) Communication and the evolution of society. Heinemann, London
Habermas J (1986) The theory of communicative action, Volume 1: Reason and the
rationalization of society. Polity Press, Cambridge
Habermas J (1989). The theory of communicative action, Volume 2: The critique of functionalist
reason. Polity Press, Cambridge
Hardin G (1968) The tragedy of the commons. Science 162:12431248
Jackson MC (2000) Systems approaches to management. Kluwer Academic, New York
14
Khisty CJ, Mohammadi J (2001) Fundamentals of systems engineering with economics,

probability, and statistics. Prentice-Hall, Upper Saddle River
Leleur S (2008) Systemic planning: principles and methodology for planning in a complex world,
2nd edn. Polyteknisk Forlag Lyngby, Denmark
McCarthy T (1981) The critical theory of Jrgen Habermas. The MIT Press, Cambridge
Qvortrup L (2001) Det Lrende Samfund. Gyldendal, Copenhagen
Senge P (1990) The fifth discipline: the art and practice of the learning organization. Random
House, London
Simon HA (1968) Administrative behavior. The Macmillan Company, New York
Simon HA (1969) The sciences of the artificial. The MIT Press, Cambridge
Stacey RD (1993) The chaos frontiercreative strategic control for business. ButterworthHeinemann Limited, Oxford
Chapter 2
The Condition of Complexity
2.1 The Systems World of Luhmann

The German sociologist and science theorist Niklas Luhmann (19271998) is an
interesting theorist whose work has attracted great attention in very different
professional fields. Attention in this context will naturally focus on his possible
impact on the formulation of concepts and principles that can contribute to ground
a framework for systemic planning (SP).
Luhmann has not least with his theoretical opus magnum Social Systems from
1984 (Luhmann 1995) formulated a coherent theory of social systems based on a
number of assumptions and key concepts that allow him to illuminate issues that
have impacted on later social science development in many ways (Moeller 2006).
The use of Luhmann in this book is highly selective, but reference can in addition
to those mentioned above be given also to (Leleur 2008a, b), where his work is
treated in a more comprehensive way with its value for planning theory in mind.
For the purpose pursued here the following quotation from Moeller can introduce
Luhmanns radical way of thinking:
From the constructivist perspective of systems theory, reality is not something given, but
an effect of cognitive construction. There is not one reality, no one realm of being, as in
the traditional Old European ontologybut rather a plurality of realities created through
cognition. Cognition produces reality by producing system/environment distinctions.
(Moeller 2006, p. 70).
System/environment distinctions are a main focus in the part of Luhmanns

theory that I will make use of here. Other import key concepts are autopoiesis and
contingency.
Autopoiesis of Greek origin means something that creates itself, i.e. selfcreation, while contingency describes conditions that are the result of a random
outcome, i.e. something that also under the given conditions could have been
otherwise. Those concepts are discussed further below.

15
16
2.1.1 Systems According to Luhmann

To understand the constructivist perspective of Luhmann, we first need to come to
grips with the way he actually perceives a system. In fact, he makes use of a new
important perception.
Within the field of systems theory we first had a system as an entity consisting
of a set of elements, which can be called the classical perception (Bertalanffy
1972, 1973; Leleur 2008b, c).
Next, associated with the work of von Bertalanffy, a system is perceived as
open because it thrives on energy and information flows across the system/
environment border to develop/steer towards a higher, more negentropic, state.
Finallywith the work presented by Luhmann in his Social Systems from 1984
(Luhmann 1995)we have based on what is called Luhmanns autopoietic turn
systems as isolated entities. However, in the vocabulary of Luhmann, such systems
can enter into structural coupling with each otherand be irritated (i.e.
affected)but basically systems in his perception are to be seen as operationally
closed. I will give examples, but first we need to understand that Luhmann
operates with four different types or categories of systems set out as follows
(Ibid., p. 2):
Social systems which are subdivided into: societies, organisations and interactions
Psychic systems
Organisms and
Machines
For the upper three types of systems, Luhmann sees the autopoietic perception
as relevant, whereas machines are categorised as allopoietic, i.e. their different
parts or units are not maintained and produced by themselves during the operation
of the machinethey are excluded as part of the machine system in such a way
that they can be replaced if, for example, a certain unit is worn out. This is why the
originators of autopoietic systems thinking, the Chilean biologists Varela and
Maturana, see only living systems as autopoietic.
Social and psychic systems are prominent as concerns the attention they are
given in Luhmanns work. Basically, these systems are characterised by their use
of meaning, something which is not attributed to the categories of organisms and
machines. For our purpose I will concentrate on the social systems category, but
the perception of a system being operationally closed can be illustrated in the
psychic systems category: the nervous system of the brain is a self-referential,
closed system consisting of neurons. An event does not directly determine neural
activity, but may have an effect that can be seen as a kind of irritation. The closed
nature of the system is then not to be understood as a kind of preclusion but rather
as a possibility for openness. The neural system does not mirror the surroundings
but constructs an image of these through internal operations.
2.1 The Systems World of Luhmann
17
The basic features of autopoietic systems, briefly introduced above, make it

possible for Luhmann to formulate highly interesting views on systems and
complexity.
2.1.2 Systems and Complexity

It should be clear by now that due to his autopoietic turn Luhmanns approach to
systems thinking differs in many ways from positions held by other researchers in
the social science field, in this respect notably Habermas. The Danish philosopher
and communication researcher Ole Thyssen has described the differences in the
following way:
Autopoietic systems are outside each other and external to each other. They never overlap.
Human beings are not part of society, and the society does not exist for the purpose of
human beings. Language and culture are notas with Habermasthe sea in which
human beings swim around and which comprises social and psychic systems, but tools to
scan the streaming consciousness, to reduce the complexity and render communication
probable by supplying the participants in speech with resources they swiftly and easily can
draw upon. (Thyssen 1992, p. 26 in transl.)
How can operationally closed systems interact?This apparent difficulty is

answered by seeing them as structurally coupled. This means that the systems are
still to be seen as autonomous, but they make available their complexity in
temporary interlockings. In this way, through a kind of structured emergent
system, the occurring open possibilitiesi.e. presupposed free, unbound material
and energy, or the not-yet-fully-determined possibilitiesare to be used as
providers of meaning (Luhmann 1995, p. 221).
As noted Luhmann subdivides social systems into societies, organisations and
interactions. The interaction system appears simply when a group of people is
together, for example attending a seminar at a university, and it is constituted by
approval of agenda, oral presentations, a person in the audience talking with the
person next to him during a break, etc. The actions and interactions of people
outside the seminar room belong to the surroundings, or the system environment,
of the interaction system. At the end of the seminar, the interaction system is
dissolved, at least until the group of people meets again for a new seminar (Kneer
and Nassehi 1997, pp. 4647).
The organisational system forms another type of social system that is seen as
organised when certain conditions are fulfilled with regard to being a member of
the system. Organisations, such as public and private companies and enterprises
universities, engineering consultancies, kindergartens, IT companies, the Danish
Royal Theatre, Wagner Societies and Greenpeace that may seem to go beyond this
public/private demarcationare all seen as organisations under Luhmanns
perspective. This means they are engaged in determining special courses of
action, which could not be expected in the surroundings of the organisational
system in exactly that way, which is how it becomes possible for members and
18
non-members of the organisation to get an understanding of it and act accordingly

(Ibid., p. 47).
For Luhmann the most comprehensive social system is the societal system. All
interaction systems and organisational systems belong to the societal system, but
the societal system is not a kind of super-interaction system or super-organisational system; the societal system also includes actions between people who are
not present together in a group, and people cannot resign their societal belonging.
The societal system is made up of much more than the plenitude of interaction and
organisational systems, which meansdue to its multiplicity and varietyit
becomes, according to Luhmann, a system of a higher order, a system of a
different type (Ibid., pp. 4748 in transl.).
A striking insight set out in Luhmanns theory is that the societal system is seen
as functioning, among other things, for the purpose of reducing the complexity of
the societal world. Where previous system conceptions have been concerned with
the system as more than its parts, Luhmanns conception implies decentring
systems thinking by also seeing the system as less than its parts. By this he
means that we have no integrating way of interpreting social systems; there is no
centre of social systems endowed with some kind of power, from where social
systems could be reformatted. In fact the systemic world and systems as such are
without a centre (Harste 1992, p. 64).
2.2 System and Environment

Luhmann perceives complexity as being forced to select. In fact, this is only the
first step in a way of reasoning that points towards the concept of contingency
which is essential for understanding current and future planning conditions. The
meaning of contingency is laid down by Luhmann as follows:
Complexity, in this sense, means being forced to select; being forced to select means
contingency; and contingency means risk. Every complex state of affairs is based on a
selection of relations among its elements, which it uses to constitute and maintain itself.
The selection positions and qualifies the elements, although other relations would have
been possible. We borrow the tradition-laden term contingency to designate this also
being possible otherwise. It alludes, too, to the possibility of failing to achieve the best
possible formation. (Luhmann 1995, p. 25).
Recognising contingency as a condition of planning means recognising that

many explanations and many alternatives are possible at the same time. Does
planning not dissolve as a management strategy under such circumstances? What
might at first glance look like a blind alley when we reflect upon the meaning of
contingency and its possible influence on planning thinking, however, turns out to
have a kind of Midas touch (Ibid., p. 44). Since contingency denotes states
anywhere between what is necessary and what is impossible, the golden touch of
Midas that Luhmann refers to as characterising it can be understood as the richness
it makes possible in theorising about concrete manifestations in social systems;
19
these manifestations could have been otherwise and have led to other configurations with other interpretations, or with other understandings of these interpretations; other perspectives could have been chosen and in that case this or that
consideration or reflection would change. When we supplement the view of the
decentred approach to systems thinking with the views expressed above on the
notion of contingency, we arrive at some of the basic analytical tools in
Luhmanns treatment of social systems.
2.2.1 The Midas Touch of Contingency

A basic view expressed in Ashbys Law of Requisite Variety is that the variety in
response from a system should match the variety of the challenge in the environment, see, for example, (Checkland 1981, p. 88). But at the same time it should
be noted that according to Luhmann a social system is always less complex than its
environment. So one might expect it, for instance, to be defensive all the time as
a strategy of coping with the greater variety surrounding it. However, according to
Luhmann certain other possibilities exist. Noting first that temporal autonomy
indicates the capacity of bringing relationships of complexity into the form of
meaning by actualising what is temporally not actual, with the risk of
remembering and anticipating incorrectly, he outlines these possibilities as
follows:
If the relative temporal autonomy of a system is secured by one or another combination of
distancing mechanisms, then a system can use the temporal dimension to better solve the
problems of its own complexity and, above all, to increase its own complexity through
the use of time. We will call this the temporalization of complexity
Temporalization of complexity leads to a selective ordering of the connection between
elements in temporal succession. In a more abstract formulation, the capacity to make
selective relations can be greatly expanded if a system can establish an ordered difference
between connections in temporal succession, a change of relational models according to
internal and external demands. (Luhmann 1995, pp. 4647).
There seems to be some promise here for planning, if planning as a discipline

and activity can enable and empower the system by proactively increasing its owncomplexity so the system can better meet the upcoming challenges of contingency
and otherwise. The conditions for planning in a complex society can generally be
matched by its organisations, seen as a variety of socio-technical systems with the
common characteristic, however, that their present state is occasionally overwhelmed due to contingency. Such situations can also loosely be referred to as
information overload.
From a societal viewpoint, contingency ensures compatibility among numerous
subsystemsseen indirectly by the many rules and regulations we do not
needwhereas from the viewpoint of a single social system, for example some
kind of an organisation, contingency implies in principle that its deficit in system
complexity, compared with the complexity of its environment, will cause the
20
system to be on alert all the time. Being on alert on this basis and acting in
accordance is what management is about. SP, at least in those contexts where this
type of approach may be applied, can be seen as one of the tools available for
management, but an important one, as it is precisely aimed at assisting the management and the organisation in facilitating what Luhmann calls the temporalization of complexity.
Systemic planning, therefore, has a role that in some ways is different from
conventional planning: while the latter aims at controlling the march into the
futureseen as part of the present systems environmentSP accepts that this
environment is to some extent uncontrollable. More practice-oriented, it makes use
of a framework and methodology that make the organisation better prepared to
meet the future by recognising both its knowns and its unknowns.
2.2.2 Contingency and Decisions

Organisations need to produce decisions. Before a decision, we have contingency
represented as open options. Afterwards, we have contingency now relating to the
options closed and what the decision means in terms of change. So we can look
upon a decision as a transformation of contingency where uncertainty is changed
into risk, because there is always uncertainty about what will follow and whether
a decision can really be carried through or will be resisted, either by factors
actualised due to the alternatives foregone or more simply because some surprises
will appear (Thyssen 1997, p. 75). There is no way in which the organisation can
avoid making decisions because not acting should also be seen as a decision.
Everything in an organisation is successively being created by the organisation as
the result of ongoing decision making which is in need of decision support.
Planning provides one kind of decision support. The picture of the environment
that is used when taking the decision is not part of the environment, but part of the
organisation as it is constructed by the organisation. It is such a construction that
underpins the decision and as such it is the responsibility of the management
(Ibid., p. 85).
Since we live in a complex world, we cannot know everything. How should we?
But does that mean that the management is free of responsibility? When nobody
knows the consequences, how can we blame anybody? There seems, however, to
be a kind of paradox here as stated by Thyssen in the quotation below:
Non-knowledge seems to free people from responsibility, because how could one be
responsible for something that one in principle cannot know about? Nevertheless, it is
precisely therefore decision-makers are hired. We do not accept that they overplay their
non-knowledge and make themselves victims of the circumstances. Their task is to
undertake the responsibility and risk of the decision. Part of this is to construct the space
and time of the decisionits frame (Thyssen 1997, p. 85 in transl.).
The function of systemic planning is precisely to assist in such a construction in

cases where the organisation faces a complex planning task.
21
In the following the organisation will be perceived as a socio-technical system.

Of the many decisions defining the specific organisation some will relate to
strategic issues, and where these are set in a context of complexity, for example by
having long-term consequences and influenced by interwoven global/local (the
glocal) factors, the organisation will as part of its strategic work face futureoriented problems involving complex strategic choices. To examine the possibility
of supporting the decision-makers of the organisation in this respect is the main
focus in this book for which reason it is necessary to address how the sociotechnical system is related to its future. As regards the treatment below this may be
referred to in the plural as its futures.
2.3 The Socio-Technical System and Its Futures

The British management theorist Ralph Stacey has set out a relevant schema for
addressing different types of change. As noted in Sect. 1.3 Stacey perceives
business units as complex, dynamic systems. Altogether he operates with three
types of change: closed change, contained change and open-ended change, as
described below (Stacey 1993a):
Closed change. The key features of closed change are unambiguous problems,
opportunities and issues, clear connections between cause and effect, and the
possibility of accurately forecasting the consequences of change. Faced with such
change, people tend to behave in easily understandable ways. The decision-maker
can make use of rational decision-making techniques, and the processes of control
are formal, analytical and quantitative. There is a clear purpose with clear preferences and alternative ways of achieving the purpose are known.
Contained change. The key features of contained change derive from those change
situations where it is possible to make probabilistic forecasts based on actions
taken now and their most likely consequences. This is made possible because the
consequences appear to some degree as repetitions of what has happened in the
past or they relate to large numbers of essentially the same event. As a manager
looks into the future, accurately predictable, closed change declines in relative
importance, while less reliably predictable, contained change increases in relative
importance.
Open-ended change. Control in open-ended situations in practice means something
completely different from what it means in closed and contained situations. In such
situations, the future consequences are unknown and forecasting is totally
impossible due to an ambiguous purpose or equivocal preferences of the actors
involved. The whole situation being confronted is ill structured and accompanied
by inadequate information, more or less subjective, and conditioned by personal
ambitions, beliefs and values. There are problems with interpreting data and
applying statistical techniques in uniquely uncertain conditions, for which reason
forecasting and simulation become problematic. In open-ended change situations
we do not know the consequences of what we are doing until we have done it.
22
The latter statement about not knowing the consequences in advance of actions
to be taken is really one of Staceys strong points. His considerations have led him
to speak about the unknowable. His viewpoint is expressed in the quotation
below:
Everyone admits that the future is basically unknowable, particularly in the case of an
innovative product or course of action. This prospect, however, makes many managers
uncomfortable, and they then ease their discomfort by assuming that even innovative
futures are nonetheless approximately knowable. One can at least, they say, have a vision
or make some assumptions about the long-term future. One can give shareholders, or
others in a controlling position, meaningful information on future rates of return and risk
levels.
I argue that this is a soothing fantasy that distracts attention from, and weakens the
resolve to deal with, the real world. Instead of sidestepping the issue of unknowability,
managers must learn to face it head on. That means accepting that you really have no idea
what the long-term future holds for your organisation; forming visions and making
assumptions are not realistic possibilities. It means accepting that no individual or small
group can be in control of an organisations long-term future (Stacey 1993b, p. 7).
One can agree with many of Staceys findings in his comprehensive writings
about organisations and issues relating to change (Stacey 1993a, b; Stacey et al.
2000). However, I cannot agree with his very principal meta-finding quoted above
that one should recognise and accept that no individual or small group can be in
control of an organisations long-term future. On the contrary, the viewpoint
argued in what follows is that proactive effortseen as planning in its broadest
terms and handling complex strategic choices properlyis worthwhile. What is
basically meant by this is that it is necessary to scrutinise the consequences of an
action in advance of any concrete action when the consequences can be identified
and assessed. This is exactly the job of managers assisted by planners and decision
analysts.
Conventionally such a mapping of consequences has relied in a comprehensive
way on modelling and quantitative assessment, in this book referred to as systematic planning. This conventionalsystematicperception of planning implies
causality: if these means are used in a specific way, certain ends will result from
or be caused bythe set of planned actions. Such if then thinking belongs to the
generic idea of planning and strategy development.
However, since chaos and complexity theories entered management thinking in
the 1990s, the concept of causality in socio-technical systems has been under
attack. Above I used the work of Ralph Stacey to exemplify what may come out of
basing management thinking on these new theory constructs. I could have chosen
several authors of management literature, but Staceys books about chaos management published in the early 1990s stand out due to their penetrating insights
about organisations and not least their in some respects radical conclusions (Stacey
1993a, b). For our purpose here, seeking to come to terms with planning under
complex conditions, I highlighted the finding from this type of literature that no
individual or small group can be in control of an organisations long-term future.
This is certainly not the common belief among board members and CEOs and
23
therefore there is a need to dig deeper into this question. Summarising in Chap. 8
some plausible answers will be presented based on the findings throughout the pages.
With a focus now on causality, I will make use of Staceys more recent work
about complexity and management (Stacey et al. 2000). Due to debates with
several well-known researchers in the field, among others Jonathan Rosenhead,
Staceys argumentation on the implications of chaos and complexity theories for
management has somewhat changed. Stacey and his collaborators now recognise
that a deterministic, chaotic system may not resemble the ongoing affairs of human
organisations where many uncertainties could better be comprehended as being of
a stochastic type. The early writings of Staceythose used to introduce the
concept of dynamic complexity in Sect. 1.3focus some attention on the idea of
the unknowable future to be accepted on the premises chosen; we may see the
more recent book from 2000 by Stacey, Griffin and Shaw as a continued examination of the unknowable which is, among other things, carried out by digging
deep into the concept of causality. In fact the book, entitled Complexity and
Management, contains a thorough description and interpretation of causality in
Western thinking from Aristotle to the present day, in which, for example, the
theories of self-organisation worked out by the Belgian physicist Ilya Prigogine
and others are dealt with.
The major finding of Stacey and his collaborators centres around the concept of
transformative teleology. This is contrasted with the concepts of formative
teleology, which are concerned with the actualisation of form or self already there
in some sense and rationalist teleology concerned with realising chosen, pre-set
goals. Telos is the Greek word for goal or end, so teleology can be understood
as preoccupied with the overarching source of change (Ibid., p. 196). We can get
an understanding of the causal framework Stacey and his collaborators associate
with transformative teleology from the following quotation:
When Prigogine considers the wider implications of his work, we think he makes a clear
move to Transformative Teleology. At the beginning of his book, The End of Certainty
(1997), he poses what he sees as a central question: Is the future given, or is it under
perpetual construction? In the terms we are using, this translates into Is causality in
nature (including humans) better understood as Formative Teleology, or is it better
understood as Transformative Teleology? His answer to the question is very clear: he
sees the future for every level of the universe as under perpetual construction and he
suggests that the process of perpetual construction, at all levels, can be understood in
nonlinear, nonequilibrium terms, where instabilities, or fluctuations, break symmetries,
particularly the symmetry of time. (Stacey et al. 2000, p. 97).
The breaking of time symmetry plays an important role in the more technical
parts of Prigogines work (Prigogine and Stengers 1985, pp. 249341). The
interest pursued by Stacey, Griffin and Shaw, however, is in exploring their
conceptual implications as expressed in the quotation below:
[Prigogine] says that nature is about the creation of unpredictable novelty where the
possible is richer than the real. When he moves from focused models and laboratory
experiments to think about the wider questions of evolution, a move that many scientists
would question, he sees life as an unstable system with an unknowable future in which the
irreversibility of time plays a constitutive role. He sees evolution as developing bifurcation
24
points and taking paths at these points that depend on the micro details of interaction at
those points. Prigogine sees evolution at all levels in terms of instabilities, with humans
and their creativity as part of it. For him, human creativity is essentially the same process
as natures creativity and this is the basis for his call for a new dialogue with nature.
These features, unknowable futures emerging in here-and-now interactions, are essentially
what we have defined as the causal framework of Transformative Teleology.
Central to Prigogines approach, at all levels, is the distinction between individual
entities and populations, or ensembles, consisting of those entities. (Stacey et al. 2000,
p. 97).
When the self-organisation phenomena are interpreted later in the text with
emphasis on human organisations, the following conclusion is reached:
the dominant management discourse, including systems thinking, is built, explicitly or
implicitly, on Rationalist Teleology as an explanation of choice This is expressed in
psychological theories that accord priority and primacy to the choosing individual over
the social. It is a view of minds as information processing devices that make representations of a pre-given world, formed into maps and models that are the basis of subsequent
action. Alternatively, individuals may be thought of as having deep, true identities and
they are motivated, ultimately, by contexts that allow them to express their true natures.
The socialthat is, the cooperative and competitive relating between peopleis
important as an enabling context (Stacey et al. 2000, p. 181).
The shift in understanding of causality invoked with the perceived causal

framework expressed in transformative teleology is remarkable. From the primacy
of a choosing individual, we shift to a view that emphasises the context of the
social by which people are motivated. In model language, one may say that the
actions at the micro-level cannot be aggregated as if it were to produce a representation of the social. Instead, the micro-level and the social, as different and
separate entities, interact in a self-organising process of change. In specific
domains we may see concrete outcomes as being sympoietic, i.e. produced/created
in an interplay between the micro-level and ensembles we can refer to, dependent
on context, as either a meso- or a macro-level. Later on we will make use of these
insights by introducing in SP theory the concept of subworld and in SP practice the
decision conference as an operational arena for the unfolding of SP and strategic
decision making.
With a basic view of an organisation as a socio-technical system that undergoes
a transformative change towards its futures (being in the plural until realised) and
such change seen as related to the making of important system/environment
distinctions (at least for some strategic decisions related to situations with complex
strategic choices involved) systematic planning has to be renounced as being at
best insufficient; in some cases it may even be counterproductive. At the end of
Sect. 1.1 a framework for systemic planning was introduced as four interlinked
levels, of which level 1 concerned coming to grips with complexity by combining
complementary ways of seeing as two basic epistemic lenses for SP. This chapter
has served to encircle the condition of complexity. The next chapter on linking
complexity with simplicity will serveby adding to this chapterto finalise the
filling-in of the SP frameworks level 1.
25

Complexity and its meaning can be understood by making use of social systems
theory as developed by Niklas Luhmann. Luhmann is difficult to read for a
newcomer to his systems world. Certain of his insights are, however, important: A plurality of realities is created through cognition; cognition produces
reality by producing system/environment distinctions. These insights are a
backdrop for the stage-wise construction of SP principles and methodology
described in the following chapters. Their function may become visible already
now if reality above is replaced by decision knowledge.
The nature of complexitya central theme in Luhmanns writingsis reflected
upon by him by introducing concepts such as autopoiesis and contingency. For
this context it suffices to note that a theory of action is offered by Luhmanns
systems thinking which is not based on a transparent knowledge of the systems
functioning. According to Luhmann system changes are basically contingent
and aim to reduce complexity.
Complexity is interwoven with open-ended change and according to Ralph
Stacey open-ended change cannot be predicted or modelled by use of causal
relationships. Open-ended change is an expression of self-organisation.
On this theoretical basis companies and organisations can be seen as sociotechnical systems that when they engage in strategic decision making have to
confront the complexity that is representative of the systems world they belong
to. With strategic decision making as an expression of a socio-technical system
engaging in open-ended change, the planners and decision-makers are in need of
what is later on in the book referred to as choice intelligence. How such a kind
of intelligence is obtained is dependent on chosen applied epistemic lenses
(Chap. 3), how knowledge and competencies can increase (Chap. 4) and how
particular insights of importance relating to the concrete planning task can be
achieved through adequate use of different modes of enquiry (Chap. 5).
References
von Bertalanffy L (1972) General system theorya critical review. In: Buckley W (ed) Modern
systems research for the behavioral scientist. Aldine Publishing Company, Chicago
von Bertalanffy L (1973) General system theoryfoundations, development, applications.
Penguin books, London
Checkland P (1981) Systems thinking, systems practice. Wiley, Chichester
Harste G (1992) Niklas Luhmanns konstruktion af samfundsteori. In: Jacobsen C (ed)
Autopoiesis: En introduktion til Niklas Luhmanns verden af systemer. Forlaget politisk
revy, Copenhagen
Kneer G, Nassehi A (1997) Niklas Luhmannintroduktion til teorien om sociale systemer. Hans
Reitzels Forlag, Copenhagen
Leleur S (2008a) Systems science and complexity: some proposals for future development. Syst
Res Behav Sci 25(1):6779
26
Leleur S (2008b) Systemic planning: principles and methodology for planning in a complex
world, 2nd edn. Polyteknisk Forlag Lyngby, Denmark
Leleur S (2008c) At navigere mod fremtiden: Systemisk planlgning som ide og metode.
Polyteknisk Forlag Lyngby, Denmark
Luhmann N (1995) Social systems. Stanford University Press, Palo Alto
Moeller H-G (2006) Luhmann explained: from souls to systems. Open Court, Chicago
Prigogine I, Stengers I (1985) Den nye pagt mellem mennesket og universet: Nye veje i
naturvidenskaberne. Forlaget ASK, Copenhagen
Stacey RD (1993a) The Chaos Frontiercreative strategic control for business. ButterworthHeinemann Limited, Oxford
Stacey RD (1993b) Managing the unknowable. Josey-Bass Publishers, San Francisco
Stacey RD, Griffin D, Shaw P (2000) Complexity and managementfad or radical challenge to
systems thinking. Routledge, London
Thyssen O (1992) Forhold som forholder sig til sig selv. In: Jacobsen C (ed) Autopoiesis:
En introduktion til Niklas Luhmanns verden af systemer. Forlaget politisk revy, Copenhagen
Thyssen O (1997) Vrdiledelseom organisationer og etik. Gyldendal, Copenhagen
Chapter 3
Linking Complexity and Simplicity
3.1 Ways of Seeing

A major concern regarding the treatment of complex strategic choices is that nonintended constraining with regard to the decision space should be avoided.
Therefore perceiving the strategic choice in a proper way becomes of utmost
importance already in the beginning of the process that will lead towards concrete
decision making. How we perceive this choice thus deserves our attention, which
is the field of epistemology concerned with cognition, i.e. in our context: how can
we address and get to know about the complex planning problem at hand?
The challenge arising from complexity has already been made obvious from
previous chapters. The introduction of detail, dynamic and preference complexity
that are types of complexities which are at least of relevance for complex foresight
tells us that change being related to the strategic choice to be made can in no way
be tamed. Going back to Staceys types of strategic change, the type of change
will belong to the category of open-ended change and not to the categories of
closed or contained change. Therefore our decision space at the beginning of the
preparation of a complex strategic choice is in principle unbounded, which
means that we cannot determine a finite set of choices that we can be certain will
exhaust the choice possibilities we ought to consider. To put it more technically we
cannot conduct a complete mapping of the possibilities.
Therefore, when addressing complex strategic choices we need to initially
examine different ways of seeing, which will then set the background for the
process and methodology set out later on for handling complex strategic choices.
On a meta-methodological level, which is the focus in this chapter, we will refer to
such ways of seeing as paradigms to be described and explained below also as
epistemic lenses. Later on in the following chapter these lenses are again the
background for what will be called modes of enquiry (MOEs). These MOEs are the
first practical results to be part of the SP methodology.

27
28
3 Linking Complexity and Simplicity
3.2 Two Epistemic Lenses Concerning Simplicity

and Complexity
To deal with the issue about ways of seeing, we will make use of the viewpoints
formulated by the French sociologist and philosopher of science Edgar Morin
(Morin 1974, 1985; Leleur 2008).
According to Morin, we need to examine the overall research patterns made use
of in scientific explanation. Since the publication of The Structure of Scientific
Revolutions by the physicist and philosopher of science Thomas Kuhn (1962), such
patterns are referred to as paradigms, so we need to address relevant paradigms.
Morin says that classic scientific explanation is based on a Simplicity paradigm,
prescribing that complexity in the world of phenomena should be sorted out by the
establishment of simple principles and general laws. Thus, in this view, complexity
is perceived as the basic mode of appearance and simplicity the underlying true
essence. Not surprisingly, these considerations are exemplified by the Newtonian
physics of gravity and planetary movements. The content of the Simplicity paradigm is outlined as a set of various principles which govern sound scientific
endeavour: science must concern universal matters and reveal invariance. Objects
are separated, but deterministic laws can be discovered which explain their
behaviour. Predictability thus becomes a characterising feature. Moreover, a distance exists between the perceiving subject and the objects being perceived, so that
the objects are not affected or changed in any way during the examination process.
The picture coming out of this is that of an automaton with linear causality. The
language of the Simplicity paradigm is one of objectivity and quantity.
It is well knownas is also the case of the recent advances in systems science
and its use of analogies from non-equilibrium physicsthat physics and cosmological thinking are, and have always been, major suppliers of ideas to other
branches of science. It is therefore quite interesting that Morin sees the insufficiency of the Simplicity paradigm as revealed in the field of subnuclear physics,
where newly discovered ephemeral particles cannot be satisfactorily described.
Against this background, he argues for a Complexity paradigm to be formulated
with the purpose of enriching not only natural science but also social science and
the humanities. He formulates the Complexity paradigm as a set of principles to
complement those in the Simplicity paradigm.
In the Complexity paradigm, focus is set upon local and unique matters instead
of invariant forms of universal validity. Emphasis is given to organisation,
autonomy and possibility, instead of determinism, dependence and necessity.
Relating to physics, the Complexity paradigm recognises asymmetric time irreversibility as an integrated part of natures multiplicity. Other concerns are that
prediction, separation and identity have to be complemented with surprise,
wholeness and individuality. Instead of subject-object relations between perceiver
and an object element, subjectsubject relations need to be given attention. The
picture is no longer that of the automaton, but one of an organism in its broadest
sense, in a context of self-organising multi-causality. The language of the
3.2 Two Epistemic Lenses Concerning Simplicity and Complexity

Table 3.1 The two
paradigms concerning
Simplicity and Complexity.
Adapted from (Morin 1985,
p. 19)
29
Simplicity paradigm
Complexity paradigm
Universality
Determinism
Dependence
Necessity
Lawfulness
Prediction
Separation
Identity
The general
Objects
Elements
Matter
Quantity
Linear causality
The automaton
Objectivity
Multiplicity
Organisation
Autonomy
Possibility
Self-organisation
Surprise
Wholeness
Individuality
The particular
Subject
Interactions
Life
Quality
Multi-causality
Time
Culture
Complexity paradigm is not objectivity and quantity, but cultural interpretation

and quality. Table 3.1 shows the two paradigms as formulated by Morin.
One point being made by Morin of great relevance for our understanding of the
meaning of systemic is that neither the Simplicity nor the Complexity paradigm is
right per se to underpin our deliberations in a concrete decision situation, with this
seen as a choice between two competing approachesor better, meta-approachesfor validating our concepts, procedures and models.
Therefore the paradigms should not be thought of in the way that one should be
adopted and not the other or vice versa; on the contrary, Morin argues that the
paradigms should complement each other. They thus become remedies for each
other: uncertainties invoked by making use of just one of these can be dealt with
by adopting additional strategies for examination based on the other one.
3.2.1 The Complementarity of Simplicity and Complexity

The concepts of uncertainty and complementarity were worked upon by the
physicists Werner Heisenberg and Niels Bohr in the mid-1920s. Fritjof Capra
known for his Tao of Physics from 1975 and also educated as a physicistlater on
turned to systems science to formulate viewpoints on society and ecology in his
book The Turning Point from 1982. Here he gives a really broad sweep of the
societal aspects of putting more emphasis on wholeness and holistic approaches in
medicine, energy and other sectors of society. Of particular interest in this context
is the way he recalls the achievements of Heisenberg and Bohr concerning the
notion of complementarity:
30

It was Heisenbergs great achievement to express the limitations of classical concepts in a
precise mathematical form, which is known as the uncertainty principle. It consists of a set
of mathematical relations that determine the extent to which classical concepts can be
applied to atomic phenomena; these relations stake out the limits of human imagination in
the atomic world. Whenever we use classical termsparticle, wave, position, velocityto
describe atomic phenomena, we find that there are pairs of concepts, or aspects, which are
interrelated and cannot be defined simultaneously in a precise way. The more we
emphasize one aspect in our description, the more the other aspect becomes uncertain,
and the precise relation between the two is given by the uncertainty principle.
For a better understanding of this relation between pairs of classical concepts, Niels
Bohr introduced the notion of complementarity. He considered the particle picture and the
wave picture two complementary descriptions of the same reality, each of them only partly
correct and having a limited range of application. Both pictures are needed to give a full
account of the atomic reality, and both are to be applied within the limitations set by the
uncertainty principle. The notion of complementarity has become an essential part of the
way physicists think about nature, and Bohr has suggested that it might also be a useful
concept outside the field of physics (Capra 1982, p. 68).
3.2.2 Systemic Thinking Further Defined

I have adopted the ideas of the Simplicity and Complexity paradigms and their
complementarity as issues relevant for a first and most basic orientation towards a
given complex strategic decision situation.
In the introductory Chap. 1 the concept systemic was introduced as being more
comprehensive and inclusive in contrast to systematic described to be schematic
and based on a given prefixed procedure. From the paradigm theory of Morin it
becomes possible to attach a more full-fledged meaning to systematic vs. systemic,
namely by seeing systemic thinking generally as rooted in the Complexity paradigm
and systematic thinkingexemplifying in this respect the way that many topics are
taught in the various subject curricula at universities or vocational schoolsas
rooted in the Simplicity paradigm. Consequently in our context, we will not see
systemic thinking as isolated from systematic thinking but as unfolding from an
interplay with more systematic considerations for each particular decision situation.
As a consequence systemic exploration related to strategic decision making can
take on rather broad interpretations. With wholeness as just one of the constituting
concepts of the Complexity paradigmsee Table 3.1we obtain a wide basis for
the explorations to be carried out relating to concepts, methods, processes, etc.
This wide basis we will carry on as two basic epistemic lenses, each of which can
frame our deliberations in a particular way.
3.3 Combining Simplicity and Complexity Thinking

A fundamental characteristic of the systemic approach is that we draw deliberately
on both a simplification orientation and a complexity orientation; this means that
we apply the two epistemic lenses in combination. Earlier on we stated that with
31
regard to complex strategic choices the decision space cannot be defined

exhaustively. We therefore seek to build insight and understanding through an
interaction of the two paradigms. In this context one can ask the crucial question
whether such an effort can be worthwhile.
3.3.1 Towards Interaction of the Whole and the Parts

The answer to this question was begun in the previous chapter with the description
of Luhmanns theory and the herein described complexity reduction as necessary
but in principle impossible. Morin explores the meaning of complexity in the
following way by stating that complexity is not just a surface noise of the real
making it necessary to take precautions about explanations based on
simplification. Necessarily these are insights to bear in mind with regard to
providing decision support for complex strategic choices. In Morins own words it
is stated as follows:
we must question the adequacy of all explanations based on simplification of principles.
Complexity is not a surface noise of the real, but is the very principle of the real
Uncertainty, indeterminism, randomness, and contradictions occur, not as residues to be
eliminated by explanation, but as non-eliminable ingredients of our cognition/perception
of the real. (Morin 1992, p. 130).
Later on he describes more closely how an interaction between the whole and
the parts can take place:
we can make a higher level of understanding based on the constructive circularity of
the explanation of the whole by the parts and the parts by the whole, in which the two
explanations become complementary in the movement associating them without losing
their simultaneous and opposing characteristics (Morin 1992, p. 131).
3.3.2 Reflection-in-Action
Focusing on a practical application of theoretical insights from Luhmann and
Morin, we can draw on the American organisation theorist Donald Schn, who
developed a reflection-in-action inquiry that deals with learning processes at
individual and group level (Schn 1983). In Chap. 4 we will explore the role
learning processes can play within the systemic approach by making use of the
brothers Hubert and Stuart Dreyfuss five-step learning model.
It is a striking feature that the insights contained in the theory elements from
Luhmann, Morin, Schn, and Dreyfus and Dreyfus in many ways resemble each
other. In this respect Schns work seems to constitute a good and practical bridge
between the general theoretical insights of Luhmann and Morin to the concrete
model of learning by Dreyfus and Dreyfus, who like Schn focus on individuals
learning processes.
32
A fundamental idea in Schns work is that problem understanding can

successively be built-up using reassessments and new hypotheses within an
interaction between what he refers to as reframing and back-talk. In his work
reframing means the establishment of an interim framework of understanding,
while back-talk relates to the specific feedback that is obtained using the actual
framework of understanding (applied to either a problem area through practical
investigation or in conversation with a person). Hansen and Kolmos (1998, p. 4)
give the following understanding of Schns reframing/back-talk schema:
Any formulation of a problem is a result of what the observer of the problem can see,
what intentionality is behind the way that the problem is understood. An engineer will
focus on other parts of the same problem than a sociologist. Reflection-in-action involves
reflecting on how your problem understanding affects the problem itself. A good understanding of a problem is a broad problem understanding based on past experience,
whether its your own or that of others. Reflection-in-action is mainly concerned with first
formulating the problem and then through the back-talk of the problem reflecting on the
consequences of the chosen formulation. Are all aspects taken into consideration? Is there
perhaps still a significant difference to the present problem and those it was originally
related to? What values lie behind the chosen assumptions and choice of analysis theory
and method? (Hansen and Kolmos 1998, p. 4 in transl.).
Problem formulations are necessarily biased by methods, objectives, perceptions, etc. As stated above, we do not expect an engineer and a sociologist to
approach a problem in the same way. Rather schematically, we may expect the
engineers approach to be closer to the Simplification paradigm than the approach
taken by the sociologist and vice versa as concerns the Complexity paradigm. With
a systemic approach we have, however, paved the way for a combination of the
two basic epistemic lenses represented by the paradigms of simplicity and
complexity.
As already noted above, learning takes on a particular role when engaging in
complementing conventional thinkingreferred to as being systematic by naturewith a type of thinking referred to as being systemic by nature. These ways
of thinking about a complex strategic issue are our first level of the systemic
framework introduced as four interconnected levels in Sect. 1.1. By means of the
two basic epistemic lenses of simplicity and complexity we have established ways
of seeing, which will link complexity and simplicity represented later on by
learning in association with quantitative modelling. Chapter 4 will address
learning processes.
According to Edgar Morin the exploration of reality makes two basic epistemic
lenses necessary. By referring to the epistemic lenses as paradigms these are the
Simplicity paradigm and the Complexity paradigm.
The Simplicity paradigm and the Complexity paradigm are not meant to replace
each other but should be made use of in a complementary way.
33
Furthermore Morin states that complexity is not a surface noise of the real, but is
the very principle of the real.
A fundamental principle is set out by Morin as constructive circularity, where
the explanation of the whole by the parts and the parts by the whole constitutes
our cognition of the real.
The considerations about the epistemic lenses lead to the findings that systemic
planning should make balanced use of simplicity and complexity thinking and
adopt constructive circularity as stepping stones in formulating the SP
framework.
References
Capra F (1982) The turning pointscience, society and the rising culture. Flamingo by Fontana
Paperbacks, London
Hansen S, Kolmos A (1998) Projektvejlederen som mesteren i en gensidig forstelsesdialog,
Pdagogisk Udviklingscenter, Aalborg Universitet
Kuhn TS (1962) The structure of scientific revolutions. Chicago University Press, Chicago
Leleur S (2008) Systems science and complexity: some proposals for future development. Syst
Res Behavioral Sci 25(1): 6779
Morin E (1974) Complexity. Int Soc Sci J 26(4): 555582
Morin E (1985) Kompleksitetens bud. Paradigma 1(1):1820
Morin E (1992) The concept of system and the paradigm of complexity. In: Maruyama M (ed)
Context and complexity: cultivating contextual understanding, Springer, Heidelberg,
pp 125138
Schn DA (1983) The reflective practitioner. Basic Books, New York
Chapter 4
The Systemic Process
4.1 From Novice to Expert

The perception of learning to be outlined in this chapter is based on a model set out
by Dreyfus and Dreyfus in 1986. Their model assumes that learning takes place in
five stages, which they call: (1) novice, (2) advanced beginner, (3) competent,
(4) proficient and (5) expert (Dreyfus and Dreyfus 1988, pp. 1651).
In the first stage, the novice learns to recognise the objective facts and features
that are relevant for the particular skill to be acquired and learns rules that can
determine actions based on such facts and features. Characteristic of this initial
learning stage is that the rules for determining action can be applied in an
unambiguous and context-free way. The two Dreyfus brothers, one a philosopher
and the other an operations researcher, use the well-known example of learning to
drive a motorcar. At the novice stagewe assume a car not equipped with an
automatic gearshifting the gear is an action the learner needs to become familiar
with and the instructor relates it to speed. In the same way, the minimum distance
to the car in front prescribed by the instructor is also related to speed. When first
introduced, these rules will ignore context by not referring to traffic density, for
example, or anticipated stops.
In the following stage, advanced beginner, the learner has accumulated some
experience by using the rules. What happens now is that by handling issues with
meaningful elements in concrete situations, which neither the instructor nor the
learner can define in terms of objectively recognisable context-free features, the
advanced learner starts to recognise these elements when they are present. This is
done by noticing similarity with prior examples. Dreyfus and Dreyfus call the new
elements situational to distinguish them from context-free elements. Rules for
behaviour now relate to both the new situational and the context-free components.
In the case of the car driver, the learner now uses both situational engine sounds as
well as context-free speed to determine gear-shifting.
At stage 3, competence, what happens is that further accumulation of experience leads to an overwhelming number of context-free and situational elements in

35
36
4 The Systemic Process
a real-world setting. To cope with such problems, people have to learn to adopt a
hierarchical procedure of decision making to organise the situation. By first
addressing a small set of the most important factors, a persons performance can
both be simplified and improved. As Dreyfus and Dreyfus explain, a competent
performer with a goal in mind sees a situation as a set of facts whose individual
importance may depend on the presence of other facts. Given a situation with a
particular constellation of facts, a certain conclusion should be drawn, a decision
made, or an expectation investigated. In the car-driving example, the safe and
smooth operation of the vehicle is related, for example, to the goal in mind of
going quickly from A to B. The route is then chosen with attention to distance and
traffic, which may result in manoeuvring the car in a way that disregards passenger
comfort.
Important changes take place from the novice stage via advanced beginner to
the competence stage according to Dreyfus and Dreyfus:
the novice and advanced beginner recognize learned components and then apply
learned rules and procedures. As a consequence, they feel little responsibility for the
outcome of their acts. Assuming that they have made no mistakes, an unfortunate outcome
is viewed as the result of inadequately specified elements or rules. The competent performer, on the other hand, after wrestling with the question of the choice of a plan, feels
responsible for, and thus emotionally involved in, the product of his choice. While he both
understands and decides in a detached manner, he finds himself intensely involved in what
occurs thereafter. An outcome that is clearly successful is deeply satisfying and leaves a
vivid memory of the plan chosen and of the situation as seen from the perspective of the
plan. Disasters, likewise, are not easily forgotten (Dreyfus and Dreyfus 1988, p. 26).
In many ways, stage 3, where competence is achieved, seems to describeor

more simply matchthe kind of decision making that can be related to the
decision support provided by successful conventional planning. The understanding
and deciding in a detached manner, combined with what we might call best
practice performance clearly dependent on actual involvement, are basic characteristics of good planning performance that a great number of experienced planners
would probably subscribe to. Therefore, for my purpose of outlining a systemic
planning practice with a broader view of planning, it is highly relevant to see how
Dreyfus and Dreyfus relate and assess the competence learning level to problemsolving in general:
When cognitive scientists, psychologists, and others who think about thinking speak of
problem-solving, they have in mind the thought processes that characterize competence. Herbert Simon is typical of such information-processing psychologists, for his
concern is to understand how we choose plans, goals, and strategies, and how situations
represented as sets of facts and figures can be transformed by rule-like procedures into
new sets that conform with our goals. Those psychologists have produced convincing
evidence that we act as problem-solvers when confronted by puzzles or by unfamiliar
situations. However, they typically go on to generalize their results too far, accepting as
essentially true, without supporting this claim by any arguments or empirical evidence,
that all intelligent behavior is of the problem-solving form. They thus uncritically accept
the information processing assumption that intelligence consists in drawing conclusions
using features and rules. (Dreyfus and Dreyfus 1988, pp. 2627).
4.1 From Novice to Expert
37
Table 4.1 Overview of the five stages of skill acquisition

Skill level
Components
Perspective
Decision
(1) Novice
(2) Advanced
Beginner
(3) Competent
Context-free
Context-free and
situational
Context-free and
situational
(4) Proficient
Context-free and
situational
Context-free and
situational
(5) Expert
Commitment
None
None
Analytical
Analytical
Detached
Detached
Chosen
Analytical
Experienced
Analytical
Experienced
Intuitive
Detached understanding
and deciding. Involved
in outcome
Involved understanding.
Detached deciding
Involved
Adapted from Dreyfus and Dreyfus (1988, p. 50)
The following stage 4 in the five-stage model, proficiency, is transitional between

analytic and intuitive thinking. The proficient performer is seen as deeply involved in
the specific task. Because of recent events, certain features of the situation will stand
out as salient and will be scrutinised in new perspectives, while others will recede into
the background and be ignored. No detached choice or deliberation is seen to occur,
and intervention seems to happen based on similarity with previous situations in the
proficient performers experience. The skill applied here by intuitively responding to
patterns without decomposing them into component features is termed holistic
discrimination and association. Intuition is seen as a product of deep situational
involvement and holistic discrimination (Ibid., p. 29).
The final stage 5, expertise, is seen as the stage where deliberation does not
require calculative problem-solving but where critical reflection will address the
experts intuition. Typically, however, the performance of the expert will be
ongoing and nonreflective, concerned with what normally works. In our cardriving example, the expert driver has become at one with the car and experiences
the situation as driving and not driving a car (Ibid., p. 50). Table 4.1 presents an
overview of the five-stage learning model.
4.2 Towards Intelligence Beyond Calculative Rationality

The major concern of Dreyfus and Dreyfus when they published their book, Mind
over Machine, in 1986 was to discuss the potential of the various research programmes in artificial intelligence (AI). Their main point was that these could not
be expected to progress beyond stage 3 about competence in their formulated fivestage learning model whereas human intelligence can achieve this. Human
learning therefore makes possible:
the progression from the analytic behavior of a detached subject, consciously
decomposing his environment into recognizable elements, and following abstract rules, to
involved skilled behavior based on holistic pairing of new situations with associated
responses produced by successful experiences in similar situations. (Dreyfus and Dreyfus
1988, p. 35).
38
With regard to rationality and intelligence, this makes room for arationality.
This is argued in the following way (Ibid., p. 36):
The moral of the five-stage model is: there is more to intelligence than calculative
rationality. Although irrational behaviorthat is, behavior contrary to logic or reason
should generally be avoided, it does not follow that behaving rationally should be
regarded as the ultimate goal. A vast area exists between irrational and rational that
might be called arational. The word rational, deriving from the Latin word ratio, meaning
to reckon or calculate, has come to be equivalent to calculative thought and so carries
with it the connotation of combining component parts to obtain a whole; arational
behavior, then, refers to action without conscious analytic decomposition and recombination. Competent performance is rational; proficiency is transitional; experts act arationally (Dreyfus and Dreyfus 1988, p. 35).
4.3 Arationality and Subworld

The concept of arationality deserves particular attention. If the usual rationality
linked to analytical decomposition expresses the essential features of intelligent
simplification thinking, does similarly arationality then express a kind of synthetic
composing or in more everyday language: a kind of intelligent, holistic behaviour?
And were this the case would it then be possible to benefit from this type of
behaviour when dealing with a complex problem?
These questions can be answered in a confirmative way. In this respect the
concept of subworld will be introduced by another quote from Mind over Machine
written by the Dreyfus brothers in 1986. As mentioned the purpose of their book
was to argue that artificial intelligence (AI) would face difficulties progressing
beyond the intelligence to be associated with the learning level 3 of competence as
contained in the formulated five-stage learning model. In hindsight, AI research
afterwards seems not to have contradicted this statement. Their distinction
between universe and subworld is set out as follows:
physical ideas about the universe can be built up by modeling relatively simple and
isolated systems and then making the model gradually more complex and integrating it
with models of other domains. So much is possible because all the phenomena are presumably the result of the lawlike relations of a set of basic elements
This idea doesnt work in AI. There workers confused two domains, which we shall
distinguish as universe and world. A set of interrelated facts may constitute a universe,
like the physical universe, but it does not constitute a world. The latter, like the world of
business, the world of theater, or the world of the physicist, is an organized body of
objects, purposes, skills, and practices on the basis of which human activities have
meaning or make sense. Thus one can contrast the meaningless physical universe with the
meaningful world of physics. Subworlds, like the world of physics, the business world, and
the theater world, make sense only against a background of common human concerns.
They are local elaborations of the one commonsense world we all share. That is, subworlds are not related like isolable physical systems to larger systems they compose, but
are rather, local elaborations of a whole, which they presuppose. (Dreyfus and Dreyfus
1988, p. 76).
4.3 Arationality and Subworld
39
This contrasting view of world opposite the view of universe indicates a

highly relevant issue to observe when considering how to deal with complex
strategic issues. Specifically it shows the necessity of having to presuppose a kind
of underlying world complexity in principle by the local elaboration which the
dealing with the complex strategic decision task represents; the wording dealing
with is deliberately chosen as solving is not meaningful under such a perspective. The local elaboration of the one commonsense world we all share is a
situation very much different from a situation where we are engaged in a solving
approach (by systematic thinking solely) as such a solving approach is characterised by concretely modelling the representative socio-technical system by
establishing component-like links to the larger environment system. For the same
reason the subworld concept is of great principal importance for the idea of systemic planning as described below. In brief the subworld notion emphasises that
the elaboration of a complex strategic decision task must have a truly holistic
orientation in its unfolding.
4.4 Unfolding the Systemic Process

The SP approach was initially developed by making use of the generic structure
shown in Table 4.2. In a most simple, generic version, planning can be seen as
switching between a mode where the planning environment is scanned for relevant information and a mode where the information is assessed. In the systemic
process these two modes are perceived as being complementary: assessment will
via clarified preferences enable a new, more focused scanning, and scanning will
via new types of information enable a new, more detailed assessment. As the
scanning and assessment modes cannot be problematised at the same time, these
alternating modes are similar to systemic versus systematic perceived as being
complementary. Thereby the generic SP structure in Table 4.2 is obtained as a
combination of systemic versus systematic and scanning versus assessment.
Table 4.2 also indicates some methods and techniques. Presently they are only
to be seen as examples to illustrate the SP structure (Leleur 2004, 2007). In the
chapters to follow they will together with other relevant methods and techniques
be treated with regard to their potential as part of providing decision support for
strategic decision making.
SP is based on applying appropriate operations research (OR) methods and
systems techniques in a kind of self-organising process that embeds conventional
optimisation in a wider process of exploration and learning. The ongoing searchlearn-debate process drawing on the skills, competencies and purposes for the
strategic work moves on by contrasting and interpreting the different findings and
insights. The process aims at converging into a satisfactory end result for the
decision-makers.
Generally hard OR methods can be seen to provide so-called first-order
findings based on calculative rationality, whereas second-order findings (or even
40
Table 4.2 The SP generic structure as four interrelated modes of exploration and learning
SP generic
Systemic
Systematic
structure
Scanning
Assessment
Example: Critical systems

heuristics
Example: Futures workshop
Example: Scenario analysis

Example: Multi-criteria analysis and
simulation
In the figure different methods are indicated to illustrate some possible method choice that can
assist in the systemic process. Adapted from Leleur (2003, p. 262)
higher ones) are associated with systems techniques such as some soft OR
methods. The methods and techniques will be the main topic in Chap. 5.
Wrapping up a number of findings the application of systemic planning (SP) for
complex strategic choices is based on having recognised that the challenge facing
the strategic decision problem is related to open-ended change. This means we
must abstain from relying solely on an analytical approach. Instead the SP
approach consists of embedding the final decision to be taken in a process of
building knowledge about the concrete decision task within a subworld that
unfolds in the course of events. Therefore we need to set a team that can be
expected to handle this. Generally this team will consist of analysts and decisionmakers. Later on in Chap. 6 we will consider the use of decision conferences,
where stakeholders of different kinds are also involved. The subworld notion as
introduced above sets focus upon the importance of avoiding unintended closures
when the systemic process unfolds.
The validity of applying SP on complex strategic choices by using the outlined
systemic process will be returned to in the final Chap. 8. At this later stage the
discussion will be underpinned by its application on a number of presented cases.
Hubert and Stuart Dreyfus provide a convincing theory about learning as five
consecutive stages leading from (1) novice via (2) advanced beginner and
(3) competent towards being (4) proficient and (5) expert.
A finding based on this theory of learning made use of in systemic planning is
the importance of going from detached understanding to involved understanding. In SP the systemic process has this transformation of understanding as a
major goal.
Another finding for systemic planning from the Dreyfus five-stage model is that
rationality (what has been referred to as systematic thinking) is but one type of
reasonable cognitive behaviour. At stage 3 in the model, rationality gives way to
what is termed arationality (which is not to be confused with irrationality not to
be recommended). Arationality denotes a kind of intelligent, holistic behaviour
that the Dreyfus brothers associate with the most developed stages of their
model. They simply find that experts act arationally.
Another finding concerns the difference between the cognitive notions subworld
and universe. The latter assumes a mechanistic view where smaller systems (a
4.4 Unfolding the Systemic Process
41
company for example) are linked into the larger system (a market segment for
example) as a kind of component. The view behind the subworld is opposite to
the mechanistic view as it is seen as a local elaboration of the one commonsense
world we all share. In systemic planning the subworld notion has been adopted
to indicate the collective knowledge about a complex planning problem that
evolves in an SP decision conference. This is exemplified by the description in
Chap. 7 where a subword unfolds around the activities to look for the best
relocation site for a company headquarters. The modes of enquiry to be filled
into the systemic process set out in this chapter are dealt with in the following
Chap. 5.
References
Dreyfus HL, Dreyfus SE (1988) Mind over machine. The Free Press Macmillan, New York
Leleur S (2003) SCOPEan integrated framework for multi-attribute decision making.
Innovation, Carfax publishing 17(3):259270
Leleur S (2004) Systemic decision support in a complex business environment. In: Yucesan E,
Chick S (eds) Proceedings of the 1st future business technology conference (FUBUTEC) at
INSEAD, Fontainebleau published by EUROSIS, March 2004
Leleur S (2007) Systemic planning: dealing with complexity by a wider approach to planning.
Emerg Complex Organ, 98(12):210
Chapter 5
The Systemic Toolbox
5.1 Systems Science as Three Waves

In recent reviews of the development of systems science, a staged development has
been recognised (Midgley 2000). There seems to be a kind of agreement that a first
wave covered by a fifty-year period ending around 1980 can be viewed as an
expression of a functionalist approach, while a second wave, among others
introduced by researchers such as Ackoff and Checkland, unfolded in the 1980s
within an interpretive research orientation. The period from the 1990s until the
present time is characterised by more uncertainty as concerns the wave
categorisation. At least, however, what has been termed emancipatory and postmodern
approaches present themselves as important research orientations to feature in a
current third wave. Furthermore, recent ideas stemming from complexity research
are found to be relevant also in this context.
In this context we will make use of the above findings not for systems research
as such but for the way the research orientations may guide efforts when dealing
with complex strategic choices; therefore we will refer later on not to different
research approaches but to either just approaches or to modes of enquiry (MOEs).
The MOEs become relevant in this book as a way of informing and supporting a
systemic process.
The third wave of systems science is based on the previous functionalist and
interpretive orientations, whichquoted from Systems Approaches to Management by Jackson from 2000can be characterised as follows:
From the functionalist perspective the theories [the functionalist systems thinkers]
produce and test, using the scientific method, clearly relate to some real-world outside of
discourse. Moreover, because of the understanding they gain about the nature of reality,
functionalists believe that they possess expertise that can be employed to ensure efficiency
and efficacy through an enhanced capability to predict and control. As a result they see
themselves contributing both to an increase in knowledge and to the progressive
improvement of the human condition

43
44
5 The Systemic Toolbox

Interpretive systems thinkers wish to promote mutual understanding and learning
through the widest possible participation in decision making. They encourage open debate
and believe that language is a vehicle which can be used to arrive at a consensus, or at
least accommodation, about improvements that can be made to the existing situation.
(Jackson 2000, p. 349).
On the platform of the functionalist and interpretive approaches Jackson and his
collaborators in the 1990s developed critical systems thinking (CST) with
improvement of managerial thinking in mind. They attached CST to Habermass
three cognitive interests: the technical, practical and emancipatory (Habermas
1979; McCarthy 1981). In combination with critical systems heuristics (CSH)
developed by Ulrich (1983), CST constituted a new framework that made it
possible to deal with strengths and weaknesses of the available approaches and the
relations between them, which contributed to emancipatory systems thinking
(Jackson 2000). A great merit is that the work by Habermas hereby was made
accessible and methodologically operational, for example with the total systems
intervention (TSI) methodology from the early 1990s, for professionals involved in
organisational problem solving and decision making. With focus on diversity and
heterogeneity and their meaning and value for managerial decision making,
a postmodern approach has also been formulated which draws on theories by
Lyotard and Foucault (Ibid.; Leleur 2008b).
To support the formulation of the SP framework I have included Tables 5.1 and 5.2
with systems science as three waves on the basis of Jackson (2000, p. 42). The
table was initially set up by Burrell and Morgan in 1979 and later it was extended
by Alvesson and Deetz in 1995 with focus on postmodernism (Ibid., p. 41).
In 2005 I added a fifth column with the purpose of incorporating complexity issues
as part of the third wave (Leleur 2007). On this basiscategorised as three
development waves of systems sciencethe following five modes of enquiry can
be presented in an overview by adequate keywords for a set of features: the
functionalist, the interpretive, the emancipatory, the postmodern and the complexity research orientation. Specifically, the latter is presented in accordance with
the features of a complexity research orientation, which is based on my work with
Systemic Planning (Ibid.). The main sources behind this stem from the previously
described works by Morin and Luhmann. The four other approaches shown in
Tables 5.1 and 5.2 have not been changed in any way as concerns their formulation. In this way Tables 5.1 and 5.2 overview the contents of the first and second
waves of systems science together with ideas which are seen as a tentative profile
of a third wave of systems science that is still unfolding.
In his highly engaging presentation Systems ThinkingCreative Holism for
Managers from 2003 Jackson gives surveys of the functionalist, interpretive,
emancipatory and postmodern research orientationsor paradigms as he prefers to
address them. Over the next pages these surveys are quoted as a background to a
survey of the suggested complexity research orientationor complexity paradigm
to apply this term in the context of paradigm descriptions. Tables 5.1 and 5.2
represent a kind of cognitive billboard, where the patches are the individual
45
Table 5.1 Characteristics of functionalist and interpretive modes of enquiry (MOEs) representing the first and second waves of systems science
Features defining the modes Functionalist mode of enquiry
Interpretive mode of
of enquiry
enquiry
Basic goal
Method
Hope
Demonstrate law-like relations

among objects
Nomothetic science
Problems addressed
Efficiency, effectiveness, survival

and adaptation
Machine, organism, brain, flux
and transformation
Inefficiency, disorder
Narrative style
Time identity
Organisational benefits
Scientific/technical, strategic
Modern
Control, expertise
Mood
Social fear
Optimistic
Disorder
Organisation metaphor
Display unified culture

Hermeneutics,
ethnography
Recovery of integrative
values
Culture, political system
Meaninglessness,
illegitimacy
Romantic, embracing
Premodern
Commitment, quality of
work life
Friendly
Depersonalisation
Adapted from Jackson (2000, p. 42)

Note Table 5.1 and 5.2 should be put together so the features in Table 5.1 can be seen as
continued in Table 5.2
messages that specify a paradigm. Some of the patches have been found especially
interesting for the work on SP and have been highlighted in the quotations below.
For the functionalist paradigm, which represents the first wave in Table 5.1,
Jackson gives the following description:
The functionalist paradigm takes its name from the fact that it wants to ensure that
everything in the system is functioning well so as to promote efficiency, adaptation and
survival. It is optimistic that an understanding can be gained of how systems work by using
scientific methods and techniques to probe the nature of the parts of the system, the
interrelationships between them and the relationship between the system and its environment. The expertise it provides should put managers more in control of their operations and organizations, and enable them to eliminate inefficiency and disorder.
Associated with this paradigm can usually be found the machine, organism, brain, and flux
and transformation metaphors. (Jackson 2003, p. 38 with accentuation added).
For the interpretive paradigm, which in this context is seen as the second wave
of the development schema for systems science, Jackson has the following
description:
The interpretive paradigm takes its name from the fact that it believes social systems,
such as organizations, result from the purposes people have and that these, in turn, stem
from the interpretations they make of the situations in which they find themselves.
Organizations happen, and people act and interact in organizations, as a result of their
interpretations. This paradigm wants to understand the different meanings people bring
to collaborative activity and to discover where these meanings overlap, and so give birth
to shared, purposeful activity. Managers can be guided to seek an appropriate level of
46
Table 5.2 Characteristics of emancipatory, postmodern and complexity research modes of

enquiry (MOEs) representing the current third wave of systems science
Emancipatory mode of
Postmodern mode of
Complexity mode of enquiry
enquiry
enquiry
Unmask domination
Cultural and ideological
critique
Reformation of social order
Psychic prison, instruments
of domination
Domination, consent
Therapeutic, directive
Late modern
Participation, expanded
knowledge
Suspicious
Authority
Reclaim conflict
Explore unknown territory
Deconstruction, genealogy Integrate complexity and simplicity
thinking
Claim a space for lost
Contingent insights that will mean
voices
a difference
Carnival
The panopticon with a restricted
view
Marginalisation, conflict Open-ended, wicked and
suppression
hypercomplex problems
Ironic, ambivalent
Multi-dimensional, eclectic
Postmodern
Hypermodern
Diversity, creativity
Awareness, alertness
Playful
Totalisation,
normalisation
Curious
Constrained reasoning and living
Adapted from Jackson (2000, p. 42)

Note Tables 5.1 and 5.2 are in SP seen as a kind of cognitive billboard, where thepatches are
the individual messages that read vertically produce a specific MOE (or paradigm). When read
horizontally the messages produce a feature-line which allows a cross-paradigmatic comparison
of the MOEs
shared corporate culture in their organizations. They can take decisions, on the basis of
participative involvement, that gain the commitment of key stakeholders. Usually associated with this paradigm are the culture and political system metaphors. (Jackson 2003,
pp. 3839 with accentuation added).
The following two descriptions of the emancipatory and postmodern paradigms

cover essential research orientations for the current third wave:
The emancipatory paradigm takes its name from the fact that it is concerned to
emancipate oppressed individuals and groups in organizations and society. It is suspicious of authority and tries to reveal forms of power and domination that it sees as being
illegitimately employed. It criticizes the status quo and wants to encourage a radical
reformation of, or revolution in, the current social order. It pays attention to all forms of
discrimination, whether resting on class, status, sex, race, disability, sexual orientation,
age, etc. Usually associated with this paradigm are the psychic prison and instruments of
domination metaphors. (Jackson 2000, p. 39 with accentuation added).
The postmodern paradigm takes its name from the fact that it opposes the modernist
rationality that it sees as present in all the other three paradigms. It challenges and
ridicules what it regards as their totalizing attempts to provide comprehensive explanations of how organizations function. From the postmodern perspective organizations
are far too complex to understand using any of the other paradigms. It takes a less serious
view of organizations and emphasizes having fun. It also insists that we can learn much by
bringing conflict to the surface, claiming a space for disregarded opinions and thus
encouraging variety and diversity. The carnival metaphor fits well with this paradigm.
(Jackson 2000, p. 39 with accentuation added).
47
The quotation below presents a survey of the fifth column about the complexity
paradigm and its features. Similar to the above four paradigm descriptions I have
indicated a metaphor for this paradigm:
The complexity paradigm takes its name from the fact that it believes that many insights
of importance (that make a difference) are contingent in nature due to the complexity of
a messy problem space. It seeks to explore unknown territory in the organizational
landscape with the hope of gaining such insights. The method orientation it makes use of
consists of integrating what it terms complexity thinking with simplicity thinking, where
the latter in many ways resembles the functionalist paradigm. Complexity thinking consists of getting a higher-level understanding based on adopting a constructive circularity
of the explanation of the whole by the parts and the parts by the whole, in which the
complementary nature of this process plays an important role. A demonstrated practice of
such a principle used in planning and referred to as systemic planning (SP) sets focus
upon a self-organising exploration and learning cycle, where alternating systemic perceptions gained as insights by the participants are fed by a combined use of soft and hard
operations research (OR) methods. A toolbox of especially relevant methods and techniques has been worked out consisting of seven soft and seven hard OR methods. The
organisation metaphor is the panopticon with a restricted view. The person in the panopticon gets his viewing capability geared by mirrors and the intricate way that these
can be placed to enable still changing perspectives and surprises. This greater viewing
capability, however, is also tied to an unfolding discovery of hidden corners and sometimes vast areas of dark space that come around at the same time. Thus we have to
recognise that a gearing of knowledge-gaining concurrently implies becoming conscious
of new issues that we will also need to address. Benefits of the complexity paradigm are
then not to be associated with obtaining control or having fun but with benefits due to
awareness and alertness. The mood is therefore not necessarily optimistic, friendly,
suspicious or playfulit might, however, be sobut is basically seen as curious. This
orientation of mood is in good agreement with perceiving the social fear feature in
Table 5.2 as relating to constrained reasoning and living. (Leleur 2009, p. 52 with
accentuation added).
The four paradigms functionalist, interpretive, emancipatory and postmodern

are seen by Jackson (Jackson 2000, p. 39) as providing different perspectives that
can encourage creativity when managers deliberate how the organisation should be
perceived and how strategic issues should be dealt with. Adding to this view, it is
my opinion that the complexity paradigm has a role of its own in this respect. By
shedding light on the function that altering systemic perceptions may hold with
regard to creativity enhancement, the complexity paradigm should be part of the
paradigm foundation that is available for managers and other professionals in the
organisations for dealing with strategic planning and decision making (Leleur
2008a).
Each line of features in Table 5.1 continued in Table 5.2 allows, when read
horizontally, a detailed comparison in one particular section or dimension across
the five research approaches and thus also across the three waves. This can provide
a background for reconsidering the development in the perception of what a
system actually is. At this stage, in the midst of the third wave, a system can
tentatively be seen as a kind of representation which can be contrasted to perceptions of a system as reality in the first wave and as construct in the second
48
wave. In the current third wave, systems as phenomena are no longer clearly out
there or in the mind. As such, one could get the impression that the system
concept per se has become more opaque with regard to explanatory power. This,
however, is not the casein fact the opposite view may be expressed with the
statement that previous understanding of the system notion can be seen as contained in the perception of seeing systems as representations of differences in
complexity (Leleur 2008b).
5.2 Formulating Five Modes of Enquiry

The idea of making use of comprehensive knowledge gathering for planning and
problem solving stems back to Churchman, who back in 1968 published Design of
Inquiring Systems (Churchman 1979), which influenced, among others, Linstone
and Mitroff. Their work, published in 1994 entitled The Challenge of the 21st
Century led to applying a multiple perspectives approach built on combining what
they called the Technical, Organisational and Personal perspectives (Linstone and
Mitroff 1994). Recent work in this area has been carried out by Jackson in his
books Systems Approaches to Management from 2000 and Systems Thinking from
2003, which were both made use of above (Jackson 2000, 2003).
In this context focus will be on formulating knowledge gathering as modes of
enquiry, where a set of features (see Tables 5.1 and 5.2) enable the user to raise
still new questions instead of just following a fixed type of perspective or procedure. We may then say that the five modes below can be seen as cognitively
open, which is exactly the purpose they should fulfil in the systemic process; they
are meant to inspire the team undertaking the task of setting up decision support
for a complex strategic decision situation, which is treated in Chap. 6. The
dimensions of inspiration for deliberate consideration and discussion are indicated
in Table 5.3.
Of these modes of enquiry the complexity mode is as stated above the most
recent and relates to adopting a focus on uncertainty and complexity and their
possible influences on process and methodology for strategic decision making.
Seen together the five modes of enquirybacked by the cognitive billboard
provide a vast platform on which the subworld-unfolding described in Chap. 4 can
take place.
What basically creates the SP subworldseen as a key notion both in the
learning model and in a systemic planning process (comprising both systematic
and systemic activities)is a successive recasting of systemic perceptions constituted by A: the various insights referred to as findings of first, second and maybe
even higher order, and B: the way these insights are confronted, interpreted and
combined. A new systemic perception of the problem is achieved, when a
difference to the previous one is established, i.e. this is being realised/recognised
by the planning team, maybe in combination with presentation and consultation
with various stakeholders. The shifts in systemic perceptions where a new holistic
5.2 Formulating Five Modes of Enquiry

Table 5.3 The five modes of
enquiry
49
Mode of enquiry
Aims at inspiring and supporting

deliberations about
Functionalist
Interpretive
Emancipatory
Postmodern
Complexity
Goal seeking and viability

Exploring purposes
Ensuring fairness
Promoting diversity
Recasting systemic perceptions
impression is actualised may be both smooth and more disruptive. The panopticon with a restricted view in Table 5.2 may illustrate the occurrence of the
latter: sometimes even a very small change in the angle and shape of a mirror,
symbolising maybe an actual change of methodology, may cause considerably
changed reflections and images due to intricate linkages. The changeto remain
within the metaphoris due to the actual architecture of the allegoric, subworldencompassing mega hall of mirrors and relates to the actual planning conditions.
The latter can be perceived as the problem-specific impacts stemming from at least
detail, dynamic and preference complexity. The images are what we refer to as the
systemic perceptions.
5.2.1 The Two Promises of Applying Systems Science

The application of systems science for the improvement of problem solving and
planning holds two promises:
By seeing our problem or study object as a system, we may be able to make use
of the systems concepts to make a better representation of it and here capture
(and model) various interrelations among elements, etc. in a more qualified way.
By seeing our problem as a system, we may be able to focus less on step-by-step
approaches and capture more holistic impressions which can qualify our study.
The first view expresses what is sometimes referred to as systems analysis: we
proceed by defining our problem and determining the objectives, then turn to
envisaging the consequences of various alternatives (often helped by models of
various types), after which we appraise and select the best alternative. This is
finally implemented and afterwards we may decide to monitor the implemented
alternative (Leleur 2000, p. 18). This almost generic process is well known in both
engineering design and planning. Here it suffices to note that the application of
systems science as systems analysis is very much tied to the ideal of rational
decision making, where complete information is available and is processed analytically to lead to an optimal result (design, plan, etc.). Earlier on we referred to
this as a systematic approach.
The second view simply states, as a corrective to the first one, that wholeness
matters. While the systematic approach above was tied to a step-by-step approach,
50
we have earlier defined a systemic approach as an approach that in contrast to the

systematic approach is concerned with holistic views.
In Sect. 4.4 the generic structure of SP was stated as the dual pairs of
systematic vs. systemic and scanning vs. assessment, respectively, see Table 4.2.
The task now is to encircle methodology that can empower the outlined five
modes of enquiry (MOEs). The purpose of these MOEs is to maintain openness
but at the same time to inform the planner/planning team about the complex
strategic task in a way that assists in the further identification of it. The latter is
required for dealing with assessment of consequences and risks relating to the
strategic alternativesthe choicesthat have come forward as a relevant set of
options for the final decision. In Chap. 6 we will treat activities concerning the
strategic task by formulating scoping of alternatives together with an assessment
of their consequences and risks and thereby present the way that SP can be
applied for making complex strategic choices. In the remaining part of this
chapter the focus will be on establishing a toolbox that is suitable for sustaining
the five MOEs.
5.3 Selecting Methods and Techniques for the Toolbox

To inspire the work concerning selecting the right methods and techniques for the
systemic toolbox a long list of candidate methods has been produced based on a
literature review, see (Troncale 1985; Mingers and Gill 1997; Midgley 2000;
Khisty and Mohammadi 2001; Jackson 2000, 2003; Rosenhead and Mingers 2001)
and (in Danish) (Vidal and Srensen 2001). This list is shown in Table 5.4 with the
methods and techniques indicated in alphabetical order.
In the long list in Table 5.4 the methods have simply been classified into hard
and soft, with the soft ones indicated by a.
Based on Table 5.4 a detailed assessment of the different methodologies can be
made. In this respect we need to ensure that the selected methods and techniques in
addition to systematic and systemic qualities also generally enable and empower
the use of the five MOEs.
In short the five MOEs concern:
Goal seeking and viability.

Exploring purposes.
Ensuring fairness.
Promoting diversity.
Recasting systemic perception.
Against this background, a selected set of an appropriate number of methods

and techniques have been chosen (Leleur 2008b). The total of 14 methods and
techniques consisting of seven hard and seven soft methods and techniques are
shown in Table 5.5.
51
Table 5.4 Selected methods and techniques

Hard and soft methods and techniques
Analytic hierarchy process (AHP)
Computer-aided design (CAD)
Conflict analysisa
Cost-benefit analysis (CBA) and cost-effectiveness analysis (CEA)
Critical systems heuristics (CSH)a
Critical path method (CPM)
Cross-impact analysis
Decision analysis (DA)
Delphi conferencing techniquesa
Environmental impact assessment (EIA)
Expert systems
Forecasting
Futures workshop (FW)a
Fuzzy set theory
Game theory
Graph theory
Inputoutput analysis
Interactive planning (IP)
Intuitive exploration/brainstorming/metaphor and analogy buildinga
Linear programing techniques
Multi-criteria analysis (MCA)
Multiple perspectives (MP)a
Network theory
Optimisation theory and heuristics
Statistics, probability and queuing theory
Scenario building
Sensitivity analysis
Simulation
Soft systems methodology (SSM)a
Strengths, weaknesses, opportunities and threats analysis (SWOT)a
Systems dynamics (SD)
Total systems intervention (TSI)a
a
Soft methods
It is characteristic of the two columns in Table 5.5 indicating the hard and the
soft methods, respectively, that there is a general movement towards more
demanding methods and techniques when reading down through the columns.
One fundamental question is how many hard and soft methods, respectively,
must be involved to ensure a smooth running of a systemic process. There is no
clear-cut answer here.
In the simplest cases, a combination of cost-benefit analysis (CBA) and brainstorming (BS) may lead to a conviction among the decision makers involved that
the factors found by brainstorming and afterwards used to inform a more or less
standardised CBA have provided a sufficient basis for their decision making.
52
Table 5.5 Seven hard and seven soft methods and techniques
The systemic toolbox sustaining the five MOEs
Hard methods
Soft methods
Cost-benefit analysis (CBA)

Simple multi-attribute ranking technique
(SMART)
Scenario analysis (SA)
Preference analysis (PA)
Risk analysis based on Monte
Carlo simulation (RA)
Composite methodology for assessment
(COSIMA, SIMDEC)
Brainstorming (BS)
Mind mapping (MM)
Strengths, weaknesses, opportunities
and threats (SWOT)
Critical systems heuristics (CSH)
Soft systems methodology (SSM)
Stakeholder analysis (STA)
Futures workshop (FW)
Adapted from Leleur (2008b, p. 132)
In less simple casesmethodically speakingit will be necessary to draw on

the methods and techniques that highlight some particular aspects of the actual
strategic decision problem. This could concern the best set of decision criteria or
the relevance of monetary versus non-monetary consequences. In such cases it
may be appropriate to draw on, for example, AHP or SMART. In other cases there
may be special circumstances relating to uncertainty and risk which can make it
relevant to draw on scenario analysis and/or risk analysis. In the case of conflicting
goals and fairness issues, it may be reasonable to focus on preference analysis and/
or stakeholder analysis. In some cases it may be difficult to structure the problem
contained in the actual planning task. Here it may be a good idea to use methods
such as critical systems heuristics (CSH) or soft systems methodology (SSM),
which have a general exploratory nature: What do we know? What do we want?
And why?
Of course there is nothing wrong with drawing on methods from the more
inclusive list in Table 5.4 or for that matter on methodologies not mentioned here.
The rationale is that when an issue of systematic nature is becoming increasingly
complex, it may be advantageous to make use of special methods and techniques.
In the systemic toolbox COSIMA and SIMDEC refer to such methods. Both will
be treated in Chap. 6 and in a more technical way in Appendices A and B
respectively.
Endowed now with the outline of the systemic process from Chap. 4 and the
systemic toolbox from this chapter, the next chapter concerns setting up the
decision support.
Based on systems science and three recognised waves of development of the
theory and methodology within these fields, altogether five modes of enquiry
(MOEs) have been identified: Functionalist, interpretive, emancipatory, postmodern and complexity.
53
The five modes have been scrutinised with regard to their potential to inform
and assist the group learning taking place in the SP decision conference. This
has paved the way for two SP components: one concerns the cognitive billboard
and the other the systemic toolbox, see below.
By posing the five paradigms behind the five MOEs together, a kind of billboard
appears with a multitude of patches. Each patch conveys a message/viewpoint
which may be a discussion trigger in an ongoing group process. In this way the
billboard can contribute to a wide range of deliberations, which of course should
relate to the problem dealt with.
The systemic toolbox consists of seven hard and seven soft operations research
(OR) methods. The methods have been selected from a long list of OR methods
with the purpose in mind that they in various ways should be able to empower
the five MOEs. For the participants in a decision conference this means that a
suitable range of methods and techniques are available which can assist the
group in its deliberations.
References
Churchman CW (1979) Design of enquiring systems. Basic Books, New York
Habermas J (1979) Communication and the evolution of society. Heinemann
Habermas J (1986) The theory of communicative action, volume one: reason and the
rationalization of society. Polity Press, Cambridge
Habermas J (1989) The theory of communicative action, volume two: the critique of functionalist
reason. Polity Press, Cambridge
Jackson MC (2000) Systems approaches to management. Kluwer Academic, New York
Jackson MC (2003) Systems thinking: creative holism for managers. Wiley, New York
Khisty CJ, Mohammadi J (2001) Fundamentals of systems engineering with economics
probability and statistics. Prentice-Hall, New Jersey
Leleur S (2000) Road infrastructure planninga decision-oriented approach, 2nd edn.
Leleur S (2007) Systemic planning: dealing with complexity by a wider approach to planning,
Emergence: Complex. Org. 98(12):210
Leleur S (2008a) Systems science and complexity: some proposals for future development. Syst
Leleur S (2008b) Systemic planning: principles and methodology for planning in a complex
world, 2nd edn. Polyteknisk Forlag Lyngby, Denmark
Leleur S (2009) The meaning of system: from CAS to CHESS. The systemist 33(23):4762
(United Kingdom Systems Society)
Linstone HA, Mitroff II (1994) The challenge of 21st century. State University of New York
Press, New York
McCarthy T (1981) The critical theory of Jrgen Habermas. The MIT Press, Cambridge
Midgley G (2000) Systemic intervention: philosophy methodology and practice. Kluwer
Academic, New York
Mingers J, Gill A (1997) Multimethodology. Wiley, New York
Rosenhead J, Mingers J (2001) Rational analysis for a problematic world revisited. Wiley,
New York
54
Troncale LR (1985) The future of general systems research: obstacles potentials case studies
systems research 2(1):4384, Pergamon, London
Ulrich W (1983) Critical heuristics of social planning. Wiley, New York
Vidal V, Srensen L (2001) Strategi og planlgning som lreprocesseks blde metoder, 2nd
edn. Handelshjskolens Forlag, Copenhagen
Chapter 6
Setting up the Decision Support
When facing a complex planning problem requiring strategic decision making the
concerns initially relate to gaining insight into the nature of the problem and to
sweeping-in all aspects of relevance. This initial phase is of utmost importance as
it will influence all activities along the way towards making the best strategic
choice later on in the process by assessing the decision alternatives that have been
identified. Ideally we can consider the following main steps:
Scanning
Scoping
Assessing
Choosing
Scanning may be perceived as a kind of broad, principally non-bounded search.

In this respect Tables 5.1 and 5.2 can inspire us generally as a kind of multifaceted
cognitive billboard and more specifically by the way it frames and makes
available different, explorative modes of enquiry. In practicerecognising the
previously stated circularity between scanning and assessmentscanning with an
intent, seen even loosely as just getting informed about the actual strategic decision task, expresses the commencement of the scoping that aims at finally producing a range of decision choice alternatives. These can be seen as the scoping
condensation of the horizon of possibilities.
Setting up decision support will require that we, in principle, avoid any constraining of the way we deal with the problem. In practice on the basis of the
principles on which the systemic planning (SP) approach has been set out this means
that all the five modes of enquiry (MOEs) should be adopted. Hereby we seek to
include in principle a sufficient range of matters that we judge as important.
The validity of proceeding like this will be discussed more closely in the final
Chap. 8. From a practical viewpoint the MOEs in the SP framework are seen to
represent a rather wide range of possible ways for exploring a complex planning
problem.

55
56
6 Setting up the Decision Support
Table 6.1 The five modes of SP enquiry

Systemic planning (SP) mode of
enquiry (MOE)
Mainly involves the following methodology
Core performance
Wider performance
Fairness
Diversity
Robustness
Hard methodology
Hard and soft methodology
Soft methodology
Below the five MOEs are reiterated with some consideration of their main use
when placed in a context of planning and strategic decision making:
MOE
MOE
MOE
MOE
MOE
1:
2:
3:
4:
5:
FUNCTIONAL aiming at improving goal seeking and viability

INTERPRETIVE aiming at exploring purposes
EMANCIPATORY aiming at ensuring fairness
POSTMODERN aiming at promoting diversity
COMPLEXITY aiming at recasting systemic perceptions
Based on Tables 5.1 and 5.2 each MOE is meant to inspire and guide exploration along one particular axis of insights; evidently the cognitive search in this
respect can in no way be finished. Therefore we have to speak of modes and not of
activities. As each mode in the tables is set out in a general way and not particularly for planning relating to complex strategic choices, we have to come to terms
with how the MOEs can become most cognitively effective when applied in a
context of planning and strategic decision making.
As concerns MOE 1 and MOE 2 goal seeking and viability together with
exploring purposes may become precise if seen as relating to core performance
and wider performance respectively. These formulations set focus on efficiency
and effectiveness as will be described below. MOE 3 and MOE 4 about fairness
and diversity respectively, can enter directly into methodology, while MOE 5 is
seen as having its focus on robustness, also to be treated further below.
With these modifications the five SP MOEs in Table 6.1 have been outlined,
which will guide the practical aspects of applying SP for complex strategic
choices.
The following sections will specify the methodology. First scoping will be
treated in Sect. 6.1 and afterwards assessing in Sect. 6.2. Both sections serve to lay
a basis for the treatment of the topic of choice intelligence addressed in Sect. 6.3.
Choice intelligence is coined in this presentation as a collective term for judging
more closely the overall effort and capability relating to decision support.
6.1 Scoping the Strategic Choices

The main purpose of scoping is the sweeping-in of all relevant information for the
determination of a preliminary set of strategic choice alternatives. The choice
alternatives are described as preliminary as each of these can be modified later in
57
the process if such alteration is desirable for some reason. What matters, however,
in this early stage is that we do not exclude an option that could later onin case it
had been includedturn out to be a serious competitor among the alternatives to
come forward as the best choice. In a way then scoping is concerned with a kind of
boundary setting between what may loosely be described as included in the
decision space vs. excluded from the decision space. In a more mathematical
language we would describe this as omitting the fallacy of engaging in
suboptimisation.
Experience shows that sometimes scoping is paid too little attention as the
alternatives to be worked on more or less seem to present themselves. This may, of
course, be true in some cases. In complex, strategic decision making, however,
scoping becomes important as the boundary setting it represents is generally a
main element of the challenge represented by a complex plan problem.
On this basis all relevant modes of enquiry ought to come into play as they
explore our cognitive capability for what may facilitate the transformation of a
complex strategic decision problem into a set of choices that represent some initial
constraining of the decision space. Therefore this constraining should generally be
influenced by concerns of core and wider performance, fairness, diversity and
robustness, see Table 6.1.
Typically the strategic decision challenge arises on a background triggered by
either an opportunity or a threat or a combination of both. As it can be assumed
that the team addressing the strategic decision will have some preknowledge about
the possible core performance of alternativessome screening may be available
with rough estimatesthe wider performance and the fairness, diversity and
robustness issues come into the fore. Scoping therefore makes it reasonable to dig
deeper into the MOEs exploring such issues, and as already indicated in Table 6.1
this is the field of the soft methods in the toolbox.
6.1.1 The Soft Methods in the Toolbox

The following soft methods are included in the toolbox:
Brainstorming (BS)
Mind mapping (MM)
Strengths, weaknesses, opportunities and threats (SWOT)
Critical systems heuristics (CSH)
Soft systems methodology (SSM)
Stakeholder analysis (STA)
Futures workshop (FW)
The first three methods are well known and practised in a number of more or
less formal versions (Leleur 2008). Brainstorming may range from free-and-open
discussion to a version based on rules, where a facilitator conducts the session that
will typically contain a sequence of questions; mind mapping is also relatively
58
well known, where ideas and especially how they interrelate are brought forward
successively as the process goes on and ends up presenting the team involved with
what is sometimes called rich pictures. Especially wider concerns and diversity
issues can be shed light on using brainstorming and mind mapping in combination.
A more structured way of thinking about the complex, strategic problem can be
obtained by a SWOT analysis, where internal and external factors are approached
by imagination and consideration of respective strengths versus weaknesses and
opportunities versus threats leading to a SWOT-matrix that can facilitate the
further process of scoping the alternatives.
What constitutes the differences between the methods is their balance between
being unstructured and thereby allowing open discussions and being more structured and thereby securing a relevant result (in some respect) to come out of the
efforts; needless to say that no right balance in this respect can be prescribed.
A concern always present in processes preparing decision making is the amount
of time and resources being consumed. If the team includesin addition to analystsa number of senior people from either the middle or maybe even the toplevel of the organisation this concern becomes even more outspoken. Typically
this will inhibit the use of methods more demanding in time than brainstorming,
mind mapping and SWOT. However, there may even on this background be good
reasons to apply a more demanding soft method that has shown a capability to dig
out knowledge about the decision problem in hand and have a critical influence on
the outcome of the process. In the systemic toolbox the available methods are
critical systems heuristics (CSH), soft systems methodology (SSM), stakeholder
analysis (STA) and futures workshop (FW). Of these four methods the first two
will be described below with regard to their potential. The latter two methods are
less demanding in prescribed content but not in time, and they are treated
explicitly as part of the demo-case in Chap. 7. As this is not the case with regard to
CSH and FW, the following two subsections will concern CSH and SSM, both of
which have a strong record of practical applicability, for which reason they are part
of the SP toolbox.
6.1.2 Critical Systems Heuristics

Critical systems heuristics (CSH) was developed by Werner Ulrich in the 1980s
and has had a profound influence on management thinking based on systems
theory. It presented around 12 questions which enable the team to reflect upon
circumstances of critical importance, see Table 6.2 (Ulrich 1983, pp. 240264;
Jackson 2000, p. 318).
It should be noted that the 12 questions in Table 6.2 are given in is mode
(Who is the actual client , etc.) and that Ulrich in parallel to this uses the same
questions in an ought mode (Who ought , etc.). By using CSH, the answers to
the questions lead towards a first mapping of critical issues of relevance for the
formulation of decision choice alternatives.
59
Table 6.2 Critical systems heuristics (CSH) as methodology based on 12 questions

Critical systems heuristics as 12 questions
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Who is the actual client of the systems design?

What is the actual purpose of the systems design?
What is the built-in measure of success?
Who is actually the decision-maker?
What conditions of successful planning and implementation of the system are really
controlled by the decision-maker?
What conditions are not controlled by the decision-maker (i.e. are in the environment)?
Who is actually involved as planner?
Who is involved as expert, and of what kind is the expertise?
Where do the involved seek the guarantee that their planning will be successful?
Who among the involved witnesses represents the concerns of the affected? Who is or may
be affected without being involved?
Are the affected given an opportunity to emancipate themselves from the experts and to take
their fate into their own hands?
What world view is actually underlying the design of the system? Is it the view of (some of)
the involved or of (some of) the affected?
Adapted from (Jackson 2000, p. 318)
The intention behind the CSH tool is that all 24 questions consisting of the 12
is-questions and the 12 ought-questions should be carefully dealt with.
However, in practiceevidently depending on the actual problemit is the
experience of applying CSH as part of the SP framework that it is perceived as
cumbersome (at least by some people participating in the team) to pay the method
full respect in this way. Therefore an alternative way of making use of CSH is to
concentrate on the questions that seem to be most relevant and productive with
regard to obtaining new insights concerning the particular problem dealt with.
However, CSH used in this selective way can also be really worthwhile.
6.1.3 Soft Systems Methodology

Soft systems methodology (SSM) was developed by Peter Checkland in the 1980s
in the same years as CSH originated. While CSH is basically a critically reflecting
type of questionnaire, SSM is set up as a learning cycle (Checkland 1981; 1985),
with a prescribed content of process to be carried out.
The structure of SSM is made up of seven interrelated activities, which proceed
from finding out to taking action. The activities are: (1) Problem situation:
unstructured, (2) Problem situation: expressed, (3) Root Definitions of relevant
human activity system, (4) Conceptual Models of the system concepts named in
the Root Definitions, (5) Comparison, (6) Changes: desirable and feasible and
(7) Action. As a principal trait of the methodology, it should be noted that
60
activities nos. 1, 2, 5, 6 and 7 are embedded in what Checkland terms the real
world, while nos. 3 and 4 are seen as systems thinking about the real world. The
structure described is shown in the upper part of Fig. 6.1.
The principle of SSM can be understood by surveying the stages of the
methodology. Stages 1 and 2 try to build as rich a picture of the problem
situation as possible. This implies collecting several perceptions. Specifically, it
has been found relevant to investigate both a slow-to-change structure and a
continuously changing process. By relating structure to process, essential
characteristics of the situation may be revealed. The function of the initial
stages is to obtain an expression which can serve as a background for relevant
choices.
The subsequent stage 3 concerns what Checkland calls Root Definitions. Their
purpose is to define one or more relevant systems in a way that makes it possible to
discuss their nature more openly. Such definitions constitute a survey of the
problem situation and provide the base from which such a survey and its implications can be further developed. It should be emphasised that we are not dealing
with real-world problems in stages 3 and 4, but with intellectual constructs or ideal
types. In this way, each Root Definition (RD) or set of Root Definitions is an
abstract ideal type of a purposive system that Checkland calls a Human Activity
System.
In stage 4, what Checkland calls Conceptual Models is formulated. This can be
done by simply bringing different verbs together, but more formal systems rules
formulated as part of the methodology can also be applied. Conceptual Models can
be seen as structured sets of activities combined logically in accordance with their
underlying Root Definitions. Thus, these models are not in the real world either,
but are a refinement or further building up of the ideal thinking.
The aim of stage 6 is to make use of the comparison results obtained in stage 5
to discuss possible, relevant changes. These should be both desirable on the basis
of the insights from Root Definitions and Conceptual models and they should also
be culturally feasible in the actual context.
In stage 7, action should be taken on the basis of the outcome of stage 6,
whereby the learning cycle is closed and a new situation obtained.
The outlined SSM stages can be based on further explanation. Among other
things a mnemonic CATWOE has been devised that can function as a kind
of checklist for further considerations. Each letter indicates a type of question
or consideration to be reflected upon, for example W for Weltanschauung,
where the German word for world view has been chosen to indicate a certain view behind a specific Root Definition and its associated Conceptual
Model. The other letters in the mnemonic are explained in the lower part of
Fig. 6.1.
Next, in stage 5, a comparison is carried out between the findings from stage 2
and the suggestions derived from the Conceptual Models in stage 4. Now we are
dealing with real-world problems once more by asking what features of the
Conceptual Models are especially different from present reality and why.

Finding out
Problem
situation:
Unstructured
61
Taking action
Action
Changes:
Desirable and
feasible
Problem
situation:
Expressed
Comparison
Real world
Systems thinking
about the real world
Conceptual Models of
the system concepts
named in the Root
Definitions
Root Definitions of
relevant Human
Acticity Systems
C onsi der ati on
Amplification
(1)
Ownership (O)
Ownership of the system, control, concern or sponsorship; a wider system which may discourse about
the system
(2)
Actors (A)
The agents who carry out, or cause to be carried out,

the transformation process(es) or activities of the
system
(3)
Transformation (T)
The core of the RD; a transformation process carried

out by the system; assumed to include the direct
object of the main activity verb(s)
(4)
Customer (C)
Client (of the activity), beneficiary, or victim, the

sub-system affected by the main activity(ies); the
indirect object of the main activity verb(s)
(5)
Environmental and wider

system constraints (E)
Environmental impositions; perhaps interactions

with wider systems other than that included in (1)
above, these wider systems being taken as given
To this list is added the sixth item which is, by nature, seldom if ever explicit in a root definition but is always implicit and always relevant:
(6)
Weltanschauung (W)
The (often unquestioned) outlook or taken-forgranted framework which makes this particular RD a
meaningful one
Fig. 6.1 Soft systems methodology (SSM): Process and a mnemonic checklist (CATWOE)
for particular considerations. Adapted from (Checkland 1985, p. 19)
62
We can sum up the meaning and potential of SSM the following way (Leleur
2000, p. 201):
It is important not to restrict the planning unintentionally, that is not to impose
any kind of unwanted closure
Basic choices with regard to concepts (Root Definitions) and their use in
modelling (Conceptual Models) fundamentally bias the whole planning process,
and
Radical considerations may be made possible if alternative ideal constructs are
cultivated and confronted with an expressed real-world problem situation.
6.2 Assessing Consequences and Risks

Assessing consequences and risks become important when scoping has produced a
set of strategic choice alternatives. With the core performance of the alternatives
addressed at best in doing a screening based on rough estimates in the scoping, it
now becomes essential to take a closer look at each alternative by asking fundamentally whether it remains attractive with a view on both its core performance
and its wider performance. Scrutinising each alternative in turn will also lay the
basis for a subsequent exploration of their relative attractiveness.
While scoping was conducted principally on the basis of the soft methods in the
toolbox, examination of the attractiveness, at least at the beginning of the
assessment process, will be dominated by analytic methods referred to as
belonging to the category of hard methods.
6.2.1 The Hard Methods in the Toolbox

The following hard methods are included in the toolbox:
Cost-benefit analysis (CBA)

Simple multi-attribute ranking technique (SMART)
Scenario analysis (SA)
Preference analysis (PA)
Risk analysis based on Monte Carlo simulation (RA)
Composite methodology for assessment (COSIMA, SIMDEC)
The examination of core and wider performance will primarily draw on the
functional and the interpretive MOE respectively.
Generally the core performance will comprise economic issues such as gains
termed benefits and expenditures termed costs. The wider performance of a choice
alternative will concern other issues that cannot be treated in money terms.
63
Therefore in SP, core and wider performance are approached as economic and
non-economic assessments respectively.
Not surprisingly economic assessment in organisational decision making is a
prominent issue and rightly so as organisational development (or in a more pessimistic mode: organisational survival) depends on a combined outcome of decisions made (strategic, tactical and operational) that at the end of the dayor better
in an accounting perspective: end of the yearshould consist of black figures on
the bottomline instead of red ones. What complicates this matter when focusing on
strategic decision making is that often it is not possible to obtain an economic
assessment that covers the attractiveness of the decision choice alternatives in a
satisfactory way. With complex strategic choices we typically face the challenge
that matters of importance are non-economic.
In the SP framework this is treated by assessing the core performance by use of
cost-benefit analysis and the wider performance by use of multi-criteria analysis.
Both methods belong to the hard methods in the systemic toolbox. To see them as
analytic (requiring a functional MOE) and semianalytic (drawing also on the
interpretive MOE) respectively, will follow from the way they are described in the
subsections below. As concerns the other hard methods, scenario analysis and
preference analysis will be part of the demo-case description in Chap. 7, whereas a
description of risk analysis as applied in SP will follow after the subsections below
about cost-benefit analysis and multi-criteria analysis.
6.2.2 Cost-Benefit Analysis

With the strategic choice alternatives laid down, the assessing of the consequences
will depend on the nature of the consequences identified. In our monetised world
the first issues addressed are often those where consequences are identified by
being described in money terms such as expenses and gains, or in the language of
economics as costs and benefits. This type of assessment is what we know as costbenefit analysis (CBA). No doubt CBA is the most commonly known methodology; in the sphere of private firms it is also known as financial analysis (FA). The
way of thinking is the same in CBA and FA. However, it should be noted that the
CBA comprises primarily societal consequences, whereas FA may delimit its
focus solely on the firm. There need not, however, be an easy way of delimiting
such a focus as, for example, the BP Explorer accident in the Mexican Golf in
April 2010 and the subsequent mitigation efforts demonstrate.
The approach of CBA is quite simple: What does it cost? What comes out of it?
The logic of the analysis consists of selecting the decision choice alternative that
gives the most for the least. This should accordingly result in a favourable
situation with a surplus by implementing the identified best decision alternative.
Considering a set of alternatives, the choice is thus determined by implementing the alternative that causes the greatest surplus. In a strict sense, this can be
modified to choose the alternative which will lead to the highest return per
64
invested monetary unit. This modification becomes especially relevant when it is

not possible to provide the investment required by the typically expensive alternative that will produce the highest net surplus. Economic theory has developed a
series of investment criteria known as net present value (NPV), benefit-cost rate
(BCR) and others that according to economic theory provide guidance on what
should be done under given conditions (Leleur 2000).
As mentioned the consequences covered by CBA are referred to as the economic consequences. This has made it relevant to address all those highly diverse
consequences that cannot be covered by a CBA as the non-economic consequences. In general it can be stated that complex strategic choices will comprise
not just economic consequences but also non-economic consequences. Actually
the latter will tend to dominate in many strategic decision situations making their
inclusion in deliberate strategic decision making simply necessary. In somewhat
self-contradictory terms one can say that non-economic consequences count (!) in
strategic decision making. The non-economic consequences are treated by using
multi-criteria analysis.
6.2.3 Multi-Criteria Analysis

Multi-criteria analysis (MCA) differs from CBA as direct pricing of the different
elements as in a CBA is not possible. Instead weights are used to replace the
missing monetary unit prices. Whereas CBA was established within economics as
a decision making approach in many countries and problem contexts back in the
1960s, a similar development has not taken place as concerns MCA. It should be
noted also that MCA compared to CBA has a different disciplinary origin as MCA
was developed within operations research as one of several specialised fields for
utilising mathematical methods and models. Thus basically MCA is concerned
with the relative importance of different criteria as opposed to CBA where unit
prices reflect some sort of objectivity. Two quotations can shed light on the differences between CBA and MCA. The first one sets focus on the importance of
user input in MCA:
Multi-criteria analysis is a fairly recent method for assessing and selecting projects
exerting complex socio-economic effects. In this method, the individual assessment elements are taken separately and measured in the appropriate dimensions. the criteria
will have to be weighted among each other because they are not of equal relevance.
Determining the weights requires much responsibility and expertise from the decisionmaker as the weights have considerable influence on the results of the assessment. (ECMT
1981, pp. 16, 23).
The second quotation has a focus on the short-comings of market pricing and
perceives MCA as an engineering approach in contrast to an economics approach.
In a comprehensive presentation of MCA methods for regional planning from
1988, it is stated that:
65
there exists the situation where the market price mechanism is not any longer well
functioning and for which alternative evaluation criteria have not yet been well established. The market price mechanism combined with the efficient allocation of resources
has not worked as the proper evaluation index for planning. This problem is known as
market failure. A major subject of MCDM (multicriteria decision methods) research is
thus to resolve the theoretical evaluation problem. this research highly intends to
take problem-solving as well as problem-findings aspects into major consideration: thus
this is an engineering approach in contrast to an economics approach. (Seo and
Sakawa 1988, p. xiii).
Hence, MCA methods are informed by preferences that are available from
decision makers. For this reason MCA accommodates an open-ended process
much better than is the case with CBA. One may say that in this way CBA takes on
a tinge of objectivity compared to MCA. The purpose of MCA is therefore not to
find some kind of correct, hidden answer but rather to assist the decision makers in
mastering the (often complex) information involved and advancing towards a
solution (Gissel 1999; DMG 2010).
Two main branches of MCA methods have been found particularly useful for
assisting decision making regarding complex strategic choices. One concerns
using multi-attribute utility theory and is represented in the toolbox with SMART
(simple multi-attribute ranking technique). This type of method consists of scaling
and weighing the different attributes of the alternatives to achieve the one which
scores the highest. The other branch proceeds by applying pairwise comparisons,
which has been found useful in the way decision makers can be involved in the
assessment. This methodology is represented in the toolbox by AHP (analytic
hierarchy process). AHP is by far the most wellknown of the pairwise MCA
methods. Another method is REMBRANDT (ratio estimation of non-dominated
alternatives); REMBRANDT can overcome certain difficulties that may arise in an
AHP session. The practical use of the SMART and AHP methods is presented in
Chap. 7, and further information about SMART, AHP and REMBRANDT is given
in Appendices A and B about the COSIMA and SIMDEC methodologies.
6.2.4 Uncertainty and Risk

Other hard methods in the toolbox address the core and wider performance of the
alternatives as regards uncertainty and risk. These types of examination draw on the
MOEs of fairness, diversity and robustness. Typically these can inspire what-ifquestions of a wide range. In this respect scenario analysis (SA) is a well-known
methodology, where critical assumptions are derived from scenarios representing
what is perceived as possible, plausible and internally consistent images of the
future (Leleur 2008). Preference analysis (PA) is a kind of hard version of the softer
stakeholder analysis (STA). Typically analysts and modellers can identify certain
parameters that they interpret as being sensitive with regard to the different decision interests involved (Ibid.). In the general demo-case in Chap. 7 it will be seen
66
that the employees and management of the company TRANS-IT Consult have
different opinions, which may have an influence on what is the best relocation
decision.
Uncertainty may hamper the expected outcome of a decision. Therefore complex strategic choices clearly involve various types of risks. The higher complexity
involved due to the open-ended type of change earlier seen as characterising the
strategic decision (with tactical and operational decisions seen as related to closed
and contained change, respectively) means that the risks involved are important. In
the SP framework risk analysis (RA) by use of Monte Carlo simulation is included
as one of the hard methods in the systemic toolbox. But as will be underscored in
what followsmainly in Chap. 8this type of calculation has to be accepted with
caution. Uncertainty and risk analysis is also treated in Appendices A and B about
the COSIMA and SIMDEC methodologies linking, respectively, cost-benefit
analysis and multi-criteria analysis in COSIMA and multi-criteria analysis and risk
analysis in SIMDEC.
6.3 Creating Choice Intelligence

In Chap. 4 a systemic process was outlined and in Chap. 5 a systemic toolbox was
presented, which in this chapter has been specified as concerns the potential of the
individual seven hard and seven soft methods to contribute to providing decision
support for complex strategic choices. How can this lead to better strategic
choices?
Some information will be given in Chap. 8 based on ten cases where SP has
been applied. At this stage the question will be reformulated and answered in a
more indirect way: Is SP able to create what I will term choice intelligence?
Answering this makes it necessary to take a closer look at intelligence as a
concept in general and try to distil what kinds of intelligence are in demand when
undertaking complex strategic choices. The following definition of intelligence has
been set out by the Board of Scientific Affairs of the American Psychological
Association (APA):
Individuals differ from one another in their ability to understand complex ideas, to adapt
effectively to the environment, to learn from experience, to engage in various forms of
reasoning, to overcome obstacles by taking thought. Although these individual differences
can be substantial, they are never entirely consistent: a given persons intellectual performance will vary on different occasions, in different domains, as judged by different
criteria. Concepts of intelligence are attempts to clarify and organize this complex set
of phenomena. Although considerable clarity has been achieved in some areas, no such
conceptualization has yet answered all the important questions, and none commands
universal assent. Indeed, when two dozen prominent theorists were recently asked to
define intelligence, they gave two dozen somewhat different definitions. (Neisser et al.
1995).
67
There seems to be no definite way of defining intelligence, but as part of the

quotation we get to see that attempts to clarify and organise complex phenomena
are a kind of indicator. Coining the more specialised concept of choice intelligence
we can define this as:
An ability to clarify and organise complex phenomena concerning foresight and
related decision making based on constructive circularity.
Specifically this clarifying and organising centre around a process that builds on
in principle unending scoping of a range of best possible choice alternatives and
assessment of their consequences and risks, which can point out the best among
the alternatives. That the process is in principle unending is due to the constructive
circularity paid attention to earlier in Chaps. 2 and 3 stating that the scoping will
frame the assessment and the assessment will frame the scoping.
6.3.1 SP as Multi-Methodology Approach

Combining methods from the SP toolbox builds on the assumption that a more
coherent and comprehensive understanding of a complex strategic choice situation
can be obtained using different ways of epistemic seeing. The contrasting of a
Simplicity paradigm with a Complexity paradigm and the introduction of the five
SP modes of enquiry, see Table 6.1, have this assumption as their rationale. We
will generally refer to this type of epistemological thinking as integrative or
cognitive pluralism (Mitchell 2004).
In practice cognitive pluralism in management thinking is behind the advocacy
of methodological pluralism, where the following three types have been categorised (Mingers and Gill 1997, p. 9):
Loose pluralism: encourages a variety of paradigms and methods but does not
specify how or when they should be used.
Complementarism: different paradigms are viewed as internally consistent and
based on different assumptions about their context of use, such that each paradigm
would be seen as more or less appropriate for a particular situation.
Strong pluralism: assumes that a situation often would be dealt with more effectively with a blend of methodologies from different paradigms.
In SP it is common for both soft and hard methods to be used. Therebygiven
the methods represented in the toolboxall five modes of enquiry (MOEs)
ranging from core performance, wider performance, fairness, diversity and
robustness can be utilised. Clearly this represents a cross-paradigmatic multimethodology approach, see Tables 5.1 and 5.2, which adhere mostly to the category of strong pluralism.
68
6.3.2 SP as Teamwork
In Chap. 4 the systemic process leading forward towards a situation where a
decision about complex strategic choices could be made was related to the
building of competence, and in Chap. 5 a systemic toolbox was formulated with
individual methods and techniques categorised as being either hard or soft. What
characterises the use of formal models and group deliberations? Gilboa has raised
the issue of how unqualified decision making can be avoided:
the use of formal models may be of great help working in groups and brainstorming
often helps. This should be qualified, because group decisions are not always better than
individual decisions. Groups that differ in their motivation may find it hard to make
coherent decisions, and if they do, the decisions may be very conservative, and may also
be swayed by charismatic personalities. But individuals who discuss a problem together
and then go their own ways to make individual decisions will generally make better
decisions than they would on their own. Groups tend to be better than individuals in sheer
analysis, with many ideas being brought up, challenged by others, compared, and analyzed. (Gilboa 2011, p. 20).
No doubt, according to Gilboa, group processes aiming at preparing and maybe

also taking a decision are beneficial if motivation is shared by all participants and
the decision problem is one that encourages the participants to come up with ideas.
This is how we may often see a situation with complex strategic choices. At the
same time, however, Gilboa makes the point that working in groups should be
qualified and thereby carefully designed and prepared.
It is essentially the purpose of SP to qualify such group processes. In the
underpinning theory some effort has been made to see such an endeavour as being
secured by drawing on and combining different epistemic lenses and paradigms.
Specifically, the cognitive billboard, see Tables 5.1 and 5.2, has helped formulate
the SP toolbox. The billboard can, however, also be used more directly in the SP
process as a means to make the individual members of the team cognitively alert.
This is done early on in the process simply by letting the team members discuss
what the individual patches really may express and whether their messages are
relevant for the complex planning problem in hand. Experience of using such
free-styling shows that this can be a way of getting the team together around the
complex planning problem to be dealt with.
Taking the advice as set out in the five-stage learning model by Dreyfus and
Dreyfus at its face value and seeking to climb the different competence levels add
up to what may be a quite cumbersome, demanding and time-consuming process.
In practical strategic decision making the process cannot in principle be unending.
In practice decisions need to be takenor they will sometimes take themselves as
no strategic decision is also a strategic decision. We can say with Stacey that the
future is under perpetual construction.
So when is a person or a team ready for a decision?with Luhmann ready to
move forward in the unending process made up of the following steps: (1) forced
to select ? (2) contingency ? (3) risk ? (1) forced to select etc.
69
Even acknowledging that highly successful strategic decisions may have been
taken by one single person with a short time to prepare it, it is the idea and message of
this book that for important complex decisions it is worthwhile to spend time and
accumulate a certain amount of efforts for preparing decision making. At the same
time it is worthwhilewith the way the SP framework has been set upto organise
the strategic decision making as a group effort. Therefore SP is best pursued as
teamwork as will also be the case in Chap. 7, which presents an example of company
relocation. For a company to relocate its headquarters, this type of problem is
certainly an issue involving a situation with complex strategic choices.
When facing a complex planning problem requiring strategic decision making
the initial concerns relate to gaining insight into the nature of the problem and
to considering in principle all aspects of relevance. With this as our point of
departure two major activities are scoping and assessment. Scoping aims at
determining a preliminary set of choice alternatives, whereas assessment aims at
identifying the most attractive of the choice alternatives.
In the scoping mainly the soft methods in the systemic toolbox assist the
planners in their deliberations, while mainly the hard methods are applied in the
assessment for the determination of the consequences and risks that relate to
each of the choice alternatives.
Scoping and assessment are necessarily interrelated activities. What matters in
scoping is that an option or choice alternative is not excluded if it later on in the
assessment could have come forward as a serious competitor to the alternative
assessed as being the most attractive one. Therefore the scoping should be
returned to and reconsidered on the basis of the assessment.
All the 2 9 7 methods in the toolbox have their particular functions and relate in
different ways to the five MOEs behind their inclusion. As the later case
descriptions bring forward they are used in combinations that were found
suitable in the specific study.
Use of group processes arranged as decision conferences and making use of soft
and hard methods in combination are major characteristics of systemic planning.
Finally in this chapter some evidence is given that the blending of methods,
known as multi-methodology, and making decisions as teamwork are both
effective and beneficial for the end result. The validation of SP as a decision
support approach is carried out after the detailed demonstration of one possible
use of SP in the following Chap. 7.
References
Checkland P (1981) Systems thinking systems practice. Wiley, Chichester
Checkland P (1985) The approach to plural rationality through soft systems methodology.
Lecture Notes in Economics and Mathematical Systems 248, Springer.
70
DMG (2010) Decision modelling group compendium decision support: theory and practice,
DMG. DTU Transport, Technical University of Denmark, Copenhagen
ECMTEuropean Conference of Ministers of Transport, Group of Experts (1981). Exchange of
information criteria applied to transport infrastructure projects. ECMT Publication, Paris
Gilboa I (2011) Making better decisions: Decision theory in practice. Blackwell, Malden
Gissel S (1999) Decision aid methods in rail infrastructure planning, Ph.D. thesis, Report 4,
Department of Planning, Technical University of Denmark
Jackson M. C (2000) Systems approaches to management. Kluwer Academic, Berlin
Leleur S (2000) Road infrastructure planningA decision-oriented approach, 2nd edn. Polyteknisk Forlag Lyngby, Denmark
Leleur S (2008) At navigere mod fremtiden: systemisk planlgning som ide og metode.
Mingers J and Gill A (1997) Multimethodology. Wiley, Chichester
Mitchell S.D (2004) Why integrative pluralism?, emergence: complexity and organization,
6 (12):8191, ISCE Publishing
Neisser U and APA Task Force members (1995) Intelligence: Knowns and Unknowns, report
from the American Psychological Association (APA)
Seo F, Sakawa M (1988) Multiple criteria decision analysis in regional planning. D, Reidel
Publishing Company, Dordrecht
Ulrich W (1983) Critical Heuristics of social planning. Wiley, New York
Chapter 7
Company Relocation as Demo-Case
7.1 A Complex Decision Task: Relocation

of TRANS-IT Consult
The purpose of this chapter is to provide an example of applying the systemic
principles to a specific case. It should be noted that the case is built around a real
case but in agreement with the company it has been modified and given a fictitious
company name. The case was developed as part of the Danish-Swedish research
project about the resund Region reported in Barfod et al. (2008).
The case concerns the consulting company TRANS-IT Consult Denmark A/S.
Due to various circumstances we are dealing with a complex strategic decision
task. Some issues such as lack of office space push towards moving to new
facilities, and other issues have also become relevant in the discussion of relocating the company. To introduce this relocation case, a company profile is given
below with indication of company vision etc.
TRANS-IT Consult Denmark A/S was established in November 2006 as a
subsidiary of the international group TRANS-IT Consult Ltd. TRANS-IT Consult
in Denmark has around 200 employees and worldwide approx. 15,000 employees;
this constitutes one of the worlds leading consulting companies in GIS, transportation, IT, environment and industry, and TRANS-IT Consult provides innovative consulting within these areas both in the domestic markets and
internationally. The focus of the Danish branch of TRANS-IT Consult is the rapid
development in the resund Region and the opportunities that this development
can lead to for the transport industry. The consulting is based on extensive
knowledge and years of experience in the industry and recognised global
capabilities and experience. The companys vision is to become the leading
Scandinavian consultant while consolidating and extending its present strong
position as an international company that is rooted in Scandinavia with a highly
recognised expertise in its field.
The decision to be taken has the character of a complex and strategic one. There
has been a growing recognition of its importance and that a decision ought to be

71
72
7 Company Relocation as Demo-Case
taken soon. There are a number of requirements to be met by the new site and
buildings ranging from a restructuring of offices into an office landscape on the
basis of a just completed, although minor reorganisation to including also the
concern that the new location should be in line with the companys image. Among
a number of other requirements are also that the economic consequences should be
sound, which means that although relocation will open the door to new opportunities, the economy is not without significance. Another requirement concerning
the relocation decision is that the staff emphasises that it must not be too difficult to
get to work. A number of factors are thus to be taken into account and some of
these are not readily measurable. A move from the present central location in
Copenhagen close to a major transport terminal should in general with all pros and
cons taken into account be considered advantageous.
The management decides to set up a group with representatives from both
management and the various departments and staff groups in order to prepare a
basis for the final decision. A newly employed engineer who has eagerly participated in the discussion and has raised the issue at the companys annual meeting
is invited to become a member of the group as an input from fresh eyes is
welcomed by the CEO of the company. Accepting this invitation he proposes that
a systemic planning (SP) approach may be useful, which he suggests on the basis
of his recent attendance in a class in planning theory at the Technical University of
Denmark. His proposal is prompted by the numerous factors and uncertainties he is
already aware of. It is agreed to give SP a chance as, by considering the long-term
consequences for the company, it seems reasonable that factors other than the
purely economic ones should influence the final decision about the relocation site.
7.1.1 Application of Systemic Principles

At the first meeting of the Move Group (as the group formed to prepare the
decision has already been baptised), the members discuss how to get started.
Actually they undertake a brainstorming and in the light of this also a stakeholder
analysis. What comes out of these two soft methods is fundamentally that the
economy is important, but should not count too much and that the basis for
decision should not only reflect the views of the management, but that staff
should also have its say.
In this way a first consensus has been created stating that the decision should be
based on both economic and non-economic issues and that these must be seen both
from a management point of view and a personnel standpoint. Furthermore, it has
been decided that the Move Group will consider the results stemming from a
management consultants report based on the economy at different locationsthis
report was actually already ordered by the management of TRANS-IT Consult
before the Move Group was established. But it is also decided that in the final stage
of its work the Move Group will organise a workshop (later on the Move Group
will refer to this as a decision conference).
7.1 A Complex Decision Task: Relocation of TRANS-IT Consult
73
On this basis the work of the Move Group falls into three phases: (I) they will
discuss the commissioned consultants report when available, (II) they will then
work to mark areas beyond the economy which should influence the decision, and
(III) they will conduct a final decision conference to establish their recommendation for the decision to be taken.
7.1.2 Consultant Report and Preparation

for the Decision Conference
The management consultant firm began its work by identifying a total of eight
alternative sites, namely four on the Danish side and four on the Swedish side of
resund, the belt between Denmark and Sweden. As presented in the consultant
report the sites have come forward as a result of interpreting the various motives
and arguments for giving up the present office location in mid-Copenhagen.
The different sites are given a first consideration by the Move Group on the
basis of the advice that scoping in SP is important to avoid an unintended constraining of the decision space. It is important that a site included in the set of
relocation alternatives meets the requirements resulting from the brainstorming
and stakeholder analysis. A representative from the management states that she
sees an alternative on both the Danish and the Swedish side of resund that in her
opinion are completely unnecessary to include and that efforts could be saved by
removing these possibilities already now. Another member of the Move Group
points out that when reading all the background material provided by the management consultant, it seems that a very close examination has been carried out.
Therefore it would not be right to exclude two alternatives from the further work.
An argument from the newly hired engineer that six alternatives instead of eight
alternatives would not considerably affect the total amount of work to be carried
out ends the discussion, and it is at this stage agreed to continue with all eight
alternatives suggested in the management consultant report.
The eight alternatives are shown in Fig. 7.1 and can be briefly characterised as
follows: on the Danish side of resund located in central Copenhagen near the
channels we find Tietgens Hus, which is an older and very prestigious building
located near Copenhagens commercial centre. Arne Jacobsen All in the modern
restaden is a rather new high-tech building, which is located in the new business
district close to motorway and airport. Vallensbk Torvevej is located outside the
city centre in a cheaper and less prestigious area; however, this is a new office
building with all modern facilities attached. An office building in the industrial
district behind lholmparken makes up the regional north alternative, which is
located outside of the regions real centre, but which is expected to have a strategic
location depending on developments in the region; the alternative is referred to as
lholmparken-Hillerd.
On the Swedish side of resund, we find Vstre Hamnen in Malm, which equals
the central location of Tietgens Hus in Copenhagen with a central location in
74
Denmark:
D1: Tietgens Hus Copenhagen
D2: Arne Jacobsens All restaden
D3: Torvevej Vallensbk
D4: lholmparken Hillerd
Sweden:
S1: Vstre Hamnen Malm
S2: Hyllie Centre Area Malm
S3: Svgertorp Malm
S4: Sder Helsingborg
Fig. 7.1 The eight alternatives for the relocation of TRANS-IT Consult
downtown. Vstre Hamnen is a brand new high-tech office building, specifically

designed to handle IT companies. Hyllie Centre Area is also a new high-tech
building, but located in a new area in Malm, which is ultimately planned to
countermatch the Danish restad. Svgertorp is in the more established industrial
area in Malm, which in prestige is comparable to the Danish Vallensbk site. Sder
in Helsingborg is the northern location site outside of the regions centre; considering
the development planned for this area it can become a relevant strategic location.
At its next meeting the Move Group once again discuss the proposed alternatives and end up concluding that the consultant has managed to screen the entire
region with regard to possible locations. According to the task formulated for the
management consultant, an overall requirement was that the new location of
7.1 A Complex Decision Task: Relocation of TRANS-IT Consult

Table 7.1 Overview of the
B/C rates for the eight
alternatives
75
Alternative
B/C rate
Tietgens HusCopenhagen
Arne Jacobsens Allrestad
VallensbkCopenhagen
lholmparkenHillerd
Vstre HamnenMalm
HyllieMalm
SvgertorpMalm
SderHelsingborg
1.00
0.80
2.40
1.25
2.07
3.05
2.83
3.64
TRANS-IT Consult should be somewhere in the resund Region, but that within
this constraint no site which ought to be considered must be left out.
Next the Move Group address the economy calculations of a cost-benefit type,
which have been created by the consultant. Based on a wealth of information about
rent levels, operating costs, removal costs, efficiency gains through redeployment
and costs associated with breaking up (disruption costs) and differences of Danish
and Swedish wage levels, etc. B/C rates have been established, which express the
attractiveness of each alternative from an economic viewpoint. These results
represent the core performance according to SP terminology and are shown in
Table 7.1.
Specifically, this table indicates that in economic terms a location in Helsingborg
will be preferable. Since almost all rates are above 1, the B/C calculations also show
that the very decision to move seems economically correct.
The B/C-rates give rise to intense discussion of, among others things, what is
currently included and what is yet to be taken into account. As concerns the latter
it is furthermore discussed whether these matters are possible to include by
calculation or only by being addressed by wording.
After the work programme earlier agreed upon, the next phase concerns getting
more clarity about the influences SP addresses as non-economic and which relate
to the wider performance of the alternatives. In reality it is hard to separate
economic and non-economic factors. One of the factors that has already, not least
from the staff side, been given much attention is the accessibility of the new
location compared to the existing location, which is quite close to commuter trains
and subway at Nrreport Station in central Copenhagen.
This situation could well be a restarting of the cost-benefit calculation as there
are well-defined transport costs and an established practice of valuing the savings
and costs related to changed travel time. There are also ways of calculating the
effects of more or fewer shifts and for waiting-time in this respect. Depending on
the specific alternative some home-to-work travelling may change to car travelling
(the present location is very accessible by public transport). This effect could be
calculated in economic terms comparing the difference between the before and
after situation as it is called in cost-benefit analysis. Actually using a geographic
76
information system (GIS) and applying the knowledge of staff residences fairly
accurate calculations become possible.
The budget-responsible person of the Move Group, however, realises that a new
extended cost-benefit analysis conducted by the management consulting firm with
inclusion of changed daily travel patterns of the employees will cost so much that
there will be no budget means left for the planned final decision conference.
However, based on a suggestion from the newly hired engineer that a less costly
approach can be used when considering also the change in travel pattern, it is
decided to stick to a work programme with a final decision conference.
Actualised by the home-work accessibility issue the Move Group asks whether
there are other methods in the systemic toolbox that could be used to gain insight
into the non-economic factors considered to influence the choice of location.
It appears that in addition to brainstorming more demanding methods exist, which
can be used to pursue the issue. The methods in this respect are soft systems
methodology (SSM) and critical systems heuristics (CSH). These methods are very
suitable if you start from scratch. But the Move Group finds that it is actually not
the situation anymore. There have been until now three sessions, and there is a
feeling in the Move Group that both phase 1 and phase 2 of the work programme
have been covered reasonably well so it is time to prepare for phase 3 with the
decision conference.
7.2 Principles and Steps of the Decision Conference

7.2.1 The Principles of Decision Conferences
Organising a decision conference is one way in which the systemic process as
group learning can be intensified. With a look back at the contents of Chap. 4 it
can be said that the decision conference is a means to create optimal conditions for
a structured discussion that can advance subworld formation around the strategic
decision task. The reason is that the decision conference brings together three
important elements: decision analysis, group processes and information technology, see Fig. 7.2.
A decision conference can generally be seen as an approach that enables a team
consisting typically of decision makers and stakeholders to work together in order
to prepare a complex decision to be taken (Phillips 2007). In the decision conference the participants bring together maybe even widely different views on a
companys vision and goals, which supported by a facilitator and analysts feed into
the collective learning that the team members undertake together.
More schematically, a decision conference is characterised by the following
issues based on (Goodwin and Wright 2010, pp. 317319):
Consists of a number of sessions conducted over one or two days (time and
number of sessions may vary depending on needs)
77
Fig. 7.2 Decision

conferences as an interplay
of decision analysis, group
processes and information
technology
Aims at analysing important issues

Involves key people who represent different perspectives on issues
Is facilitated by an impartial specialist in group processes and decision analysis
(facilitator)
Uses a decision support model operated by a decision analyst, which helps to
give the process structure
In practice the group of decision conference participants are placed around a
table with the fundamental aim to receive oral and screen-based information,
discuss the problem and give feedback to influence the upcoming activities. The
decision conference is as already mentioned conducted by a facilitator who guides
the process at its different stages. Interactive decision support information technology is made use of to undertake the on-site modelling of the decision
engineering methods applied, which actually consists of different types of decision
analysis methods driven by inputs from the participants.
In front of the table one or more large screens (smart boards in combination
with IT and decision analysis software) are positioned in full view of all conference participants. The decision analyst is not participating directly in the process
but is important for a smooth running of the software on the basis of the various
requests made to model calculations as part of the process. In addition to this basic
set up of the conference, experience has shown that generally a good physical
environment of the conference will contribute to a successful conference. Part of
this is that interruptions from the outside, if not completely avoidable, are kept to a
minimum. If the budget permits it an out-of-office residential stay is not a bad idea.
The fundamental goal of a decision conference is to create a synthesis from the
dynamic and creative deliberations of the group processes and the applied decision
analysis techniques. Essentially such a synthesis derives from a set of steps with
intermediary results which feed back into new model runs and new model outputs,
which then can trigger new deliberations. In a successful conference participants
may gain a sense of common purpose and what may be interpreted as a kind of
ownership of the decision to be recommended or taken as the overall outcome of
the decision conference.
78
In this respect it is important that the conference facilitator initially explains the
underlying decision model theory to the participants. This needs not be with all
details set out but in a way so the participants obtain an overall understanding of
how the process is organised and why the different steps taken are important. In
brief there should be no black boxes as they will detract from the purpose of a
common learning that should lead to a good decision. Therefore the decision
conference should appear as a natural way of unfolding the complex decision
problem without having to bother too much about technical details. The description below of the decision conference conducted to obtain a best choice about the
new company location in the resund Region seeks to demonstrate that this can be
achieved.
7.2.2 The Relocation Decision Conference: Decision Tree

and Pairwise Comparisons
Based on the first two of altogether three work phases the Move Group members
initiate the final decision conference. The conference begins with a summary of the
findings from the previous phases 1 and 2 and an overview of the principles to
underpin the conference, which, among other things, include that all arguments
have the right to be heard and that this will be secured by the facilitator.
The facilitator outlines how the conference is structured around a series of
questions to be debated and how the groups deliberations concerning answers to
the individual questions will influence the progression of the work. After this
introduction, which has also included the management consultants report and a
review of its results, the facilitator asks the participants the first question:
Question 1. By applying CBA it was found that SderHelsingborg is the
most attractive alternative. Do you agree?
If the participants agree with this outcome and feel that the CBA results make
up an appropriate solution to the problem, which would often be the case for
more standardised assessment tasks, the decision conference may soon be over.
Participants will, however, when having non-standardised tasks in front of them,
often disagree with the pure CBA-based choices and feel that the analysis is
inadequate in its coverage of the problem. This reflects the complexity of the
choice alternatives presented to them, which leads the facilitator to ask the
second question:
Question 2. Is it possible to explain and put words on what is currently
missing in the analysis? Can this lacking influence be expressed by formulating some criteria?
79
This instigates a larger discussion, and it becomes relevant to draw on the views
that were formulated already in phases 1 and 2. These can now be reconsidered
and further elaborated on to form a set of criteria. The facilitator points out that the
criteria should not overlap and that they will function to represent in principle all
factors of relevance for the relocation decision.
After some deliberation the participants decide that the following criteria, or
effects, should be included as they are defined below.
Proximity to customers: The companys position in relation to its primary
customers
Image: The image the location and the building(s) present, especially to
customers
Office size: The size of the new office environment (m2/person) and the layout
flexibility
Site facilities: The facilities associated with the site and the surrounding area
Parking facilities: Parking facilities for cars, etc.
Public transport: Its accessibility by public transport (bus, subway, train)
Individual accessibility: Its accessibility in relation to individual transport
(car, etc.)
Global accessibility: Its accessibility in a global perspective in terms of proximity to airport
At this stage it is important to emphasise that the list only highlights that the
mentioned criteria are expected to influence the choice of the relocation alternative, not how important each criterion is in comparison with the other criteria on
the list. Therefore the order of presentation of the criteria is not relevant as it does
not reflect a relative importance associated with the criteria. Including also the
CBA assessment from the management consultants report an overview of decision criteria can be set out in a decision tree introducing two intermediary levels
between the relocation choice problem and the criteria that have been formulated,
see Fig. 7.3.
Taking a look at the decision, it is essential to consider whether relevant criteria
are missing and that overlapping between the criteria does not occur and that they
can be regarded as truly additive to the effects in the cost-benefit calculation. The
facilitator here tells the participants that a preliminary decision foundation has
been established consisting of what is commonly referred to as the cost-benefit
analysis (CBA) effects and the multi-criteria analysis (MCA) effects. From a
methodological point of view she states that some major questions remain, namely
how important the MCA effects are seen relatively to each other and how
important the group of MCA effects combined is compared to the group of effects
included in the CBA. She introduces the term trade-off between CBA and MCA in
this respect as one of the upcoming tasks. First, however, it will be necessary to
address the MCA effects more closely. This concerns what is called the scoring of
each relocation alternative under each of the MCA criteria/effects. Characteristic
here is that for each MCA criterion a value function (VF) is defined so that the best
80
Fig. 7.3 The formulated decision tree for the TRANS-IT Consult relocation case
alternative is assigned the value 100 and the lowest the value 0. The other alternatives are graded linearly so they obtain scores between 100 and 0.
The technique made use of here is the analytic hierarchy process (AHP) (Saaty
1977). Under each of the eight MCA criteria all the eight alternatives are compared
two at a time. This process is supported by the smartboard screen and facilitated
with regard to software prompts by the analyst, and for each MCA effect in the
decision tree the participants are asked how alternative A performs compared to
alternative B under this criterion. The A versus B performance (A and Bs
attractiveness in accordance with, for example, proximity to customers) can be
better, equal or worse (the latter corresponds to B being better than A).
81
The essence of this scoring is that all preference information collected from the
group in this way can mathematically produce the VF scores for all alternatives.
For each of the eight MCA effects eight scores are thus produced, with one of these
being 100 and one 0 and the remaining scores in the range in-between. Evidently
what matters here is how the preference information is graded by the group, and
for this purpose a semantic scale is used. The grading here is explained by just
paying attention to A being better than B (as A being worse than B can be
transformed into B being better than A).
With A being better than B the group has to decide on the following semantic
gradings (Saaty 2001, p. 73):
Equal importance
Moderate importance
Strong importance
Very strong importance
Extreme importance
These are the five main gradings that the group initially has to consider. In case
that doubts are expressed about which of two neighbouring grades that best
expresses the intensity of importance, in-between gradings can also be used as
indicated below with grading-numbers also shown (Ibid.) (Note that the gradingnumbers are treated in Appendix A. Here it suffices to observe the semantic
presentation of the various intensities that can be made use of by the group
members in a pairwise comparison).
Equal importance
In-between grading
Moderate importance
In-between grading
Strong importance
In-between grading
In-between grading
Extreme importance
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
The facilitator informs the group that their promptsassisted by the analyst
feed into a numerical method that calculates the VF scores. Furthermore, asked by
one of the participants about interdependence, she explains briefly about the
concept of transitivity simply meaning that if football team I beats team II and
team II beats team III, then we should expect team I also to be able to beat
team III. What makes football entertaining is that it is not always so! In scoring,
however, she explains you have to observe consistency in the grading to some
extent so transitivity is not violated. However, she adds that the grading in fact by
answering all possible pairwise comparisons produces a kind of surplus information that helps stabilising the outcome so that minor inconsistencies can be
handled. Furthermore, she notes that the software has a capability to indicate
82
whether inconsistency in their grading surpasses a practical limit. Being very

devoted to its task the Move Group does not see this in action.
The scoring is perceived as a core task within the work and with eight alternatives and eight criteria this amounts to 28 pairwise comparisons per criterion
that for eight criteria produces altogether 224 pairwise comparisons. The facilitator
uses this observation to notice that with a growing number of alternatives and
criteria the pairwise comparison work may be cumbersome. Had the group only
had to deal with five alternatives and five criteria the total number of pairwise
comparisons would have been limited to 50 pairwise comparisons, which is
considerably lower not least from a practical viewpoint.
With this backgroundafter having covered one of the eight criteria onlythe
facilitator says that the software makes it possible to exclude some comparisons.
This will, however, happen on the expense of surplus information checking for
inconsistency. Furthermore, she tells the group that they also have the possibility
to reconsider whether all the remaining criteria really need to be included.
Actually the work in the decision conference has been organised as two
afternoons with the first afternoon drawing to an end. In the discussion some
concern is given to the point that scoring has an air of objectivity as compared, for
example to the assessment of criteria importance foreseen to take place at the last
session. At the same time the Move Group is reluctant to give up criteria, each of
which has been given notable attention in previous discussion. Even though the
facilitator says that with eight criteria it is expected that the least important
criterion will have a rather low influence on the later attractivity ranking of
location alternatives, a participant sets focus on the point that the Move Group at
this stage of the process is not able to agree about which criterion is actually the
least important to the group. Another solution, however, attracts attention, namely
that a subgroup of participants particularly interested in the methodology and
software part of the decision conference can continue the work on pairwise
comparisons and then, when the final part of the decision conference is carried out,
can introduce this with a presentation of the outcome of the scoring for the whole
group. Deciding this, the first part of the decision conference ends.
7.2.3 The Relocation Decision Conference: Scores,

Weights and Trade-Off
When the participants meet again and resume the decision conference, the following VF scores are shown, see Fig. 7.4.
No doubt this is an important step in the overall process and leads the facilitator
to ask the third question:
Question 3. Do you think that the alternatives are represented in a satisfactory way by the scores resulting from the pairwise comparisons?
83
Fig. 7.4 The scores for the eight alternatives assessed by eight MCA effects
Some issues are raised but turn out to be explainable. In fact some notes taken
by the subgroup help shed light on the image scores and the public accessibility
scores, and the underlying preference gradings that can be recalled on the screen
are accepted by the whole group. It is decided that in the reporting of the work
of the Move Group a kind of log book will be added as an appendix giving
others the possibility to look into gradings that they consider to be of main
interest for the decision making. Fortunately both the company management and
the employees are represented in the Move Group but there is a general
agreement that log book information is important and would have been even
more so, had not both main stakeholder groups been represented in the work.
The facilitator adds that had the work been conducted mainly as consultant work
supplemented with maybe a few persons from the company as informants, the
necessity of producing information about the critical parts of the process would
evidently have grown in importance.
In this way each alternative is scrutinised, and it is found that the scores seem to
represent the alternatives pretty well. The facilitator notes that the columns for
each alternative cannot just be added to point out the most attractive site as this
would reflect equal importance of each effect; however, she adds that the scores to
some extent indicate which alternatives can end up being the most interesting to
concentrate on when the final decision is to be taken.
That the MCA effects are not equal in importance is the background for the
fourth question asked by the facilitator:
Question 4. Can you agree about an order of priority for the MCA effects as
concerns an assessment of their relative importance?
84
Table 7.2 Two rankings and associated ROD weights

Priority
Management
Employees
ROD-weight
1
2
3
4
5
6
7
8
0.23
0.20
0.17
0.14
0.11
0.08
0.05
0.02
Image
Proximity to customers
Global accessibility
Site facilities
Individual accessibility
Public transport
Parking facilities
Office size
Individual accessibility
Public transport
Site facilities
Parking facilities
Office size
Image
Global accessibility
Proximity to customers
The facilitator briefly explains why this particular question is asked. In fact she
could have asked the participants directly as a group to enter criteria weights summing to 1. However, experience shows that such a process when viewed group
dynamically is quite complicated. As was the case with the previous scoring of the
alternatives, pairwise comparison could also be a possibility to set the criteria
weights. However, she states that another approach will be used. This concerns a
ranking approach which can better accommodate iterative group deliberations due to
the more subjective nature of weight-setting as compared to the more objective
comparisons of the alternatives made to assess their scores. The technique she has in
mind is rank order distribution (ROD) weights that by mathematical principles and
probability theory can provide a good estimate of the weights that lie behind a given
ranking of the MCA effects (Roberts and Goodwin 2002). The facilitator then sets
focus on whether a joint prioritisation of the MCA effects can be achieved.
However, this happens not to be the case in this group of decision conference
participants. The ranking of non-monetary MCA effects thus depends on whether
the company management representatives set the rank order of criteria or whether
the employee representatives do. Therefore it is decided to continue with two
rankings, namely a ranking or priority order devised by the company management
and a ranking devised by company staff.
Having recognised that this was the best way to move forward, the Move Group
splits into two subgroups to focus on the respective rankings without undue
interference. However, before doing this the management representatives briefly
note that at the end of the day the decision belongs to the company, which makes
the employees respond that when raising the issue for the first time some months
back the managing director had emphasised the necessity to take all interests into
account. In this way Table 7.2 is produced showing the two different rankings and
their associated ROD weights.
The two rankings can be viewed as an expression of two different attitudes or
sets of preferences towards the relocation decision. Thus the table shows that
management will prioritise criteria such as image, proximity to customers and
global accessibility. These are primarily concerned with the companys public face
and turnover, which are obviously very important to address for the management.
In this context, it then becomes less important to consider public transport, parking
and the size of the offices.
85
The staff, however, wishes to put great emphasis on individual and public
transport accessibility and the facilities at the location. This is not surprising as
these criteria have a major influence on their daily working conditions. Less
important for the staff is global accessibility and proximity to customers.
7.2.4 The Relocation Decision Conference:

An Intermediary Note
It should be noted that in the case description focus has been on process aspects.
With the focus on the group processand in this respect on the factual steps taken
to move forward in the process more than on the interpretation of individual and
group knowledge buildingless attention is given to the formal methodology and
even less attention to the information technology, i.e. the software and screen
layouts.
As regards formal methodology the reader is referred to the introduction given
in Chap. 6, whereas a more in-depth treatment of formal methodology is given in
Appendix A and Appendix B, comprising, among other things both the AHP
method and ROD weights. Appendix A also gives a thorough treatment of the
COSIMA methodology applied below to support decision making on the basis of a
combined use of CBA and MCA, whereas Appendix B concerns applying risk
analysis (RA), also given attendance to below, in combination with MCA, which is
addressed by the SIMDEC methodology. At the same time these appendices also
treat the software developed for the methodology made use of in the SP decision
conference.
7.2.5 The Relocation Decision Conference: MCA Results

and CBA Versus MCA Trade-Off Analysis
With the scores and weights determined for the MCA effects it is possible to
achieve a result for the MCA model component. Afterwards, this result will
together with the earlier CBA result lay the basis for the overall result as regards a
ranking of alternatives considering both the CBA and the MCA effects. This is
what is referred to as CBA versus MCA trade-off analysis.
The MCA result is obtained by multiplying scores with weights and for each
alternative adding these individual contributions; this is what is usually referred to
as a simple, linear-additive model. As the Move Group has produced both a
management and a staff weight set, the MCA result shown to the participants
appears as two bars for each alternative, see Fig. 7.5.
It appears that the attractiveness of the individual alternatives depends on
whether the management or the staff preferences are used. Given this difference
86
Fig. 7.5 MCA result: management and staff
which to some participants is smaller than they earlier conceived it to bethe two
stakeholder groups, however, can agree that Arne Jacobsens All is the most
attractive alternative and Sder, Helsingborg, the least attractive.
Were the final decision to be based solely on the non-economic factors and thus
ignoring the economic factors presented in the management consultants report
with the CBA results, the Move Group would be pretty close to having a foundation for a location decision. An interesting fact is, however, that the CBA
component suggests a solution that is the opposite of the one based on the MCA
component of the overall analysis as Arne Jacobsens All has the lowest B/C rate
of all the alternatives, see Table 7.1. Furthermore the least attractive location
choice according to MCA Sder-Helsingborg is when based solely on the CBA the
most attractive. This accentuates the need to consider CBA versus MCA, which is
the concern of the CBA versus MCA trade-off analysis.
The facilitator presents the principles to be applied in the CBA versus MCA
trade-off analysis, and the analyst contributes with some relevant screen shots to
illustrate how to make use of both the CBA and MCA component parts at the same
time. More technically this concerns how to weigh the two parts together.
A basic principle here is that the results of the CBA part should not be changed,
which means that the calculated B/C rates must be maintained. What is being
changed is the MCA influence on the overall attractiveness relative to the CBA
influence, see Fig. 7.6, which serves to describe the meaning of the trade-off
parameter named MCA%.
As CBA is kept fixed, what varies is the MCA%. This parameter as explained by
the facilitator can be understood in an intuitive way so that a high MCA% means a
high influence on the overall attractiveness assigned to each alternative dependent
87
Fig. 7.6 Illustration of the MCA% in the CBA versus MCA trade-off analysis
on its MCA-scores profile, whereas a low MCA% means less importance paid to the
MCA component. If an even influence is sought the MCA% is set to 50%; a very
high influence can be 80% and a very low one can be, for example, 10%.
On this basis the facilitator asks the following question:
Question 5. How do you perceive that CBA and MCA should affect the
overall result expressed by an appropriate MCA%?
As the ranking of MCA effects could not be reached based on consensus about a
single priority-order listing of criteria, it is conceivable that a single pre-defined
MCA% cannot be agreed upon either. Relatively soon the participants request to
see the consequences of a variety of chosen MCA% values, which then could be
the point of departure for further discussion. Agreeing to proceed in this way, the
facilitator, however, asks each participant to write down a value (not at this stage
to be shown to the other participants), which can be used as background for the
upcoming further discussion.
Furthermore, it is agreed to continue both with the management and the staff
MCA results. Here a large screen or even a number of screens are advantageous as
they allow results from the two decision strategies to be studied simultaneously.
First to be shown on the screen are the results according to the management
strategy. The result is indicated as a value composed as a contribution from CBA
and from MCA and is described as the total rate of return (TRR) value. Thus TRR
for a given MCA% expresses the overall attractiveness based on a CBA and an
MCA contribution. The background for expressing TRR as a rate is that MCA
88
Fig. 7.7 Management decision strategy: total rate of return (TRR) based on applying both CBA
and MCA for different values of the trade-off parameter MCA%
(similar to CBA) is considered also by taking costs into account. Therefore TRR
comprises the value of the economic (monetary) and non-economic (non-monetary) effects when compared to the total cost invested in the relocation of the
company from its location in central Copenhagen to the studied location alternatives. From the first screen that appears with all eight alternatives it is possible for
the participants to narrow down and select, for example, the four most interesting
alternatives, which has been done in Fig. 7.7. For the MCA% equal to 0 the CBA
result is shown on the TRR-axis; as MCA is weighed in with still higher MCA%values, the relative performance of the alternatives change as indicated by the
curves.
It can be seen that the TRR of the individual alternatives depends on the actual
value of the MCA%. In the interval from around 0% and until around 25% Sder
in Helsingborg is the most attractive alternative, which is in accordance with the
result of CBA. From around 25% and until around 45% the picture changes so
Hyllie in Malm becomes the most attractive alternative. Note also that in this
interval Arne Jacobsens All changes from being the least attractive alternative
and from around 45% it becomes the most attractive one for relocation.
This development is in agreement with the result of the MCA, which is more or
less the opposite of the CBA result. This is not surprising as alternatives seen as
attractive in the MCA will gain in overall performance when a higher MCA%,
which means a stronger leaning towards the MCA assessment, is applied. The
change in MCA% is also to be seen as a decrease in the consideration of the core
performance associated with CBA towards an increased influence from importance
of the alternatives wider performance associated with the MCA. Note that an
alternative as Helsingborg with low scores on the MCA effects is only moderately
adjusted with a higher MCA%.
89
Fig. 7.8 Staff decision strategy: total rate of return (TRR) based on applying both CBA and
MCA for different values of the trade-off parameter MCA%
In a similar way the TRR values are calculated in accordance with the staff
decision strategy, see Fig. 7.8, also showing only the selected four most interesting
alternatives.
Figure 7.8 shows more or less the same tendencies as Fig. 7.7. Again the
participants realise that no profound differences occur between the two decision
strategies. What really happens is that the intervals of attractivity are slightly
changed. Helsingborg is still the most attractive alternative between 0 and 25%,
and Hyllie now becomes the most attractive between 25 and 55%; then again Arne
Jacobsens All is the most attractive alternative for the higher MCA% values.
The difference in outcome as regards the management and the staff decision
strategies are due to the different priorities for the relative importance of the MCA
effects. The facilitator notes that greater differences could occur given that other
distributions of MCA scores had been found. Nobody, however, wants to reconsider the scoring made earlier.
The MCA% to be decided as recommendation for the final decision making
ought to depend on the type of decision problem. It will be relevant here to give
examples to the participants. For a large infrastructure investment maybe a
maximum of 30% would be relevant, but in the actual case higher values could be
applied as the relocation decision for the company may be the most important for a
long time. As a rule it can be stated that the uniqueness of the strategic decision
matters. This information starts a discussion among the participants about not
choosing a too small value. At the same time issues turn up about general economic development (regional, national and international) relating to the companys markets and otherwise. This leads to a wish to engage in further
consideration about taking possible alternative developments into account and also
to engage in whether the decision to be taken will be robust.
90
7.2.6 The Relocation Decision Conference: Scenario

and Risk Analysis
At this point, several things can happen in the process. There may be a sense of
being close to a basis for decision, but there may also be a wish to return to the
input given to the earlier questions in the conference. An alternatives performance
relative to the other alternatives in the result graph based on the sliding MCA%,
see Figs. 7.7 and 7.8, can make one or more participants convinced that there is a
criterion we may have completely forgotten or that maybe with respect to a
particular grading in the scoring we definitely made a wrong choice with regard to
the intensity in the pairwise comparison. It is important that such uncertainty and
possible misunderstandings are cleared up. More generally, a common recognition
may also arise that an additional criterion needs to be included and/or that one of
the criteria is in need of reinterpretation. Also a recognition of a stakeholder
viewpoint not taken into account could lead to a new ranking of criteria and to
altered performance in the result of the process.
This makes great demands on the facilitator who wants momentum in the
process, but not at the expense of a lack of consensus or dissatisfaction. At this
stage the mentioned suggestions concerning making changes with regard to earlier
input into the process could develop into, for example, a reconsideration of the
scoping carried out. This did not happen in the relocation case, but possibly with
several suggestions about changes in need of being attended to a suggestion may
arise that the scoping part of the process had been given too little effort and should
maybe be reconsidered by using one of the more heavy soft methods in the
systemic toolbox such as soft systems methodology or critical systems heuristics.
However, as regards the TRANS-IT relocation decision conference the participants end up perceiving that the results provided can broadly be agreed upon,
but something seems to be missing in our deliberations because of future
uncertainty. On this basis, the facilitator formulates the following question:
Question 6. Until now a business-as-usual situation has been assumed with
regard to company development etc. Are there other situations or scenarios
as regards different, plausible developments that ought to be considered?
Strategic decisions due to their long-term consequences need to be taken with
alternative developments in mind. Both brainstorming and SWOT analysis can be
highly recommended as methodologies to explore the opportunities and threats that
can materialise as gains and losses given the different profiles of the individual
choice alternatives. At this stage the decision conference can apply the futures
workshop method to help formulate scenarios that can be used as a basis for
examining each alternative in more detail, for example, by reconsidering the earlier
answers to the questions that were made on a business-as-usual basis. Compared to
the decision conference, here seen as a format that can include all the methods and
91
techniques in the systemic toolbox, the futures workshop more explicitly helps
establishing a particular focus by running through altogether three phases: A first
phase with emphasis on being critical about the present situation, a second phase
with emphasis on outlining overarching visions for the future, and a third phase with
emphasis on the implementation of different alternatives of action.
In the decision conference the participants choose to set focus on the development
in the resund Region. Material is available from a large regional study (RIB
2007) that has, among other things, considered three developmental scenarios for the
regions urbanisation growth pattern towards 2045: (I) business-as-usual, (II)
monocentric growth, and (III) polycentric growth. In this respect a MCA% equal to
40 is chosen to serve as a point of departure meaning a close run between all the four
alternatives shown in Figs. 7.7 and 7.8 but with Hyllie as slightly more attractive
than the other.
It is recognised that in a region that is still only partially integrated, integration
between the Danish and Swedish sides of resund will continue to evolvein fact
one of the participants notes that quite a number of people from Malm on the
Swedish side of resund have actually been employed by TRANS-IT Consult in
recent years. Continued integration is expected to strengthen a monocentric
development around Copenhagen-Malm-Lund, which favours the performance of
Arne Jakobsens All compared to Hyllie. On the other hand a polycentric development may point to Hyllie as the best choice. At the same time the monocentric
development is related to a better overall economic development towards 2045
than is the case with the polycentric development. At this stage the Move Group
recognises that better economic development may point to increasing the MCA%.
The explanation is simply that the wider performance aspects associated with the
MCA effects become more important (we can better afford our vision about
TRANS-IT Consult and the image we want to build).
With a focus now on economic development and the impact this has on the
demand for consultancy and services offered by TRANS-IT Consult a background
is created for the last question to be asked by the facilitator:
Question 7. Do you find it relevant to make a risk analysis of selected factors
and parameters that have been made use of in the previous steps?
This question makes the group discuss the CBA results once again. What will
happen if the anticipated efficiency gains are only included as determined by
using various probabilities related to the different types of efficiency gains foreseen
to materialise with the relocation, and will this examination point towards this
being a critical issue? In the management consultants report these efficiency gains
were discussed but entered the CBA calculation as a set of rounded estimates.
What happens if the Danish-Swedish wage gaps fade away with time, also an issue
to be treated in the risk analysis? Will risk analysis change the view on the
individual location alternatives if, for example, the efficiency gains as estimated
92
will be affected by taking a closer look at the differences in building(s) and

equipment?
The developed software for the systemic approach is able to handle such considerations using Monte Carlo simulation. The necessary parameters can here be set
with various degrees of involvement of the decision conference participants.
However, it is also a goal at this stage that all conference participants obtain a basic
understanding of the risk calculations that are carried out. After a general introduction to risk analysis applying Monte Carlo simulation, the group discuss the
efficiency gains that relate to the individual relocation alternatives. Then the analyst
conducts a session with parameter settings in accordance with the views expressed
by the group. In this respect the risk calculations showin addition to the uncertainty that relate to the scoring and weighing of the MCA effectsthat some risks
point towards not achieving net-economic feasibility, i.e. the B/C rate could possibly become markedly less than 1 as regards Tietgens Hus and Arne Jacobsens
All. However, the results shown in Table 7.1 already pointed in that direction.
7.3 A Summing up of Case Findings

At this stage of the decision conferencepretty close to the end of the second
afternoonthere is a feeling in the Move Group of having established a basis for
recommending a decision about the new location of the TRANS-IT Consult
company. The facilitator gives the following summary of the findings:
The results from the CBA and the MCA parts respectively have identified three
main alternatives worth considering, namely Sder in Helsingborg, Hyllie in
Malm and Arne Jacobsens All in Copenhagen. The cost-benefit analysis (CBA)
offers the decision-makers a clear picture of Sder in Helsingborg as being the
economically most attractive site for its new company location. This is in contrast
to the result of the multi-criteria analysis (MCA), which points towards Arne
Jacobsens All in restaden as being the most attractive alternative. A combined
CBA and MCA rating then determines that the overall attractiveness of the
alternatives depends on the balance between the CBA and MCA influences. At a
low MCA% (below 25%) Sder in Helsingborg remains the most attractive
alternative, but as the MCA% increases (in the interval 2545%) Hyllie in Malm
becomes the most attractive one and later again (with a MCA% larger than 45%)
Arne Jacobsen All in restaden stands out as the most attractive alternative based
on considering both the core and the wider performance of all eight location
alternatives.
This finding holds for both the examined management decision strategy
(ranking of MCA effects) and the staff decision strategy. What happened was that
the staff strategy moved the MCA% upwards from 45 to 55% before pointing
unequivocally towards Arne Jacobsens All in restaden.
On this basis the conference participants agree to propose Arne Jacobsens All
in restaden as the first priority proposal and Hyllie in Malm being its nearest
7.3 A Summing up of Case Findings
93
competitor as the second priority proposal. The proposed alternatives are supplemented with comments and arguments that have been made in the light of the
progression of the decision conference.
The office relocation example demonstrates but one example of applying SP for
making complex strategic choices. Other methods could have been applied from
the toolbox, and the process could have been designed in another way.
This indicates that SP is basically an open-ended approach. Furthermore,
testing it on a diverse range of complex planning problems over the last five
years as described in Chap. 8 indicates a high degree of adaptability to the
specific study, which should be seen as a desirable feature.
In the theory chapters of the book (Chaps. 2, 3, 4 and part of 5) some efforts
have been made to pinpoint findings that can lie behind a reasonable practice for
supporting planning and strategic decision making. The SP theory skeleton
established in this way is, even in its present rudimentary version, seen as a
worthwhile and necessary effort as it functions as a backdrop against which new
ideas about process and methodology can be illuminated by the theoretical
works that were called upon over the chapters to specify the SP framework.
References
Barfod MB, Jensen AV, Leleur S (2008) STMA beslutnings-analyse: metoder, proces og
software. In: Carlsson CM (ed) Hllbart transport system fr inre och yttre attraktionskraft,
EU interreg project about strategic transport management in the resund region (STM)
MAH Malm, final STM report by Decision Modelling Group, DTU Transport. Technical
University of Denmark (DTU), Denmark
Goodwin P, Wright G (2010) Decision analysis for management judgement, 4th edn. Wiley,
New York
RIB (2007) Project resundsregionens Infrastruktur og Byudvikling, final report resundsregionen r 2045: Scenarier fr trafik och byudvikling. resundskomiteens Sekretariat,
Copenhagen
Phillips LD (2007) Decision conferencing. In: Edwards W, Miles RF Jr, von Winterfeldt D (eds)
Advances in decision analysis: from foundations to applications. Cambridge University Press,
Cambridge, pp 375399
Roberts RJ, Goodwin P (2002) Weight approximations in multi-attribute decision models.
J Multi-Crit Decis Anal 11:291303
Saaty TL (1977) Scenarios and priorities in transport planning: application to the Sudan. Transp
Res 11(5):343350
Saaty TL (2001) Decision making for leaders: the analytic hierarchy process for decisions in a
complex world. RWS Publications, PA
Chapter 8
A Summing up: The Challenge

of Strategic Decision Making
8.1 Systemic Planning: Practice and Cases

8.1.1 From Detached to Involved Understanding
The location case described in the previous chapter can serve to illustrate the
opportunities that systemic planning (SP) can offer in complex planning tasks.
A closer look into this issue may be based on the considerations of the Move Group
and whether its efforts to prepare a decision concerning relocation of TRANS-IT
Consult by means of a decision conference have actually been worthwhile?
Basically we cannot know for sure. But we do know that the Move Group via
its work, not least in this respect the final decision conference, must have
acquired very good skills with regard to this specific task. In the systemic
parlance we would say that they have established a subworld for that problem,
which means that both individual and collective building of knowledge has taken
place. In the subworld perspective of SP a local elaboration of a commonsense
world is pursued, which is methodically carried out by a circularity of scoping
and assessment based on different modes of enquiry (MOEs) and more specifically on applying and combining a number of soft and hard methods from the
SP toolbox.
The subworld perspective characteristic of SP ensures that optimisation in a
conventional way needs to be embedded in a process including also ongoing
learning and interpretation that, however, may feedback into the more calculative
processes. Organising the process as a kind of interplay between group processes,
decision analysis and interactive ITreferred to as a decision conferencecreated
a forum of accelerated learning with regard to the successive establishment of a
subworld around the complex plan problem. But how much can learning be
accelerated in this way?
This depends on how well the IT-based group process is integrated into the
above context as accelerator in combination with the interactive format for the
decision analysis. How much is learning actually raised by this process?
95
96
8 The Challenge of Strategic Decision Making
Based on the previously described five-stage learning model (novice, advanced

beginner, competent, proficient and expert), we assume that in the Move Group
undertaking the relocation case nobody ends up being neither still a novice nor
having developed into a full-fledged expert. Dependent of course on individual
skills and knowledge relating to the actual subworld-building we could see them as
knowledgeable practitioners ranging from advanced beginner via competent to
proficient. What matters when applying the five-stage model for interpreting what
is achieved is that we may assume that a majority of the participants have moved
from what in the five-stage model is called detached understanding to involved
understanding.
No doubt the facilitator plays an important role in this respect. With knowledge
of the systemic process and its objectives the facilitator can organise and induce
that situations arise in which both first and second-order learning takes place, i.e.,
learning is triggered not only by the direct screen-results and the way the procedure is designed as a structured development by means of the seven questions in
the particular case, but learning is also triggered in the individual participant by the
way the other participants understand and interpret the results.
How is the systemic process conducted in the best possible way? A number of
factors can be considered in this respect. No doubt that it is fundamentally
important that the right people join the Move Group. Everyone in the company
should in this respect feel represented by at least one person; at the same time this
person has to be sincerely committed to the task. This will not be the case with
persons that have, for example, made up their minds that they will change jobs in
the near future, etc. The recruitment to the Move Group of only highly motivated
members is therefore important.
Furthermore, it is important that the methods and procedures to be applied are
scrutinised and prepared in the right way. This means utilising in the best possible
way the resources of knowledge and background available from the appointed
members of the Move Group. A good facilitator will be able to adjust the conduct
of the conference based on the skills and the competencies that are represented in
the group. Thus the right facilitator can decisively stimulate and improve group
learning. A good analyst is also important. Experience shows that technical
incidents are not at all conducive for the work and its progress, just as the good
analyst, in addition to preventing black screens, can also contribute to a successful session by enabling upcoming what-if questions to be answered as
quickly as possible after they have been asked by the group.
In principle, a systemic process can continue for a very long time. Experience,
however, also shows that it can be completed on time and with good results if the
facilitator is very conscious about the time spent on each phase. So should the
planned time table be suspended sometimes? Again, from experience, creative
deviation can be valuable.
The good result cannot be precisely defined. In the five-level learning model at a
certain stage learning goes as mentioned from detached understanding to involved
understanding. At the same time we have passed a knowledge level in the subworldunfolding around the complex plan problem that is solely built on rationally based
97
accomplishments. According to Dreyfus and Dreyfus (see Chap. 4) arational factors

(not to be misunderstood and seen as irrational factors) affect the learning. One way
to detect this in practice is that the group begins to take ownership of the outcome.
Thereby they are on the way to gaining a kind of expertise with regard to the
complex problem they have addressed. In this respect, after a decision conference,
which ends a groups engagement in a complex decision problem, it can be recommended that the organisers hand out an evaluation questionnaire. The answers to
the questionnaire provide information about particular issues that can be either
praised or critisisedorganisers can always learn new things from thisbut at the
same time the answers will, sometimes only indirectly, indicate whether a transformation has taken place from detached understanding to involved understanding;
in case ownership of the problem is to some extent expressed in this way this can be
taken as a practical litmus test whether the SP approach has been worthwhile in the
actual problem context.
8.1.2 A Review of SP Cases

In Chap. 7 the SP approach was illustrated by the way it could support a strategic
decision about a new company headquarters location in the Danish-Swedish
resund Region. In the end a number of location possibilities were reduced to
considering four alternatives and finally one of these came out as being more
attractive than the other ones. Before seeking to validate the SP approach and
comparing it to a well-established way of supporting strategic decision making
based solely on the analytic hierarchy process (AHP), introduced as one of the
methods in Chap. 6, a number of SP-based cases will be described. With a purpose
of demonstrating the flexibility of SP and to avoid duplication of the SP steps
covered already in the case description in Chap. 7, the case descriptions will
emphasise the particular lessons learnt from each case. For each of the cases
described below, the year in which it was carried out has been indicated.
8.1.2.1 Bypasses for Hng and Allerd, Denmark (2000 and 2006)
Assessing and prioritising Danish bypasses have initiated the work on methodology for making comprehensive assessments based on both economic and noneconomic impacts, which in the present context has been referred to as core
performance and wider performance. For a detailed description of this planning
problem the reader is referred to Leleur (2000).
Generally described the purpose of a bypass around a town is to improve the
level of service for the through traffic and to relieve the town of traffic. Core
performance is assessed by using a cost-benefit analysis (CBA) comprising
impacts such as improved (reduced) driving time and driving costs (petrol savings
among other things), while relieving the town of traffic is assessed by estimating
98
impacts relating to the improvement of noise and local emissions, etc., which is
handled by a multi-criteria analysis (MCA). The latter can also include improved
conditions for pedestrians and enhanced urban quality, etc.
No doubt the design of different bypass alignments, sometimes leading to
alternatives with quite different amounts of construction costs, and the selecting of
the best alternative make up a strategic study as the outcome can have long-term
consequences both locally with regard to, for example, urban development and
often also regionally and nationally by impacting on the trunk road network, etc.
The strategic decision regarding the bypasses depends on how much CBA and
MCA should count individually as the design solutions with their individual
alignment and design standard are engineering solutions that typically seek some
kind of compromise between the objectives inherent in CBA and in MCA
respectively. The methodology from the systemic toolbox is the composite model
for assessment (COSIMA), which builds on a linking of CBA and MCA.
In Appendix A COSIMA is presented in detail together with AHP, which is
used in the MCA part of COSIMA. It is demonstrated how COSIMA can be used
to organise strategic decision making in an interactive way drawing on decisionmaker preferences. At the same time it shows how different stakeholder viewpoints
can be treated with a preference analysis.
8.1.2.2 Fixed Link Between Copenhagen and Malm (2003)

The resund Fixed Link connecting Copenhagen and Malm with a motorway and
rail line has been open since 2000. Therefore the case carried out in 2003 consists
of both ex-post and ex-ante elements: the investment equal to $3.2 billion had been
made but the long-term consequences were both complex and uncertain. This
made it an ideal planning research laboratory in the CLG study (Leleur et al.
2004a) for the exploration of planning and evaluation methodology aiming at
complex, societal problems.
Organised as a pairing of scanning and assessment with systemic and systematic exploration and examination (see Sect. 4.4), a first-round of SP searchlearn-debate activities consisting of both systematic interrelations among elements
and systemic holistic impressions led to the findings shown in Table 8.1. The
findings are listed for the four modes of examination. At this early stage of work
the findings aimed at guiding the further exploration and learning described below.
The systemic scanning comprised both general concerns and specific concerns,
see Table 8.1. One method applied was critical systems heuristics (CSH) from the
group of soft methods in the toolbox. CSH functioned as a relevant checklist but
not all questions were relevant. A similar experience with CSH has been found by
Midgley (2000, pp. 298299) but in agreement with him such an omission of some
of the CSH questions were not found to subtract from its usefulness. Collected
material about the resund Fixed Link was also used as it evidently ought to be.
Taking a wider approach to planning was substantiated by interpreting various,

Table 8.1 SP first-round examination
SP first-round findings
Systemic scanning: Issues of identification
and demarcation
99
Systematic scanning: Issues

relating to scenarios
General concerns:
resund region one of several
spheres
The meaning of national
barriers
Drivers: market, clusters,
culture, etc
Infrastructure and development
Regional scenarios:
Economy, regulation,
transport, etc
Local integration versus
non-integration
Baltic Sea development: trade, etc
Competitive transport development
Specific concerns:
Limitations of cause-effect
model
Interpreting expressed
expectations
EU-wide scenarios:
Economy, regulation, transport
Trends: resources and technology
Trends: modal policies, etc
Systemic assessment: Issues relating

to stakeholder preferences
Systematic assessment: Issues relating

to multi-criteria analysis
Ex-ante:
Local pro-coalition
Local environmental anticoalition
National interest
International pro-coalition
Core feasibility (CBA):

Investment
Time savings
Cost savings
Local emissions and accidents
Ex-post:
National interest
resund region citizens
resund companies
International interest
Wider feasibility (MCA):

Network and mobility
Global emissions (CO2)
Employment
Logistics and goods effects
stakeholder expectations, and it was found that a conventional systematic planning

approach would not be suitable to deal with these.
The systemic scanning spilled-over to influence both systemic assessment
(What motives were really guiding the industrialists in Sweden first suggesting
the resund Fixed Link?, etc.) and systematic scanning (What are the EU-wide
development trends and their implications for goods transport?, etc.). In relation
to European freight transport scenarios, a futures workshop (FW) had been
organised as part of the CLG study, and its results helped inform the scenarios
applied to the resund case (Drewes Nielsen and Homann Jespersen 2003).
As indicated by keywords in Table 8.1, scenario issues comprised a wide range of
constituent factors both on a regional and above-regional level; these together with
stakeholder groupings and their associated specific concerns (Rnnest et al. 1997)
strengthened the systemic scanning finding that conventional planning would be too
100
limited an approach to deal with the complex consequences of the resund Fixed
Link (Leleur et al. 2000).
The systematic assessment pointed to an examination including both core and
wider feasibility based on CBA and multi-criteria analysis (MCA) respectively.
The results produced by COSIMA, see bypass case above, in a first-runbased on
combining some study assumptions, for example that high integration in the
resund Region would be obtained over the next decade, and that deregulation and
high economic growth would apply generally to Europeare shown below
(Leleur et al. 2004a, b, c):
CBA: Time savings, cost savings, accidents, etc
MCA: Network and mobility
MCA: Global emissions (CO2)
MCA: Employment
MCA: Logistics and goods effects
CBA and MCA:
0.85
0.11
0.02
0.32
0.05
1.35
The sole function of such first-run results is to give feedback to and guide
the further process. They provide, among other things, the interesting result that
the figures indicating benefit-cost rates (BCRs) as used in COSIMA, show no core
CBA-feasibility (0.85 \ 1) but a wider MCA-feasibility (1.35 [ 1, with CBA as
part of the result). In other words, the feasibility is dependent on factors that are
complex and uncertain.
Such a first-round examination lays the foundation for further unfolding of the
SP process. In the actual case where the implementation had taken place, further
analysis can assist in making decision support for possible follow-up initiatives
(regulation, other investments or interventions), but in other complex cases before
the primary decision about whether to implement the project or not has been taken
(an ex-ante study), a result where narrow and wider analyses point in opposite
directions will make it relevant for the planners to resume the process.
The case demonstrates the possibility of SP to produce scanning and assessment
information that can provide an overview of the many issues concerning the
decision about the fixed link. In transport planning the planning and deciding about
large infrastructure investments are among the highly complex problems.
8.1.2.3 New Airport For Nuuk, Greenland (2007)

The decision about a new airport in Greenland servicing the capital Nuuk was to
be made among three possible alternatives:
A town-near extension to 1,800 m airfield at the present location
A town-near extension to 2,200 m airfield at the present location
A more distant location with a new 3,000 m airfield at much higher costs
101
No doubt the 3,000 m airfield is the ideal solution due to various considerations
including air technology and servicing (possible types of airplane, landing and
takeoff etc.), business, tourism, land-use with regard to developmental issues and
local environment (noise).
A decision tool was established to deal with the many economic and noneconomic issues. The tool showed that the 3,000 m alternative (from CBA
including also effects on Air Greenland) was basically not economically feasible
due to the high construction costs (low core performance), whereas both the 1,800
and 2,200 m alternatives could be defended economically but both did less well
with regard to the wider performance represented by the MCA effects included
(DTU 2007). In addition to the main study a master thesis (Mortensen and
Andersen 2007) found by combining COSIMA and CSH that the 2,200 m could be
a reasonable compromise. Specifically they conducted a stakeholder analysis that
confirmed that the 2,200 m solution would be a robust solution.
Similar to the previous case this case demonstrates that combining a number of
methods from the systemic toolbox can provide strategic decision making information that can explicate the design objectives behind the different alternatives in
a way that makes these comparable and accessible for structured debates across a
range of very different influences on the decision to be taken.
8.1.2.4 Vessel Traffic Service in resund (2008)

The following case about vessel traffic service (VTS) in resund concerns the
localisation and determination of the level of service to be offered by a new VTS
unit in collaboration between Denmark and Sweden. VTS is the collective term
for, among other things, traffic information and service offered to ships going
through the resund strait between Denmark and Sweden. The strait is made up of
a Danish part, an international part and a Swedish part. The purpose of VTS in
resund is to secure safe navigation in a narrow strait with much traffic. In connection with the construction of the Fixed Link Copenhagen-Malm a special VTS
had been established and operated in collaboration between Denmark and Sweden
to assist the extra shipping that the construction work made necessary. After the
opening of the fixed link in 2000 the unit was laid down. In 2005, however,
proposals were made to establish a permanent Danish-Swedish VTS. This soon
established itself as a complex planning problem with a number of issues to
consider.
A decision conference was conducted and what makes this especially relevant
in this context is that the soft systems methodology (SSM) was applied, with
explicit use of the rich pictures and root definitions belonging to SSM. The VTS
case was organised as two sessions and at the first meeting the issues and questions
below were made use of to produce a first version of a rich picture, where, among
other things, the concerns below about implementing and operating a common
VTS unit were addressed (Barfod et al. 2008):
102
Many institutions are involved. Where should the final responsibility be placed?
With both civilian and military organisations involved will this give raise to
invisible barriers?
What does it really mean that the cross-border collaboration will be permanent
as compared to continued operation of two national units?
How is the regulatory EU legislation affected?
How are running, operational costs affected (economies of scale)?
For some staff members a colocation in one country will mean increased time
spent on commuting. How should this be considered?
Consideration of specific location possibilities
Identification of advantages related to cross-border collaboration
Constraints relating to issues of politics
The answering of these questions gave rise to a series of rich pictures (the
facilitator happened to be highly skilled at drawing and relating statements), which
produced the root definitions in SSM (where plain verbs express functionality in a
complex undertaking). This helped to form a practical range of options given
different attributes defined on the basis of the discussion among the participants.
The case demonstrates how soft systems methodology can assist problem
structuring. This can then lead to a range of options, which express the demands
from the stakeholders to the decision about the setting up of a common VTS. In
this case a number of methods from the systemic toolbox were made use of, but the
main experience here was obtained with SSM as a suitable tool to deal with the
complex problem in the first of two sessions.
8.1.2.5 High Speed Rail in Sweden (2008)

The high speed rail case in Sweden made an interesting case in the way that it
helped to make a huge amount of background material operational for the choice
among four different alignments for a new northsouth high speed railway in the
central part of Sweden (close to Norrkping); along this stretch a number of
conflicting issues make the choice of alternative very difficult.
Due to time constraints of the participants the session was only one halfday
meeting. The session had, however, been prepared so the main focus could be set
on exploring the four alternatives on the basis of the background material made
up of a comprehensive examination of the influence of different alignments on
the surroundings. Specifically the following issues were examined (Hiselius et al.
2010):
Impact on urban land-use and landscape

The cultural heritage
Ecology and natural diversity
Recreational life
Impacts on health from noise, vibrations, pollution and severing effects
Demand on natural resources and land-take
103
Risk and safety issues including also transport of dangerous goods

Construction period with consideration of disruption, etc.
At this conference a lot of effort was put into the scoring of alternatives. This
scoring was based on the mentioned background material and the way this was
interpreted and made use of in the pairwise comparisons.
The case demonstrates that it is possible to score and establish agreement about
what appears to be relatively complex criteria such as the impact on urban land-use
and cultural heritage.
8.1.2.6 New Light Rail Service in Malm, Sweden (2009)

The new light rail service case is particularly interesting in the way it addressed the
task of reducing an initial comprehensive long list of criteria. The list appeared
from a brainstorming, with the criteria split into the following subgroups (Hiselius
et al. 2010):
Impacts on the functioning of the rail network (altogether eleven different criteria: Capacity, comfort, costs, relief of road users wear and tear on the streets,
noise and vibrations, alignment, impact on network extension flexibility, climate
impact on operation and other possible disruptions, linking to regional system,
accessibility, ease of users to navigate the system)
Impacts on urban planning (altogether six criteria: Enabling increased density of
housing, coordination between transport and land-use, diversity in townscape,
impact on modal split seen as reduced car traffic, integration of townscape,
systems thinking so services as medical centres, etc. are related to the transport
system)
Impacts on the urban environment (altogether four criteria: Influence on the
aesthetic appearance (the urban picture), development of urban spaces,
enhanced greening, fewer cars)
Urban air: Reduced emissions
Impacts on traffic: Time savings, reduced congestion
Impact on public transport: Increased number of travellers, higher esteem
related to using public transport
Social impacts (altogether four criteria: Social integration, increased accessibility to leisure activities, increased possibility of treating all people equally,
gender aspects such as securing safe transport for women)
Impacts on image (altogether four criteria: Increasing the image of the town as a
public transport town, image value to companies, the image of the town as
being modern and undergoing development with regard to technical solutions
that incorporate the environment, generally increasing the overall town
attractivity)
Political impacts: Making public transport improvement easy to defend politically, making public transport easier to prioritise
104
Altogether 36 criteria are contained in this long list behind the strategic decision
making of this case related to choosing the best new light rail investments. As can
be noted some overlaps occur, which had to be dealt with. Furthermore, the
number of criteria had to be reduced to a practical number that should be less than
ten and at the same time with a minimum loss of important information.
This was taken care of in the decision conference as a number of impacts were
included in a CBA and a number of criteria included in an MCA.
The case demonstrates that it is possible to handle even long lists of criteria and
considerations. An important finding in this decision conference worth noting is
that the ranking according to the CBA turned out later on not to be affected by the
ranking produced by the composite use of CBA and MCA. To some of the participants this was really a surprise, but the facilitators explanation that MCA
incorporation need not alter the ranking of alternatives ended up by the participants to be perceived as a good result providing the interesting information that
core and wider performance may be correlated.
8.1.2.7 Initiatives and Projects Concerning Biking,

Denmark (2009 and 2010)
The following case about prioritising initiatives and projects for promoting biking
in Denmark has several interesting aspects. Characteristic for this case, which is
part of the Danish Parliaments decision to achieve a sustainable transport system,
is that in principle all types of proposals can be made, ranging from new bike
facilities as a cycle path, etc. to proposals about health and biking to be promoted
by a proposed campaign. The number of proposals added up to 133 in 2009 and to
154 in 2010.
To deal with this immense task the mind mapping technique was applied to
identify main criteria and sub-criteria. Furthermore the mind mapping had to be
adapted to the nine different groups to which the proposals were distributed.
However, even within each group of proposals the proposals were quite different
which constituted a challenge with respect to the adaptation of criteria and
sub-criteria from the mind mapping (Beslutningsmodelgruppen 2009). Another
challenge was that CBA could not be made use of, not only due to lack of data but
also due to the lack of knowledge about traffic-economic principles similar to those
available for the appraisal of bypass projects (Trafikministeriet 2003).
The methodology applied was pairwise comparison making use of the results
from the mind mapping. As mentioned nine groups of projects were identified.
Below four of these are listed to exemplify the variety of criteria made use of.
Urban-town projects (altogether four criteria: Influence on modal split, network
aspects, linking to other transport means, perceived risk reduction)
Cycle path projects (altogether four criteria: Network aspects, perceived risk,
prioritising compared to other transport means, savings in time and money)
105
Campaign projects (altogether four criteria: Relevance for cyclists, visibility,

specific potential for promoting biking, change of behaviour with regard to how
cyclists relate to other people in the traffic)
Knowledge and research projects (altogether three criteria: Innovation, the
influence on spreading the message that biking is an attractive transport mode,
the potential of communicating the results of the project to achieve a societal
impact)
Clearly this case demonstrates the possibility of handling a case with only a
relatively vague knowledge of the core performance (examined by using CBA
typically). Thus the pairwise technique was used to enable prioritising of the
proposals based on their wider performance with regard to a number of identified
most important MCA criteria (Barfod 2010).
8.1.2.8 ITS Projects, Denmark (2010)

Promoting projects for implementing intelligent transport systems (ITS) is also,
like the promoting of biking described above, part of the new Danish transport
policy initiated as Green Transport in 2009. Making decisions about ITS projects is similar to the work on initiatives and projects for making biking more
attractive characterised by many-sided effects relating to a set of different goals.
As with biking the ITS work was commenced with a workshop laying the basis
for the further work by mind mapping a long list of criteria, which were then
shuffled and reduced to criteria sets relating to different groups of projects.
However, in this case the latter consisted of two groups only, namely projects
ready for implementation and projects in need of further development to reach a
stage where they are ready for implementation. Below the criteria sets for both the
implementation projects and the developmental projects are shown (Beslutningsmodelgruppen 2010):
Implementation projects (altogether five criteria: Accessibility including
avoidance of congestion, safety, comfort, readiness to be accepted by the users,
visibility and political support)
Developmental projects (altogether eight criteria: Consideration of chosen
platform for ITS, interoperability, IT information safety, adaptability in
upcoming projects, savings with regard to R & D costs in upcoming projects,
reliability in operation, savings in operation, competitiveness with regard to the
IT component foreseen in upcoming projects)
The government funding assigned as part of the Green Transport Policy
consists of annual funds until 2013. Due to a later start of the work on ITS as
compared to biking as a possible explanation, but mainly due to the much more
complex and demanding design work necessitated by ITS projects as compared to
the biking proposals, only a limited number of ITS projects were prioritised in the
106
2010 round. The work on the ITS projects gave a possibility of going into depth
with the individual ITS projects.
As for the biking initiatives and projects also the work on the appraisal of the
ITS proposals had to be done without the traffic-economic knowledge that is
applied with ordinary highway projects such as the bypasses treated above. Even if
efforts to develop such knowledge are given high research and development
priority, it is doubtful whether a sufficient CBA foundation can be established for
these new types of projects. In the terminology of SP we can say that the noneconomic wider performance will continue to play an important role in upcoming
work. Not just in transport planning but also in other types of planning such as, for
example, energy planning with projects competing for funding but with lack of
sufficient data, this points to the need to develop assessment based on multi-criteria
analysis.
In Scandinavia, compared to mainland European countries such as the
Netherlands and France, there has been some reluctance to give up cost-benefit
analysis (CBA) and base transport decision making on multi-criteria analysis
(MCA). Some method proposals have suggested applying CBA as one of the
criteria in an overarching MCA. One prominent example here is the EUNET
methodology developed in a European strategic transport research with a number
of European countries represented in the research team (Tsamboulas et al. 1998).
However, the EUNET approach has not gained acceptance (authors personal
interpretation) because the CBA is transformed away from money into an index
value representing the CBA score.
On this basis the COSIMA (composite model for assessment) has been
developed, which has as its unique feature that the CBA result is maintained all the
time. Some responses to COSIMA have been that the approach is appealing as the
decision-makers can still keep their eyes on the CBA. Even when disagreeing
with the view that CBA information is more valuable for decision support than
MCA information, it is a point taken. What can be learned in this respect is to
objectivise the MCA procedures. In connection with both the biking and ITS
work, a log book documenting each pairwise grading with the arguments behind
has been kept. This means that each and every pairwise comparison can be
reconsidered and changed if new arguments are in favour of this.
As already noted COSIMA is treated in detail in Appendix A based on appraisal
of the Danish bypass alternatives for Hng and Allerd, which are towns on
Zealand located about one hour west and north of Copenhagen respectively. In
addition to COSIMA, based on combining CBA and MCA (Andersen and Petersen
2006; Leleur et al. 2007) another more recent method combining MCA and risk
analysis is treated in detail in Appendix B. The method called SIMDEC is
introduced below and is based on the assumption that decision-makers can handle
CBA information not just as a point estimate as in COSIMA but also as an interval
estimate that results from a risk analysis based on Monte Carlo simulation.
107
8.1.2.9 Fixed Link Between Elsinore and Helsingborg (2010 and 2011)
The case concerning the alternatives for the new fixed link between Elsinore and
Helsingborg is the most recent SP case. The methodology applied is based on the
SIMDEC approach, where robustness of socio-economic feasibility is one of the
decision criteria (Larsen and Skougaard 2010; Leleur et al. 2010).
Altogether four alternatives have been proposed and sketch designs have been
carried out to enter the further planning and political process. In addition to
specific purposes (for car or train only or for both) and to have preliminary
investment cost estimates, the sketched alignments, etc. make it possible to begin
an examination of the attractiveness of each of the alternatives. With a decision in
either 2012 or 2013, the new fixed link can be expected to open around 2020,
which will be a much needed relief of traffic from the current fixed link between
Copenhagen and Malm that opened in 2000.
This case demonstrates the use of risk analysis (RA) belonging to the hard
methods in the systemic toolbox. As appears from the description in Appendix B,
a new concept called certainty graphs can be used to provide the users with CBA
information including information about uncertainty relating to the estimate of
construction costs and to the prognoses made for the future traffic; the latter is also
referred to as the demand estimate. It can be noted that only these two elements are
included in the risk analysis. As mentioned in the appendix they have, however,
been found to be the main causes for large transport investment projects failing to
achieve a feasible socio-economic result in many cases (Priemus et al. 2008). The
type of risk analysis conducted is referred to as feasibility risk analysis (FRA)
(Salling 2008; Salling and Leleur 2009).
The perspective of developing SIMDEC is to have a methodology that can be
based on MCA and RA. A main issue here concerns how certainty graphs representing FRA-information will perform as decision-maker input in a decision
conference. Experience so far is promising based on, among other things, a
decision conference conducted in October 2011.
8.1.2.10 Overview of Cases

Including also the relocation case described in Chap. 7, altogether ten cases have
been described. The relocation case gave a stepwise general introduction to SP
practice, whereas the cases reviewed above have been treated with emphasis on
what has been seen as special for each particular case. Together the work cases
are an important part of the background for the focus in the next section set on
seeking to answer questions about the validity and potential of applying a systemic
approach such as SP.
108
8.2 The Validity and Potential of the SP Framework

As can be seen the SP approach has been developed in the transport sector and
applied on strategic transport problems in Denmark and Sweden. Furthermore
some of the techniques in the toolbox have been applied on projects concerning the
Trans-European Transport Network. This especially concerns the COSIMA
methodology (TetraPlan 2008). As noted in the introduction to this book it aims,
however, at presenting the systemic planning (SP) approach to strategic decision
making as being generally applicable for planning problems relating to complex
strategic choices, i.e., not tied specifically to the transport sector.
At the time of writing steps have also been taken in this direction as SP is being
applied to planning problems in the sector of resources and the environment. With
a rough categorisation we may call problems belonging to the transport sector, the
sector of resources and environment and other public administrative areas for
governance problems, while a problem such as the relocation case treated in
Chap. 7 can be categorised as belonging to what may be termed business problems.
Here I follow a categorisation suggested by Bhushan and Rai (2004), but with the
difference that defence is not singled out as an individual category but is here seen
as belonging to the governance category. It can, however, be noted that interest has
been expressed by Danish military authorities to apply SP for tasks relating to
contingency and emergency planning.
In their magnificent book Strategic Decision Making: Applying the Analytic
Hierarchy Process Bhushan and Rai argue for the need of a formal decision
making framework in the following way:
The complexity of the modern world is a much-acknowledged fact. As the human race
develops, complexity increases. Technology has created various artefacts to relieve us of
manual, routine and time-consuming tasks. The predictable and deterministic world of the
past has been replaced by the uncertain, random and disorderly world of today. Technological advances in multiple fields of human activity have created a planet on which
things happen at electronic speed. Rapidly increasing complexity and information overload have schemed together to drastically reduce the time available for making decisions.
The decision-maker is stressed, overloaded with unsolicited information, has not enough
time to analyse the situation, and yet must make decisions that have high-risk implications
or consequences. What does the decision-maker need? Human decision-making in the
world characterised above needs a quick-response analysis of the situation that somehow
captures the decision-makers intuition, judgement and experience.(Bhushan and Rai
2004, p. 11).
In their book Bhushan and Ray give a convincing description of how the analytic
hierarchy process (AHP) can assist the decision-makers when engaged in evolving
strategic decisions. They take their point of departure in describing what they see as
the generic decision making problem in the following way (Ibid., p. 11):
Studying the situation

Organising multiple criteria
Assessing multiple criteria
Evaluating alternatives on the basis of the assessed criteria
109
Ranking the alternatives

Incorporating the judgements of multiple experts
In many ways the AHP methodology as presented by Bhushan and Rai (2004)
and the SP approach as outlined in this book can be seen to resemble each other as
ways of decision engineering. This is not surprising bearing also in mind that AHP
is actually one of the methods contained in the systemic toolbox. The potential and
validity of SP can, however, be explored based on the way the two approaches
seem to diverge in their approaches both as concerns their premises and their
practical ways of dealing with a complex plan problem. The latter to some extent
is tied up with the adopted views behind AHP and SP. With respect to AHP
Bhushan and Rai make the following observation:
It is the lack of adequate quantitative information which leads to dependence on the
intuition, experience and judgement of knowledgeable persons called experts. (Bhushan
and Rai 2004, p. 13).
One can read this statement as expressing that quantitative information is

preferable to qualitative information, with the latter being related to dependence
on intuition, experience and judgment. The way SP has evolved and has been
presented in this book is, however, contrary to this perception. A major strength
behind SP is the exploration robustness that is obtained by applying a multimethodology approach (Mingers and Gill 1997; Mitchell 2004; Leleur 2008a).
This is one of the points to bear in mind when assessing SP on the basis of the
presented theory, the systemic process and toolbox, and the cases. These individual elements constitute the SP framework for making complex strategic
choices.
8.2.1 The SP Framework

In Chap. 1 the presentation of the framework of systemic planning (SP) was
organised as the following interlinked levels:
Level 1: Coming to grips with complexity by combining different ways of seeing
Level 2: Designing the SP learning process
Level 3: Specifying the SP toolbox
Level 4: Demonstrating SP on a strategic decision making case
The findings made by exploring these levels have led to the following findings
about SP:
I. Quantitative information and qualitative information represent two different ways
of thinking about a problem (Simplicity-paradigm and Complexity-paradigm rooted thinking). Quantitative information should not generally replace qualitative
information. These two types of information should accommodate each other
making room for the best possible way of confronting strategic real-world problems.
110
II.
As complexity is not a surface noise of the real that can be swept away, it
must instead be paid attention to by considering the fact that strategic realworld problems are expressions of open-ended change. In practice this
implies that learning can never be replaced by doing overall modellingbased optimising.
III. To deal with open-ended change and avoid unintended closure as regards
problem understanding it is necessary to apply different paradigms that less
formally but in a more practice-oriented way express different modes of
enquiry (MOEs). Altogether five different MOEs have been identified to
produce what was termed a cognitive billboard having the purpose of
bringing cognitive awareness and alertness to SP typically early on in the
process.
IV. The systemic process can be organised in different ways, however, always
with the purpose of triggering search-learn-debate activities around the
complex plan problem. In principle one person could undertake this (with
debate then becoming reflection) as a desk study but there are obvious reasons
for setting a team covering a wide range of competencies and skills. Furthermore there are good reasons to engage a moderator, coach or facilitator.
The latter is the term used when organising the process as a decision conference (see below).
V. The systemic toolbox has been developed to contain a number of hard and soft
operation research methods that can complement each other. In principle other
methods than the seven hard and seven soft methods presented can be used.
These methods have, however, been found to cover the actual demand for
methodology in the SP cases conducted so far. A not too high number of
different methods and techniques makes it possible for the facilitator and
analyst to acquire and master such a level of knowledge as regards the individual method that overall this will contribute to making the running of the
decision conference smoother.
VI. Ideally defining a number of adequate alternatives and afterwards assessing
these to determine the best choice should be based on a circularity between
scoping and assessment. As scoping and assessment are interdependent, the
identification of a number of relevant alternatives is in practice a processfinding based on the achievement among the participants of what has been
described as involved understanding (see below). Formally untimely scoping
can be expected to lead to suboptimising within the decision space. Such
suboptimisingleading possibly to a reasonably good decision may not later
on be recognised as such.
VII. The decision conference is undertaken as a way of accelerating and structuring learning given a specific complex plan problem. The decision conference in SP is seen as an interaction between a group of participants, the
application of formal decision analysis methods, and interactive IT, typically
large projected screens or smart boards supported by an analyst-operated PC.
When this interaction is working at its best the decision conferences made in
111
connection with the SP cases described above have been successful and have
obtained good evaluation questionnaire responses afterwards by the participants. Currently, however, no fixed recipe exists for organising and optimising SP decision conference sessions.
VIII. Applying SP was in theory seen as unfolding a subworld around the problem.
This makes sense as the subworld maintains a full perspective; at no stage in
the process should it turn into a solely calculative task that can be optimised.
Optimising can be provided in various respects typically as what-if questions behind specific model runs; however, the results obtained in this way are
still to be framed into a wider set of considerations. In practice the latter may as
was the case in the relocation case result in a wish to undertake a reexamination
of some of the issues debated and decided earlier. Even if such reexaminations
are not undertaken, the possibility to do so functions as a way to maintain a
practice-oriented comprehensive perspective.
Comparing SP to AHP very generally there seems to be less belief in SP than in
AHP that the generic decision making problem as described above can satisfactorily be based on one main methodology. The multitude of methods in the SP
toolbox simply indicates this viewpoint. In SP there also seems to be more explicit
concern about avoiding suboptimising and this is paid attention to by setting focus
on what has been described as the scoping-assessment circularity. In practice this
may not always lead to factual reconsideration of the choice alternatives once
again, but the awareness of the problem of untimely closure has in some of the
SP cases led to debates that have paid attention to considering going through the
alternatives once again; thereby the participants are made aware that it might be
relevant to reconsider some of the basic issues in the light of the way the process
has been conducted. Such reconsideration will help them to form an opinion
whether to accept (or to restrain from) ownership of the perception that the best
choice of alternative has been achieved.
Is SP a valid approach? Comparing it with AHP, which today is generally
accepted and tested through numerous cases (Saaty 2001; Bhushan and Rai 2004),
SP can borrow so-to-say validity from AHP as SP can be interpreted as a kind of
corrective and/or supplement to the problem handling in some of the steps in the
generic decision making problem as it was described above.
Practically SP has proved to be a flexible way of dealing with complex
problems; in this respect the underpinning theory was of value with regard to the
outline and design of the systemic process and toolbox. At the same time, however, the theory can be useful in another way, namely by dealing with validity and
potential and hereby more generally the relevance of strategic decision making
approaches such as AHP and SP in a wider context. In the following final section
of the book we will deal with this question of relevance by addressing the
possibilities and limitations of supporting complex strategic choices in a modern,
globalised and complex world.
112
8.3 Complex Strategic Choices in a Wider Perspective

8.3.1 Known and Unknown
It can be stated that strategic decision making in todays world is both necessary
and impossible.
It is necessary because strategic development is based on long-term commitment of resources in both public organisations and private firms and for that matter
also in hybrid types of companies that span the public/private categories.
In economic jargon the making of a strategic long-term choice means resources
foregone for other competing possibilities and initiatives, with such resources not
just measured as investment means but also perceived more broadly as the variety
of strengths and opportunities not possible to pursue now due to the implications of
the strategic choice that has been made.
But at the same time strategic decision making is impossible. In the theory
chapters laying the ground for the presented and demonstrated systemic planning
(SP) practice consisting of process and tools, the orientation was to make strategic
viewing as broad as possible. All the time the company is seen as constructing its
future as a selection among futures presenting themselves as the horizon of possibilities. Due to complexity this horizon is, however, not clearly visible. Among
other things, the visibility is tied to our cognitive capability and even sometimes
with several epistemic modes of enquiry we cannot achieve clear sight or certainty.
In this respect at least detail complexity, dynamic complexity and preference
complexity contribute to blurring our viewing.
What constrains our strategic viewing was in an opening statement to the press
in 2002 formulated very elegantly in the following way by the former United
States Secretary of Defense Donald Rumsfeld:
There are known knowns; there are things we know we know. We also know that there are
known unknowns; that is to say there are some things we do not know.
But there are also unknown unknowns the ones we dont know we dont know.
(Rumsfeld 2002).
The unknown unknowns (UUs) represent the huge challenge of strategic

decision making. The precaution we need to take is whether or not UUs should
affect the way complex strategic choices are prepared. But this we cannot know!
But acting on the basis of the knowledge that has been acquired in the subworld
established around the complex strategic problem, at least the volatility of our
decision may become visible as the viability of the decision is found to be conditioned by a set of uncertainties and contingencies. These can be ascribed to both
local and in the recent decades less local factors, with the latter gaining in influence. This shift in influence is due to the fast changing national and global decision
environments around the decision-producing socio-technical systems. This kind of
systems is in this book used as the common denominator for the rich variety of
113
Problem types Four different configurations and related approaches
Means/ends
Certain
Certain
A: Simplistic
Uncertain
Uncertain
C: Complicated with compromises; in

need of being unfolded with circular
scoping and assessment
B: Complicated with judgments;
D: Complex with judgments and
in need of being unfolded with
compromises and ??; in need of being
circular scoping and assessment
unfolded with circular scoping and
assessment
companies and organisations that are engaged in present-day strategic decision

making.
In Chap. 1, see Table 1.1, different problem types were categorised into four
categories by pairing certain and uncertain means with certain and uncertain ends.
This led to the following four ways of problem solving: Computation (A),
judgement (B), compromise (C) and chaos or inspiration (D). Wrapping up the
book we can present it in the following way by applying some of the main terms
and concepts that have been introduced, see Table 8.2.
What is being offered with the SP approach is a way of dealing with
situations B, C and D. No doubt the D situation is in need of much precaution
with regard to the validity of its outcome, with the double question marks in
Table 8.2 indicating, among other things, the possible influence of the unknown
unknowns.
8.3.1.1 Black Swan Theory

In dealing with the D situation above it becomes relevant to address the Black
Swan theory set out by Nassim Taleb in his magnificent book The Black Swan
(2010). In this book, published in its first edition in 2007, Taleb sees a Black Swan
event in the following way:
What we call a Black Swan (and capitalize it) is an event with the following three
attributes.
First, it is an outlier, as it lies outside the realm of regular expectations, because
nothing in the past can convincingly point to its possibility. Second, it carries an extreme
impact (unlike the bird). Third, in spite of its outlier status, human nature makes us
concoct explanations for its occurrence after the fact, making it explainable and predictable. (Taleb 2010, pp. xxi-xxii).
Taleb has developed his Black Swan theory based on a background of a career
comprising both being a business investment trader and having a university career
as a professor in risk engineering. One can read his book as a major warning about
relying on long-term predictions that expose your decisions to unexpected events
such as extreme market moves. Specifically, he has set out a warning against
114
Talebs Fourth
Four quadrants ranging from extremely safe
Quadrant theory
to Black Swan domain
Domains
Simple payoffs
Complex payoffs
Certain (closed and contained change):

Q1: Extremely safe Q2: (Sort of) safe
Mediocristan
Uncertain (open-ended, possibly radical change): Q3: Safe
Q4: Black Swan domain
Extremistan
Adapted from (Taleb 2010, p. 365)
ending up in what he refers to as the Fourth Quadrant, see Table 8.3 above, where
radical change is possible and with severe consequences. Taleb communicates the
immanent nature of the environment of change by using the expressive labels
Mediocristan and Extremistan.
Taleb describes the quadrants in the following way:
First Quadrant. Simple binary payoffs in Mediocristan: forecasting is safe, life is easy,
models work, everyone should be happy. These situations are, unfortunately, more common in laboratories and games than in real life. We rarely observe these in payoffs in
economic decision making
Second Quadrant. Complex payoffs in Mediocristan: statistical methods may work
satisfactorily, though there are some risks. True, use of Mediocristan models may not be a
panacea There clearly are problems here, but these have been addressed extensively in
the literature
Third Quadrant. Simple payoffs in Extremistan: there is little harm in being wrong,
because the possibility of extreme events does not impact the payoffs. Dont worry too
much about Black Swans.
Fourth Quadrant, the Black Swan Domain. Complex payoffs in Extremistan: that is
where the problem resides; opportunities are present too. We need to avoid prediction of
remote payoffs, though not necessarily ordinary ones (Taleb 2010, pp. 363-365).
An important observation made by Taleb is that positive Black Swans are also a
possibility:
Actually, the Fourth Quadrant has two parts: exposure to positive or negative Black
Swans. I will focus here on the negative one The recommendation is to move from the
Fourth Quadrant into the third one. (Taleb 2010, p. 365).
Talebs advice is that being in Extremistan (where loosely speaking normally

distributed variations are insufficient as opposite to the situation in Mediocristan)
one should concentrate on changing the situation away from the high-risk exposure
that relates to Extremistan events.
The Black Swan theory truly adds to the precaution of dealing with the situation
D in Table 8.2. Should we avoid it then? As Taleb says there are also positive
Black Swans so we need not be fully convinced that this is the final resolution to
behave like that. The answer in our context may depend on how creative we can be
in navigating in complexity towards robustness.
115
8.3.2 Towards Robustness in Strategic Complex Choices

Earlier in this book complexity was introduced by using the concepts of detail
complexity, dynamic complexity and preference complexity relating to means,
path and ends, respectively. When a planning problem complexifies along these
dimensions in the decision space, we therefore generally must foresee a growing
degree of complexity. Certainly this is the case with strategic planning problems
relating to open-ended change. To deal with this situation boundary setting was
introduced based on Midgley (2000).
In SP boundary setting never becomes an explicit activity decided upon itself;
the boundaries set are better perceived as the by-product of the unfolding of the
subworld around the complex problem. As indicated above SP is carried forward
in an actual task with a principal aim of not leading to untimely closure, which
so-to-say is inscribed in the SP approach rationale with the emphasis paid to
ongoing, circular scoping and assessment. As, however, the cases described do
witness: in SP practice a stepwise locking-in has to take place.
Complexity was by instigation of the theories of Morin and Luhmann given
another role than being representative of something that can be successively
removed by more data, more refined models, etc. In Morins view complexity
continues to reappear as it is always profoundly active in the world (and not just a
surface noise of the real) and in Luhmanns view complexity reappears as being
immanent in our unending task concerned with: complexity leading to forced
selection, which leads to outcomes influenced by contingencies and risks, which
again represents new complexity forcing the decision-maker(s) towards new
selection, etc.
Using again viewing metaphorically for our cognitive endeavours (we did that
already with the term epistemic lenses), we may understand complexity in relation
to our viewing capabilities. In a truly physically flat world the viewing can be allencompassing with no hidden places if the observer is duly equipped with a strong
telescope and is placed in a high tower; in a globally-shaped (curved) world the
telescope and tower will no doubt improve our viewing capabilities, but given that
our observations are made at a certain place, some distant places will be invisible
(hidden). Moving to a different place will change the viewing but new places will
now have become invisible.
Seeing this viewing metaphor in relation to a decision space, invisibility remains a
challenge as it will unfold in parallel with the SP unfolding of the subworld that takes
place in relation to the decision situation. The best promises by applying SP seem to be
twofold. One concerns decision awareness: SP is set out as a search-learn-debate
approach and not as a panacea that guarantees the quality of decision making; the
other concerns decision ownership: SP may have a role in providing a team with a
sense of ownership towards the SP outcome by assisting the team members in moving
from being detached to being involved in the decision recommended. Both these
promises are seen as important for the quality of factual complex strategic choices.
116
The idea of SP as an approach to complex strategic decision making is not to

enable decision-makers to make the right decision but to assist the decision
making team in such a way that an informed decision may be taken. In raising the
team awareness and moving the participants towards involved decision making in
a SP process, various principal questions are set out for the team to consider.
Typically their answers have implications for the structuring of the further SP
process ahead. Characteristic of each principal question in a decision conference is
always: could or should this be otherwise? Characteristic of the process is also that
later on, in case of new insights, the possibility exists of going back into the
sequence of considerations and altering the inputs already made. In this way
robustness is built into the process.
At the same time robustness is also associated with the selection of a subset of
methods and techniques from the toolbox. As concerns the soft methods these are
mainly to be seen as explorative and problem structuring, where the hard methods are
performance-oriented (CBA with focus on monetary core performance and MCA
with focus on wider performance) and with scenario, stakeholder and risk analysis to
address performance robustness. But in no way this guarantees what may be termed
an overall robust outcome with regard to the decision to be recommended.
One current development perspective with regard to SP consists of doing
further research with regard to the linking of CBA and MCA. With a maintained
view that CBA must be kept visible the COSIMA technique has turned out to be
very useful as case applications have indicated that decision conference participants can intuitively grasp this technique, while at the same time the responses to
the decision conferences have clearly indicated that the visible CBA result is seen
as a valuable feature. A more technical description of COSIMA is contained in
Appendix A. Still maintaining the CBA visibility (now presented to the decision
conference participants as an interval-probability result instead of as a point
estimate) is behind the SIMDEC approach which is given a technical presentation
in Appendix B. With the SIMDEC approach, risk analysis can be incorporated as
part of an MCA assessment of alternatives based on pairwise comparisons.
Another current development perspective with regard to SP consists of embroadening the experience stemming from cases. The most recent applications described
above have shown that SP can be applied on task less data-rich but more criteria
demanding than conventional appraisal methodology in transport planning. This
gives a background to implement SP also for strategic decision making in other
sectors than the transport sector. In addition to a location decision addressed in the
demo case in Chap. 7 implementation of SP is underway in environmental and energy
planning.
8.3.3 SP from Here?

The development of SP as idea and methodology has taken place mainly within
transport planning by addressing complex problems of various kinds as described
117
in Sect. 8.1. Due, however, to the general nature of the SP framework as described
and validated in Sect. 8.2 there seems to be a potential for a broader use. In this
respect I would like to give attention to the recent development in business orientation and investment strategy known as social return of investment (SROI)
(Ellis 2010; Nicholls et al. 2010). A major recognition as regards SROI and related
ideas in what is generally known as corporate social responsibility (CSR), is that
these business trends will grow in importance to become major drivers in business
and management innovation (Ellis 2010). Specifically, Ellis sees the individual
organisations and companies as part of an ongoing trend seeking to merge their
core activities with CSR activities hereby creating value to both company and
society (Ibid., p. 157). Addressing this as Strategic CSR Ellis perceives this as a
new important business platform that makes it possible to set focus upon necessary
business innovation as well as upon wider sustainability. The latter is seen as both
a societal concern and a concern of the individual organisation and company.
With emphasis in this context on SROI as a way of determining the overall
effectiveness of Strategic CSR action, it can be noted that quite a number of
similarities exist between the ideas expressed in SP and those forwarded by SROI.
One main example is the necessity to address both monetised and non-monetised
issues; another one is the importance of including individual stakeholder views and
paying attention to the whole set of stakeholders involved in a particular strategic
action. In many contexts when addressing particular complex strategic choices on
the basis of an adopted CSR strategic company orientation, SP may be of interest
for decision-makers and analysts. The demo-case in Chap. 7 demonstrates how
management and staff can be dealt with as two stakeholder groups with different
interests in the relocation decision.
As concerns the challenge of strategic decision making as regards uncertainty in
what may collectively be addressed as the general decision environment, uncertainty is not expected to diminish in coming years. The challenge will consist of
strategic choices relating to what Nassim Taleb sees as Mediocristan as well as
Extremistan. No doubt especially the conditions with radical uncertainty characterising Extremistan call for special attention as concerns the formulation of the
strategic decision alternatives and the specific scenarios they are seen to be
embedded in.
The financial crisis, which is not yet over at the time of writing, has hit companies and organisations in ways that should make it relevant for them to generally
view their long-term planning and strategic choices in a complexity context. Doing
this could well pave the way for them to address more specifically the type of
decision aiding they really need to confront a world and a future characterised by
continuously rising uncertainty, where the financial crisis in a greater perspective
is but one of a number of possible precipitating causes. Even without any
knowledge of the financial crisis in 2008 and its continued reverberations, Martin
(2007) gives a compelling view of the hybrid of new and uncertain developments
and challenges across many societal fields. These are set out in both a short time
and a long time perspective in his thought-provoking book The Meaning of the
21st Century: A Vital Blueprint for Ensuring Our Future (Ibid.). One could narrow
118
Martins vision of the future down to the need and possibility of making better
complex strategic choices being the main concern behind the formulation of an SP
approach. In this respect I find in Martins outline of upcoming strategic decision
making challenges on many corporate and societal levels a thorough argumentation for considering more explicitly what was earlier described as the wider, nonmonetary performance of decision alternatives to be put in sufficient balance with
their core, monetary performance. Such an endeavour is behind both the theory
and practice chapters of the SP framework in the previous chapters.
One of the stepping stones towards the process and the methodology proposed
with the SP framework was the cognitive billboard, see Tables 5.1 and 5.2; another
was the determination of the seven hard and the seven soft operational research
methods contained in Table 5.5 and afterwards referred to as the toolbox. Both the
billboard and the toolbox in combination with the systemic learning approach
adapted (referred to as the subworld unfolding) have in their specific contexts of
applications so farrefer back to the case descriptions in Chaps. 7 and 8served
as ways of enhancing the creativity of decision-makers when addressing upcoming
strategic choice situations. With strategic cognition (SC) as a growing research
field per se and the SC research indicating and collecting evidence that cognitive
decision aids can improve decision quality (Narayanan et al. 2011, p. 341) there
seems to be some potential in continuing to develop SP and similar approaches.
Needless to say, both the cognitive billboard and the toolbox of SP will face
modifications in this respect when they are tested in new fields of application.
Hopefully the formulation of SP as set out in this book can inspire readers to
participate in such further development.
Systemic planning (SP) has been validated among other things by comparing it
to an approach to strategic decision making based solely on the multi-criteria
method analytic hierarchy process (AHP). It has been argued that SP has some
advantage because of its more explicit concern of qualitative information and its
use of soft and hard methodology in combination.
A number of SP application cases indicate that participants in decision conferenceswhich are really the integrating core activities of strategic decision
making in accordance with SPfeel comfortable with approaching complex
strategic choices by using SP. At the end of each decision conference (DC) that
has been conducted by the Decision Modelling Group at DTU Transport an
evaluation questionnaire has been distributed to the participants and returned
with generally very positive responses. In a theory context these responses, also
containing constructive suggestions made use of in various ways, show as a
relatively stable overall finding that in the course of SP activities and the concurrent deliberations of the DC participants, detached understanding and
deciding seems to recede and be replaced by involved understanding and
deciding about the complex problem dealt with.
119
SP is considered to have potential for a variety of problems outside the ten

problem cases dealt with in this book. In this respect it is worth mentioning
complex plan problems relating to corporate social responsibility and to sustainability. In both these areas there is a growing awareness of the need to make
decisions based on an assessment of the decision alternatives, where a deliberate balance of the influence from the wider, non-monetary consequences and
the influence from the core, monetary consequences is behind the decision
alternative finally to be recommended.
The following two technical appendices present in a detailed way how the
COSIMA and SIMDEC methodologies from the toolbox can be applied to
balance core and wider impacts. A main point behind their use in SP decision
conferences is that they, when guided by a facilitator and an analyst, appear easy
for the DC participants to relate to and integrate in the group process. Both
COSIMA and SIMDEC are flexible and can be implemented to support complex
strategic choices in areas where decision engineering, decision support systems
and decision conferences are still rather new concepts.
References
Andersen M, Petersen NB (2006) Bedmmelse af omfartsveje i Allerd. M.Sc. thesis, Centre for
Traffic and Transport (CTT), Technical University of Denmark (DTU)
Barfod MB (2010) Structuring and appraising large and complex decision problems using
MCDA. In: Paper presented at the 24th European conference on operational research,
Portugal, July 2010
Barfod MB, Jensen AV, Leleur S (2008) STMA beslutnings-analyse: metoder, proces og
software. In: Carlsson CM (ed) Hllbart transportsystem fr inre och yttre attraktionskraft, EU
interreg project about strategic transport management in the resund region (STM) MAH
Malm, final STM report by Decision Modelling Group, DTU Transport. Technical
University of Denmark (DTU), Denmark
Beslutningsmodelgruppen (2009) Principper og metoder til brug for prioritering i forbindelse
med udmntning af pulje til mere cykeltrafik, DMG Danmarks Tekniske Universitet,
Denmark, Sept 2009
Beslutningsmodelgruppen (2010) Principper og metoder til brug for prioritering i forbindelse
med udmntning af ITS-puljen, DMG DTU Transport. Danmarks Tekniske Universitet,
Denmark, Dec 2010
Bhushan N, Rai K (2004) Strategic decision making: applying the analytic hierarchy process,
Springer series in decision engineering
Drewes Nielsen L, Homann Jespersen P (2003) The use of action research methods in scenario
construction. Seville workshop, Institute for Prospective TechnologiesEU Joint Research
Centres (IPTS), May 2003
DTU (2007) Dokumentationsmodel for TGB Vurderingsmodel (Trafikplan for Grnland:
Beslutningsredskab), Center for Trafik og Transport (CTT). Danmarks Tekniske Universitet,
Denmark
Ellis T (2010) The new pioneers: sustainable business success through social innovation and
social entrepreneurship. Wiley, New York
Hiselius LW, Barfod MB, Leleur S, Jeppesen, SL, Jensen AV, Hjalte K (2010) Helhetsorienterad
utvrdering av kollektivtrafiktgrder, Bulletin 246, Institutionen fr Teknik och Samhlle,
LTH, Lunds Universitet
120
Larsen, LA, Skougaard, BZ (2010) Vurdering af alternativer for en fast forbindelse HelsingrHelsingborg. M.Sc. thesis, DTU Transport. Danmarks Tekniske Universitet, Denmark
Leleur S (2008a) Systems science and complexity: some proposals for future development, Sys
Leleur S, Kronbak J, Rehfeld C (2000) The resund fixed link: evaluation issues and
development of new methodology, TRANS-TALK, ICCR, Vienna, EU 5th framework
programme, thematic network meeting, Brussels, Nov 2000
Leleur S, Holvad T, Jensen AV, Salling KB (2004a) Development of the CLG-DSS evaluation
model, midterm report presenting the CTT contribution to task 9 evaluation methodology,
Centre for Logistics and Goods Transport (CLG) at the Centre for Traffic and Transport
(CTT). Technical University of Denmark, Denmark
Leleur S, Jensen AV, Salling KB (2004b) COSIMASoftware manual, version June 2004,
Centre for Traffic and Transport (CTT). Technical University of Denmark, Denmark
Leleur S, Jensen AV, Salling KB (2004c) Modelling decision support and uncertainty for large
transport infrastructure projects: the CLG-DSS model of the resund fixed link. In: Antunes
CH, Dias LC (eds) Proceedings of the 15th mini-EURO conference in managing uncertainty
in decision support modelsMUDSM 2004, Coimbra, Portugal. INESC-Coimbra, Sept 2004
Leleur S, Petersen NB, Barfod MB (2007) The COSIMA approach to transport decision making:
combining cost-benefit and multi-criteria analysis for comprehensive project appraisal. In:
Kim K-S (Ed) Proceedings of the Korean Development Institute & World Bank Conference,
Seoul, May 2007. Improving public investment management for large-scale government
projects: focusing on the feasibility studies, Publisher KDI, pp 100122
Leleur S, Larsen AL, Skougaard BZ (2010) Strategic transport decision making: the SIMDEC
approach based on risk simulation and multi-criteria analysis, EUROSIS. In: Proceedings of
the Asian simulation technology conference ASTEC 2010 conference, Shanghai, March 2010
Martin J (2007) The meaning of the 21st century: a vital blueprint for ensuring our future.
Riverhead Books, Penquin Books, New York
Midgley G (2000) Systemic intervention: philosophy, methodology and practice. Kluwer
Academic, Kluwer
Mingers J, Gill A (1997) Multimethodology. Wiley, New York
Mitchell SD (2004) Why integrative pluralism? Emerg Complex Organ 6(12):8191
Mortensen, M, Andersen, MB (2007) Sammenfattende vurdering af luftfartsalternativer ved
Nuukmed srligt fokus p multi-kriterie analyse. M.Sc. thesis, Center for Trafik og
Transport (CTT). Danmarks Tekniske Universitet, Denmark
Narayanan VK, Zane LJ, Kemmerer B (2011) The cognitive perspective in strategy: an
integrative review. J Manag 37(1): 305351 (SAGE journals)
Nicholls J, Lawlor E, Neitzert E, Goodspeed T (2010) A guide to social return on investment, the
SROI network in association with the Scottish Government
Priemus H, Flyvbjerg B, van Wee B (2008) Decision-making on mega-projects: cost-benefit
analysis, planning and innovation. Edward Elgar Publishing, Cheltenham
Rumsfeld D (2002) http://en.wikipedia.org/wiki/There_are_known_knowns. Accessed 1 July
2011
Rnnest AK, Ohm A, Leleur S (1997) The resund fixed linkthe conflicts and the players,
WP7 case report, TEN-ASSESS research project, ICCR, Vienna, EU 4th framework
programme
Saaty TL (2001). Decision making for leaders: the analytic hierarchy process for decisions in a
complex world. RWS Publications, Pittsburgh
Salling KB (2008) Assessment of transport projects: risk analysis and decision support. Ph.D.
Dissertation, Department of Transport, Technical University of Denmark, Denmark
Salling KB, Leleur S (2009) Modelling of transport project uncertainties: risk assessment and
scenario analysis. In: Proceedings of the modelling and applied simulation09 conference.
Puerto de la Cruz, Tenerife
References
121
Taleb N (2010) The black swan: the impact of the highly improbable. Random House, New York
TetraPlan (2008) Traffic flow: scenario, traffic forecast and analysis of traffic on the TransEuropean Transport Network (TEN-T), TEN-CONNECT study, Final report for the EU
Commission, Dec 2008
Trafikministeriet (2003) Manual for samfundskonomisk analyse, juni 2003, Copenhagen
Tsamboulas D, Beuthe M, Grant-Muller S, Leleur S, Nellthorp J, Panou K, Pearman A, Rehfeld C
(1998) Innovations in decision analysis, deliverable D10, EUNET research project, EU 4th
framework programme
Appendix A
COSIMA
Overview
This appendix demonstrates the COSIMA approach that is one of the hard methods
in the seven hard and seven soft (2 9 7) methodologies in the SP toolbox, see
Table 5.5. The idea behind the composite model for assessment (COSIMA) is to
link cost-benefit analysis (CBA) with multi-criteria analysis (MCA). Thus CBA
provides an absolute, general assessment related to each alternatives core performance, whereas MCA provides a relative, context-dependent assessment related to each alternatives wider performance as set against the other alternatives
under examination.
COSIMA can be adapted to very different types of selection problems, where
criteria are available both in monetary and non-monetary terms. The Decision
Modelling Group (DMG) at DTU Transport at the Technical University of
Denmark has developed software that makes it possible to customise evaluation
models for a specific study.
Appendix A is an updated version of (Leleur et al. 2007) supplemented with
technical notes based on DMG (2010).
A.1 Purpose, Background and Outline

Project appraisal is the process of comparing the virtues and deficiencies of a
project. The task is to determine the consequences of a project and to apply this
knowledge to support decision making. It is obvious that a project is only feasible
if the virtues compensate for the deficiencies and that the best project is the one
with the largest net gain. The challenge is to find a method to describe and measure
the effects or criteria and to find a rational and trustworthy method to compare and
assess the criteria. However, not all effects can be treated in the same way. While it
is possible to estimate the quantity of time savings, for instance, and assign a
monetary value to this, aspects such as impacts on nature or general societal
impacts cannot easily be assigned a monetary value or perhaps even be quantified.
DOI: 10.1007/978-1-4471-2491-7, Springer-Verlag London 2012
123
124
Appendix A: COSIMA
Many countries either use or have used CBA for transport decision making.
However, this method includes only impacts that can be valued monetarily. The
fundamental idea behind composite modelling assessment (COSIMA) is to extend
conventional cost-benefit analysis (CBA) into a more comprehensive type of
analysisas often demanded by decision-makers (DM)by including missing
decision criteria of relevance for the actual appraisal task. The missing criteria
often address issues that have been difficult to assess by the conventional CBA but
which hold a potential of improving the actual decision support from the appraisal
if treated properly. This is the purpose of COSIMA, where the added criteria will
be referred to as the multi-criteria analysis (MCA) part of the COSIMA analysis.
The COSIMA method will be described in detail below and thereafter
demonstrated on two cases to show the features of the method and the possibilities
it offers. The two cases concern an examination of alternatives for a by-pass road
around the Danish town of Hng and for a new ring road in conjunction with a new
residential area in the town of Allerd. Finally, a discussion of COSIMA compared
to the CBA and MCA methods is undertaken, and the methods are compared with
focus on the following three issues (Andersen and Petersen 2006):
Comprehensiveness: As previously mentioned, not all effects can be treated in the
same way. This creates a challenge for the methods as they have to be able to
include all important effects. The comprehensiveness issue describes how well the
methods succeed in doing this.
Effectiveness: The effectiveness issue describes how easily a final choice can be
made on the basis of the result of the analysis. An analysis must be able to consider
all important effects but also to make the results usable for the decision-makers.
Transparency: Transparency is important with regard to transportation decision
methods because the choices made by the decision-makers must be understood and
accepted by the public. This does not necessarily mean that there must be complete
consensus about the choices made but that a decision which is difficult to
understand for the public is more likely to face opposition and thereby possibly
entail expensive delays or rejection.
A.2 The COSIMA Methodology

The examination of project feasibility in transport infrastructure planning should
be based on a relevant set of impacts (or effects or project consequences), which
depends on the type and size of the project. For some of these impacts such as time
savings, vehicle operating costs, safety, etc., valid assessment knowledge exists so
that the impacts can be included in a cost-benefit analysis, while other impact types
such as urban planning, driver convenience, network accessibility, etc. do not
qualify in this respect. Comprehensive EU transport studies like EURET
(19911995) and EUNET (19961999) have dealt with these issues
(Tsamboulas et al. 1998; Leleur 2000), and more recently they have been
given much attention in EU transport planning and assessment research as reported
Appendix A: COSIMA
125
in the scientific forum project TRANSFORUM (20052007), see www.

transforum-eu.net.
The COSIMA model aims at examining a project where a mix of CBA and nonCBA effects has been found relevant to include in the concrete appraisal study.
The structure and content of COSIMA are presented below in overview (CBA
impacts refer to effects, where pricing manuals and procedures exist, and MCA
impacts refer to remaining non-CBA effects seen as important for the appraisal
task but less known than the CBA impacts).
The first task is to determine the relevant CBA impacts for the concrete
appraisal study. Most often a standard method for the CBA is used. For Danish
appraisals the CBA calculation is described in the Manual for Socio-Economic
Analysis (Trafikministeriet 2003). Benefits and costs in the CBA are calculated
before the COSIMA procedure is begun, and they do not change during the
COSIMA calculations. This modelling feature will probably be considered
important by people accustomed to applying CBA for decision making.
The next task is to determine the MCA impacts of relevance. Where possible
these should be measured in some appropriate type of quantitative unit; for
regional economic development one example could be the number of new jobs
generated; for improved network accessibility another could be the gain in
potential contact hours for specified trip types, for details see (Kronbak 1998).
Some effects cannot, however, be measured quantitatively and must thus be
described by judgement, for example using a -5, .., 0, .., +5 point scale. Another
possibilitywith due caution to be addressed lateris the application of the
analytic hierarchy process (AHP) method, in which the alternatives under each
effect are judged by the possible pairwise comparisons to assess their relative
performance, thereby producing a score for each alternative (Belton and Stewart
2002, pp. 151159).
The quantitative units and the point and AHP scores are then translated into a
final rating or score to make use of in the COSIMA method. The COSIMA method
follows the simple multi-attribute rating technique (SMART) or SMART
exploiting ranks (SMARTER) method when assigning rating values or scores to
the MCA effects. SMART assigns ratings from 0 to 100 to the effects by using
value functions that describe how well the alternatives perform within each effect.
The SMART method (von Winterfeldt and Edwards 1986, pp. 278287) has been
developed into the SMARTER method (Goodwin and Wright 2010, pp. 6366) in
which all the value functions applied are linear, meaning that the slope
representing the actual rating-value change is constant along the curve, see
Fig. A.1. If it is possible to assess an effect quantitatively, the value function gives
the rating for each alternative directly from the actual quantity, but other units such
as the formulated point scale values or AHP scores can also be used to assign the
value function rating.
A main principle in COSIMA is that examined alternatives are assessed both
absolutely and relatively. Thus CBA provides an absolute, general assessment
related to each alternatives core performance, whereas MCA provides a relative,
context-dependent assessment related to each alternatives performance as
126
Appendix A: COSIMA
Fig. A.1 Non-linear and linear value functions. The latter are assumed in SMARTER
set against the other alternatives in the examination. For the MCA criteria a setspecific (local) scale assigns the value 0 to the worst performing alternative and the
value 100 to the best performing alternative. The rest of the alternatives are then
rated by relating them to these alternatives (Belton and Stewart 2002,
pp. 121122). With application of point scales and/or AHP, dependence on the
actual set of alternatives is introduced. In practice this can be dealt with in a
satisfactory way within the model calibration and as part of the COSIMA decision
conference described later. Hereafter three stages I, II and III remain:
Stage I. With the CBA and MCA effects specified, the so-called anchoring
part of the COSIMA model formulation can take place, which concerns
determining the importance of the MCA effects against the CBA effects, i.e. the
overall MCA versus CBA trade-off, and for the MCA effects among each other,
i.e. the determination of MCA criteria weights. With regard to the latter, several
MCA techniques can be made use of: direct weights, pairwise comparisons, swing
weights, etc. (Ibid., pp. 134143, pp. 157159). To ease the assignment of criteria
weights for the MCA effects that can represent the actual DM preferences, the rank
order centroid (ROC) or the more recent rank order distribution (ROD) weighting
technique are also applicable. The determination of rank order weights is based on
the assumption that weights assigned by the decision-makersby simply ranking
the actual MCA criteriacan be derived by using specified probability density
functions (Roberts and Goodwin 2002; Goodwin and Wright 2010, pp. 6366).
The choice of relevant MCA effects and the assignment of weights to these effects
will usually be determined during a number of decision conference sessions, where
both DM and decision analysts take part, see the description later.
Stage II. After the MCA effects and their assigned weights have been agreed
upon, COSIMA can be run. As previously mentioned, COSIMA includes the MCA
effects or criteria along with those usually treated in a CBA, thereby calculating a
total value (TV) in monetary units for alternative Ak obtained by spending the
investment cost Ck:
TVAk = CBAAk + MCAAk
A:1
Appendix A: COSIMA
127
The formulation of COSIMA introduced by (A.1) resembles CBA, but the

assessment principles used in the MCA part, generally based on the involvement
of DM, are not used in CBA and justify the denomination of COSIMA as an MCA
(ECMT 1981, pp. 16, 23). It can be noted on the basis of (A.1) that in a situation
where the investment in Ak (equal to the investment cost Ck) is not feasible seen
from a CBA point of view, i.e. CBA(Ak) \ Ck, the investment may be justified by
the wider COSIMA examination if TV(Ak) [ Ck. If examined as a total rate of
return (TRR), the latter can be expressed as TRR(Ak) [ 1, see (A.2) for a
COSIMA examination comprising I CBA effects and J MCA criteria.
"
#!
I
J
X
X

TVAk 1
TRRAk

VCBA Xik a
w j VMCA Xjk
Ck
Ck
i1
j1
A:2
with:
VCBA(Xik): Value in monetary units for CBA effect i for alternative k for
altogether I CBA effects. During model calibration it is kept fixed as bik.
VMCA(Xjk): Value function rating for MCA criterion j for alternative k for
altogether J MCA criteria. During model calibration it is transformed into a
monetary value bjk.
a: Calibration factor that expresses the specific models trade-off between the
CBA and the MCA part. It should be observed that the CBA calculation remains
unchanged, but that different values of a will change the influence of the MCA
on the TRR value. The value of a = a(MCA%) is set by specifying
MCA% = 100Rj(Bj)/[(Ri(Bi) + Rj(Bj)], where Aj denominates a subset of the
k = 1..K alternatives Ak (with this subset selected for calibration, see the
Technical Notes accompanying Appendix A below) and Bi = Rj[K(bij) and Bj =
Rj[K(bjj) enter as the value elements for the individual effects i and criteria
j summed over the j alternatives; thus Ri(Bi) and Rj(Bj) concern row summations
over the I CBA effects and the J MCA criteria and Bi and Bj the results of bik and
bjk column summations over the alternatives, where some if not all are selected
for the model calibration.
w(j): Weight that expresses the influence of criterion j.
The general COSIMA principles are presented by (A.1) and (A.2). It can be
seen that with sufficient information about the MCA part, (A.2) can be specified
into a CBA-like calculation. This will be the situation when, for example,
a conventional CBA is carried out, and it is afterwards (1) supplemented with
some extra criteria which can be specified fully by impact models that (2) can
determine net effects which (3) can be given satisfactory unit prices similar to the
assessment in the CBA part. Most often, however, this will not be possible,
because usually the MCA part will be less known than the CBA part. In fact the
purpose of COSIMA is to handle such a situation. In modelling terms, this can be
done by the determination of appropriate values for a and w(j) for the J MCA
criteria and by the determination of appropriate value function ratings VMCA(Xjk).
128
Appendix A: COSIMA
The latter supplement the determination of VCBA(Xik) that can be derived from a
CBA manual relevant for the actual assessment case.
Stage III. At this stage COSIMA is run for all the project alternatives. Then the
model inputs and the related assessment questions are addressed on the basis of
the results obtained and the assumptions behind them, and a new exchange
with the decision-makers (DM) is carried out with two principal possibilities
available now. The study may simply end here if the DM are confident about the
model outcome, or the DM may want to go back into the process and re-address
some of the previous model settings to shed light on some issues that have caught
their attention.
One important characteristic of the COSIMA approach is that the model is more
or less customised to the specific appraisal case. It should be observed, however,
that the assessment result produced is given as total rate of return (TRR) values
stemming from an objective CBA part and a more subjective MCA part, where
the CBA part represents a result in its own right, i.e. without the MCA add-on
which actually only provides the DM with some extra discriminatory
information. The CBA-like way applied in COSIMA to present both the CBA
and MCA assessment information may appeal to decision-makers who want the
possibility of refined analysis using all the available information but kept in a
simple and straightforward way. To illustrate this, the decision-makers may wish
to use only the benefit-cost rate (B/C) part of the TRR value from a base case
scenario without any further analysisthis would in fact be a conventional
analysisor they may like to inspect some or all the TRR values and their
composition as produced in a number of what-if scenarios.
Figure A.2 shows the locations of the two case examples presented in this
appendix. The first case concerns the assessment of seven by-pass alternatives for
relieving the town of Hng in a rural area in western Zealand of through traffic.
The second case deals with four alternatives for a new ring road around the town
of Allerd situated in an urban area north of Copenhagen in northern Zealand.
Both case examples are representative of a typical appraisal task in
infrastructure investment planning: a mix of hard (CBA) and soft (MCA) issues
that are relevant for the decision-makers to take into consideration when selecting
the most attractive alternative.
A.2.1 Case Example 1: COSIMA with a Point Scale

The purpose of the following example is to illustrate the practical content of
COSIMA when applying point scales for the MCA part. The case examines seven
alternatives: a short alternative 1 west of Hng and alternatives 2 and 3 as less
narrow by-pass solutions; a short alternative 4 east of Hng and alternatives 5 and
6 as less narrow ones; a final alternative 7 is a combination of alternatives 4 and 5.
It has been found relevant to add altogether three MCA impacts. This means
that the alternatives are examined on the basis of the seven Danish standard CBA
criteria (travelling time, vehicle operating costs, accidents, maintenance costs,
Appendix A: COSIMA
129
Fig. A.2 The two case towns

in Zealand, Denmark (from
Google Maps)
noise, air pollution and severance and perceived risk) and the following three
MCA criteria: network accessibility, urban planning and landscape.
By use of the methodology from the Danish Road Directorate, first-year
benefits (FYB) have been calculated for the seven alternatives (Leleur 2000). This
information has been put together with point scores for the three MCA criteria,
where the point scores are determined by thorough examination of the alternatives
based on a rating protocol. Hereby the project-effect matrix in Table A.1 has been
formulated (Steffensen and Testmann 2000).
At this stage the next step is anchoring the MCA criteria. The three MCA
effects are assigned a value describing their performance on a scale from -5 to +5,
where +5 is best.
The scores are then translated into ratings between 0 and 100, see Table A.2,
using a linear, local value function (Belton and Stewart 2002, pp. 121122).
The MCA impacts must then be assigned weights to be used in the COSIMA
analysis. As the ROD weights are used, the effects need only be ranked by the DM
with the weights themselves being predetermined according to the ROD weight
principles (Roberts and Goodwin 2002). Any ranking could have been used, but it
is assumed that the decision-makers have agreed on the ranking in Table A.3 and
hereby indirectly on the shown weights.
The ratings and weights of the MCA effects are entered into the COSIMA
software along with the normal CBA input for each alternative. It should be noted
that traffic forecasts are also included and that the COSIMA results are based on
discounting the values for a 30-year service-period back to the opening year.
130
Appendix A: COSIMA
Table A.1 The three MCA impacts together with the investment cost and first-year benefits
(FYB) from the CBA methodology
Alternatives
Cost in m DKK
FYB in k DKK
Network
Urban
Landscape
accessibility
planning
Alternative
Alternative
Alternative
Alternative
Alternative
Alternative
Alternative
1
2
3
4
5
6
7
16.7
15.3
16.9
18.0
17.2
20.8
19.9
1,310
790
350
2,900
2,460
1,790
2,330
+2
+1
-1
+4
+3
-3
+3
+1
+1
+3
-2
+2
+4
+1
-2
+4
+2
+2
+3
0
+3
Table A.2 The three MCA impacts rated using the point scale method
Alternatives
Cost in
FYB in k DKK
Network
Urban
m DKK
accessibility
planning
Landscape
Alternative
Alternative
Alternative
Alternative
Alternative
Alternative
Alternative
0
100
67
67
83
33
83
1
2
3
4
5
6
7
16.7
15.3
16.9
18.0
17.2
20.8
19.9
Table A.3 ROD weights

w(j) for the MCA effects
1,310
790
350
2,900
2,460
1,790
2,330
71
57
29
100
86
0
86
50
50
83
0
67
100
50
Effect ranking
w(j)
1. Network accessibility
2. Urban planning
3. Landscape
0.52
0.33
0.15
Afterwards a is determined so the CBA and MCA parts of the analysis are traded
off in a way found suitable by the decision-makers. In this case an MCA% = 50
split is used. The programme then calibrates the model so that the MCA fraction of
the total benefits equals the split specified by the user. The case example 1 results
are shown in Table A.4.
Figure A.3 indicates how the total rate of return (TRR) values consist of both
the CBA and MCA impact contributions for all the alternatives.
It is clearly seen that alternative 5 scores higher than the other alternatives due
to a reasonably good performance with regard to the B/C-rate and the MCA
criteria. The MCA criteria are assessed by trade-off implied unit prices, set against
the costs of the individual alternative and added to the CBA part; in principle only
the latter is a monetary return of the investment (the costs of the alternative) and
the MCA rate solely a value expression of further information to make it possible
to discriminate between the alternatives.
Appendix A: COSIMA
131
Table A.4 TRR values with MCA% = 50

Alternatives
B/C
Network
accessibility
Urban
planning
Landscape
TRR values
Alternative
Alternative
Alternative
Alternative
Alternative
Alternative
Alternative
0.58
0.64
0.96
0.00
0.76
0.94
0.49
0.00
0.60
0.36
0.34
0.45
0.15
0.39
3.59
3.51
2.29
5.49
5.81
2.91
4.71
1
2
3
4
5
6
7
1.66
1.09
0.44
3.40
3.03
1.82
2.47
1.35
1.18
0.53
1.75
1.57
0.00
1.36
Fig. A.3 TRR values with MCA% = 50
An interesting COSIMA feature is that it is possible to analyse how the results

change as the relative weight (MCA%) assigned to the MCA part changes, which
reveals whether more than one alternative should be considered. In Fig. A.4 the
results of such an analysis are indicated.
Very high percentages assigned to the MCA are not included in the graph as the
TRR values rise to very high numbers, thus making the changes in the rest of the
graph difficult to follow. However, it can be argued that if such a high importance
is given to the MCA, it might be better to use another appraisal method such as the
pure MCA.
Figure A.4 clearly shows that two alternatives can be considered as the most
attractive ones. Alternative 4 is best when a low weight trade-off is assigned to the
MCA, but if the MCA is weighted higher than 30%, alternative 5 obtains the
highest TRR value. The graph thus does not depict a single answer as to which
alternative is best, but it provides the decision-makers with an overview of which
132
Appendix A: COSIMA
10
9
Alternative 1
Alternative 2
Alternative 3
Alternative 4
Alternative 5
Alternative 6
Alternative 7
TRR
6
5
4
3
2
1
0
0%
10%
20%
30%
40%
50%
60%
70%
MCA %
Fig. A.4 TRR values as a function of the MCA%
alternatives to consider, and it can also help sort out the lowest scoring alternatives
in the entire range.
A.2.2 Case Example 2: COSIMA with AHP Impacts

The second case deals with a new ring road around the Danish town of Allerd,
which is situated north of Copenhagen, see Fig. A.2. The purpose of this example
is to illustrate the use of the AHP method for determining the MCA ratings (Belton
and Stewart 2002, pp. 151159). The new ring road is constructed in conjunction
with a new residential area, and the case examines four alternatives: a short
alternative 1 east of the new residential area, a long alternative 4 west of the area,
which entails the crossing of a railroad, and furthermore two in-between
alternatives 2 and 3 located west and east of the new residential area respectively.
The CBA was conducted following the Danish manual for appraisal
(Trafikministeriet 2003), and afterwards this analysis was extended with the
following three MCA criteria:
Accessibility (assessment of accessibility to the new residential area)
Local land use (future plans for local land use)
Regional network (improvement of the regional road network)
The ratings of the impacts are conducted by first scoring the alternatives under
each MCA criterion by using the AHP method. These scores are then transformed
into a linear value function applying generally a local scale between 0 and 100,
Appendix A: COSIMA
Table A.5 Ratings assigned
to the three MCA effects by
using the AHP method
Table A.6 ROD weights

w(j) for the MCA effects
Table A.7 TRR values with

MCA% = 50
133
Alternatives
Cost in
m DKK
Accessibility
Local
land use
Regional
network
Alternative
Alternative
Alternative
Alternative
19.2
54.9
48.6
194.9
0
46
19
100
0
100
7
100
0
23
3
100
1
2
3
4
Effect ranking
w(j)
1. Accessibility
2. Local land use
3. Regional network
0.52
0.33
0.15
Alternatives
B/C
Regional Accessibility Local

TRR
network
land use
Alternative
Alternative
Alternative
Alternative
2.03
1.05
1.52
0.50
0.00
0.10
0.01
0.13
1
2
3
4
0.00
0.66
0.31
0.43
0.00
0.89
0.07
0.27
2.03
2.70
1.91
1.33
see Table A.5. Note that alternative 1 is rated as the lowest for all the MCA effects,
while alternative 4 is rated best. It can be noticed that the AHP assessment depends
on the actual alternatives; this may be more pronounced than it could be observed
in the previous case example 1 applying the formulated point scale. Although such
dependencies can be seen as less attractive from a theoretical viewpoint, this
method aspect is due to the more subjective nature of MCA, see the earlier
discussion of the CBA and MCA part of COSIMA. The various theoretical and
practical issues relating to the application of AHP are treated thoroughly by
(Belton and Stewart 2002).
The MCA effects are ranked according to importance and weighted using the
ROD weights as seen in Table A.6. As in case example 1, any ranking could have
been used.
The CBA is based on the Manual for Socio-Economic Analysis
(Trafikministeriet 2003). The TRR values for COSIMA with 50% weight on the
MCA are seen in Table A.7.
Table A.7 and Fig. A.5 indicate how the total TRR values are made up by the
CBA and MCA effects.
Alternative 1 only has a contribution from the CBA effects, but it still scores
higher than alternatives 3 and 4. Due to a large contribution from the MCA effects,
alternative 2 has a better TRR than the other alternatives.
Figure A.6 shows that alternatives 1 and 2 are possible contenders to
finally being selected as the most attractive choice. For a low percentage
134
Appendix A: COSIMA
Fig. A.5 TRR values with MCA% = 50

10
9
8
Alternative 1
Alternative 2
Alternative 3
Alternative 4
TRR
6
5
4
3
2
1
0
0%
10%
20%
30%
40%
50%
60%
70%
80%
MCA %
Fig. A.6 TRR values as a function of the MCA%
assigned to the MCA, alternative 1 scores highest, while alternative 2 is the most
attractive alternative for a relatively high percentage assigned to the MCA.
Alternative 4 is the least attractive one for all trade-off percentages except for the
MCA% equal to 70 or 80.
The results are presented to the decision-makers, who are then able to make an
informed decision based on both the usual CBA and the COSIMA approach with
varying percentages assigned to the trade-off between CBA and MCA.
Appendix A: COSIMA
135
The MCA% values that ought to be applied in the concrete appraisal task
depend on the decision-makers and their interpretation of which amount and type
of factors or criteria they want to influence their decision. Practical experience so
far points to MCA% values in the range of 1050. Furthermore, it seems that high
MCA% values are most likely to be adopted when appraising larger and more
complex transport infrastructure projects.
A.3 COSIMA Decision Conference

As demonstrated by the results of the cases, the COSIMA language is rather
CBA-like although COSIMA seeks to go beyond CBA in making appraisal studies
more comprehensive. Thus maintaining the CBA language, which means among
other things expressing results as rates of return, etc., is seen as an important
feature of COSIMA if this approach to wider appraisal is to be accepted by CBA
users such as civil servants, business analysts and others.
With the focus so far on methodology and assessment principles, the description below concerns the process of interacting with the DM, see (Goodwin and
Wright 2010, pp. 317319). This will be referred to as the decision conference
component of COSIMA. It will be described how decision-makers can be involved
in designing the specific COSIMA model, which by intervention of the analysts is
to be set up to address their specific appraisal task. The description of the
COSIMA decision conference takes as its point of departure a set of alternatives
that has already been examined by an accepted CBA that may be available, for
example by use of a national manual (NM) or a methodology backed by an international institution such as a lending bank, etc. Thus as a first study result benefitcost rates (BCRs) have been obtained, and the decision-makers (DM) can be asked
the first question (Q1) by the analysts/conference facilitators (CF) as follows:
Q1: Through CBA we have found that alternative k is the most attractive one.
Do you agree with this?
The situation may now beand that would often be the case with more
standardised appraisal tasksthat DM agree and feel comfortable with selecting
alternative k for implementation. This indicates that the task has been treated
satisfactorily and there is no need to make the appraisal analysis more
comprehensive.
On the other handand that would often be the case with non-standardised,
more complex problemsDM do not agree with the alternative selected by use of
CBA solely. There is more to the problem than that. In brief DM feel
uncomfortable as something is missing in the appraisal. In this situation CF can
proceed with the question below.
Q2: Can you explain and put words on what you see is missing in the appraisal?
More specifically, can you formulate some criteria that express this?
DM will then, supported by a dialogue with CF, produce a list of extra
criteria. A brainstorming session can help establish a first listing, which can be
scrutinised in various ways so criteria of minor importance can be left out.
136
Appendix A: COSIMA
Technically, the CF will as part of this scrutinisation address issues of clarity,

overlapping, orthogonality, etc. Hereafter assisted by the analysts using either
quantitative measurement, point scores or AHP, scores on the new criteriathe
MCA criteriawill be determined and expressed as value functions and then
presented by CF to DM with the next question below.
Q3: Do you thinkwhen you take a closer look at the alternatives and the new
criteriathat the different alternatives have been rated in a satisfactory way?
This, of course, may lead to some adjustments but the expected end result at this
stage is that the rating of the alternatives is accepted. Then follows a question
about the importance of the new criteria when compared relatively.
Q4: Is it possible for you to formulate and agree upon a ranking of the MCA
criteria?
Depending on the problem type and the different interests represented by DM, it
will be possible to produce one or more relevant rankings of the criteria. COSIMA
rankingsinstead of direct weightingare used to reveal and express DM
preferences at this stage. With more than one ranking formulated by DM, each of
these can be used in the continued process as an expression of some particular
strategy representing, for example, a specific viewpoint or coalition of interests. In
the following it has been assumed that it has been possible to agree upon one
ranking, which leads CF to the next question.
Q5: How do you think the CBA part (the monetary issues) and the MCA part
(the non-monetary issues) should influence the appraisal?By use of percentages
adding to 100% could you express the relative influence of the CBA part versus the
MCA part?
As with Q4 different answers may be obtained. This stage of the process can be
supported simply by testing the suggested different percentages. Furthermore, the
specific percentage at which one alternative gives way to another as the best one is
information of high interest for DM. This situation may also be used to examine
how stable or robust the original top alternative based on CBA solely is when
competing with the other alternatives based on a broader set of criteria.
The idea of the COSIMA decision conference is to obtain the necessary DM
inputs to the COSIMA methodology in a straightforward, non-technical way.
At all stages it is possible to return to and answer a previous question again in case
the DM would like to do so. It is important that the process is supported by a
software system that makes it possible to incorporate the implications of changes
in input more or less instantly on a large screen in the conference room. The
COSIMA software used in this respect is described below.
A.3.1 COSIMA Software

The COSIMA software is developed in Excel with an emphasis on flexibility and
adaptability. The flexibility makes it possible to switch, for example, between
different rating and weighting techniques, and the adaptability makes it possible to
provide a customised decision model (CDM) for the task at hand. An application
Appendix A: COSIMA
137
example that was finished in the middle of 2007 is a COSIMA version (TGB) for
the planning of new airfields in Greenland (DTU 2007). In addition to the basic
features treated in this appendix, COSIMA-TGB contains features for the handling
of the complex data input from flow modelling and from accessibility modelling,
which is one of the MCA criteria (Kronbak 1998). Furthermore, the model is set up
to treat various taxation issues and growth scenarios, and it is also endowed with
features for quantitative risk analysis (QRA) based on Monte Carlo simulation
(MCS) (Vose 2002; Leleur et al. 2004a, b; Salling 2008).
One basic issue to be addressed when setting up a new COSIMA model
application concerns its calibration. In the programme it is possible to select the
alternatives that are to be used for the calibration of the unit prices. Usually, all
alternatives that are thought to be serious contenders as a final choice are used for
the calibration. Alternatively, the project alternative with the highest B/C-rate can
be used for the calibration. The two calibration methods can give slightly varying
results, but which of the two methods to apply may depend on whether the project
with the highest B/C-rate is seen as being challenged by all others or, for example,
a minor or major group is seen as more or less equal candidates.
When calibrating, the user must make sure that all MCA effects are taken into
account. Therefore, if the alternative with the highest B/C-rate does not have a
rating above 0 in all effects, this alternative alone cannot be used for the calibration
and more alternatives must be used. In the case examples 1 and 2, described to
demonstrate the COSIMA approach, it is not possible to calibrate from the
alternative with the highest B/C-rate, as these alternatives do not have
contributions from some or all MCA effects. Therefore, all alternatives are used
for the calibration as all alternatives are thought to be possible contenders for a
final choice. When relating to the appraisal of large transport infrastructure
investments, COSIMA can also be calibrated to accommodate computable general
equilibrium (CGE) analysis, see (Leleur and Holvad 2004) for an analysis of the
resund Fixed Link.
A.4 Comparison of COSIMA to Other Approaches

The different approaches to transport decision making have different strengths and
weaknesses, and in this section, the COSIMA approach will be compared with the
regular CBA and the multi-criteria approach. The three methods will be compared
with emphasis on the three previously mentioned issues: comprehensiveness,
effectiveness and transparency (Andersen and Petersen 2006).
Regarding comprehensiveness, the MCA and COSIMA approaches have clear
advantages compared to the CBA. The pure CBA can only include effects that can
be assigned a monetary value, meaning that some effects cannot be included at all.
These effects must therefore be treated separately as is the case with regard to the
method presently used in Denmark. However, this can present a problem as it is
often difficult for the decision-makers to combine the information from the CBA
and the other effects.
138
Appendix A: COSIMA
The MCA as generic method can include all possible effects as it is not
necessary to be able to assign monetary values to the MCA effects. Therefore, DM
wanting one method to include all effects could be tempted to use the pure MCA
approach. However, the composite method COSIMA can also include all effects,
but as described they will be split into effects that can and cannot in principle be
assigned a monetary value. Furthermore, when using the COSIMA method, the
result is a total rate of return which shows how the benefits are made up of the
CBA part and the different MCA effects. This means that the COSIMA method
provides the DM with CBA based information about socio-economic viability,
which cannot be provided in the MCA. Furthermore, MCA is a comparative tool
only and therefore it does not tell the decision-makers if the benefits of the project
at hand exceed the costs. COSIMA thus combines some merits, as it gives
information about socio-economic viability while at the same time being able to
include all effects like the MCA.
It is not correct to refer to the MCA approachalthough it is only done
genericallyas being one method as many methods are available (Tsamboulas
et al. 1998; Belton and Stewart 2002). In this context the MCA method therefore
refers only to some common traits across a number of individual MCA methods.
However, to ease comparison issues, one could see the SMART approach
developed by von Winterfeldt and Edwards (1986) as an exemplar MCA
methodology that is perceived as a recommendable standard approach (Goodwin
and Wright 2010, pp. 3156). Very briefly one could see the COSIMA approach
and the SMART approach as appraisal methodologies that follow opposite
strategies: where SMART translates what in this context has been described as the
CBA impacts into MCA-like value functions, COSIMA translates the MCA
impacts and their associated value functions into a CBA-like comprehensive type
of analysis, see Eqs. A.1 and A.2. Below CBA, MCA and COSIMA will be
compared further as concerns efficiency and transparency.
Efficiency describes how easily a final decision can be made on the basis of the
results of the analysis. The CBA itself gives a very clear answer as to which
alternative is the most attractive as it presents a single point estimate for each
alternative. However, as not all effects can be included in this analysis, it is
necessary to take these into account separately, and the decision will easily
become more complex.
Similarly, the MCA and COSIMA approaches per se give a final, single
appraisal result for each alternative based on considering all the effects. Therefore,
the final decision on the basis of these two methods is very straightforward,
although it is necessary to scrutinise the applied MCA scores for the different
alternatives and the appropriate CBA/MCA split. With regard to COSIMA it
should be noted that the MCA part is context-dependent by setting focus on the
relative performance of an alternative as compared to the worst of alternatives
under a specific criterion. Therefore the set of alternatives needs to be made up by
alternatives where each of these under a specific viewpoint represents a sound
candidate for being selected for implementation. At the same time it must be
required that each criterion in the criteria set should help to discriminate in an
Appendix A: COSIMA
139
adequate way between the alternatives. If this is not the case (in AHP leading to
very similar scores across the alternatives) this should lead to the exclusion of that
criterion or to the inclusion of an alternative that changes this situation but at the
same time is seen as a sound alternative worthwhile to consider.
Transparency is important in transport decision making to prevent, for example,
public opposition to the decisions taken. The traditional CBA has a high degree of
transparency as the analysis in theory can be redone by anyone. The assigned unit
prices are predetermined ensuring that the project will be appraised in the same
way no matter who carries out the analysis.
Both the MCA and COSIMA assume the presence of the DM during the
calculation process, and the decision-makers are required to make choices
regarding which effects to include and how to rate and rank them. This entails the
need for thorough discussions among the decision-makers and in order to ensure
transparency of the choices taken the reasons behind them must be presented. It is
considered an advantage that the decision-makers are not asked to determine
specific weights directly and that a COSIMA decision conference can be run in a
more or less non-technical way.
The procedures in the MCA method can be very straightforward and easy to
follow for the decision-makers. The COSIMA method may appear to be slightly
more demanding to the decision-makers, for which reason its principles should be
explained, for example, by use of a demo-case at the beginning of the first decision
conference meeting before the described DM/CF questionanswer session begins.
However, the results of the COSIMA method, the TRR values, could be easier to
understand than MCA results as the CBA method is presumed to be well known to
the DM.
The involvement of decision-makers in the entire process demands
transparency but is a great advantage as it makes it possible to adjust the
analysis to local conditions and the specific requirements of the decision-makers.
What is weighted highest in a transportation project can differ from project to
project and between decision-makers. Therefore, DM in one planning context and
location, for example, might assess a problem differently from decision-makers in
another context. Both the MCA and COSIMA approaches allow the users to
influence the setting of weights of the effects according to their wishes, for
example by determining suitable rankings.
Summing up the COSIMA method combines the comprehensiveness of the
MCA with the information given in the CBA about socio-economic viability.
Furthermore, the method allows the decision-makers to include the MCA effects of
their choice and rank and weight them according to their perceptions of importance.
A.5 Conclusions and Perspectives

This appendix has presented and exemplified an appraisal approach, COSIMA, to
assist decision-makers in exploring and appraising transport infrastructure
investments in a systematic way. Although straightforward in its design and
140
Appendix A: COSIMA
application by simply adding to (and not hiding/changing) CBA information,

COSIMA contains features that make it useful to address complex assessment
problems by incorporating relevant MCA criteria and applying different RODbased scenarios.
COSIMA seeks to address the overall feasibility/attractiveness issues of an
appraisal study comprising a number of alternatives by exploring whether some
other issues/impacts complementing the formulated CBA can possibly make a
particular alternative change from, for example, not feasible to overall attractive
compared to the other alternatives?Thus COSIMA has been formulated to deal
with the often occurring problem that the CBA says too little about the particular
problem, for which reason the decision-makers may want further, yet systematic
examinations that can extend the already available CBA information. It can be
concluded that COSIMA can be useful in this respect. Furthermore, administrative
units may consider the COSIMA approach as less of a black-box than other types
of current MCA.
COSIMA has been developed as a multi-purpose or customised decision model
(CDM), from which particular versions of COSIMA software can be set up, with
one elaborate example being COSIMA-TGB addressing airfield alternatives in
Greenland, see (DTU 2007). Compared with the early versions of COSIMA, the
COSIMA-TGB software contains some new possibilities for examination by
handling a range of both different future scenarios and of different sets of user
preferences. A specific set of user preferences affects which MCA impacts should
be included and also their ranking leading to ROD criteria weights and the specific
CBA versus MCA trade-off behind the final assessment results (Jensen et al.
2007).
Finally, it can be noted that although developed for transport planning, the
features and functioning of COSIMA provide it with a general problem-solving
scope. Therefore as a decision analysis tool it can also be applied for planning and
assessment in other areas, see (Barfod et al. 2011), which concerns a COSIMAbased customised decision model (CDM) developed to assist decision making
about strategic office relocation for an international IT company in the resund
Region. This case is treated in Chap. 7.
Technical Notes About SMART, AHP and Calibration

of COSIMA
These notes describe the principles of the SMART and AHP methods used in
COSIMA for the MCA. After this presentation the notes describe how the CBA
and MCA parts are combined. This is addressed by the way the trade-off parameter
MCA% can be calibrated dependent on alternatives. Finally a simple numerical example shows how the calculations are carried out. The notes are based on
(DMG 2010).
Appendix A: COSIMA
141
The Multi-Criteria Method SMART

The SMART technique is based on a linear additive model. This means that an
overall value of a given alternative is calculated as the total sum of the
performance rating/score (value) of each criterion (attribute) multiplied with the
weight of that criterion. The main stages in the analysis are based on (Olson 1996,
pp. 3536; Goodwin and Wright 2010, pp. 3334):
Stage
Stage
Stage
Stage
Stage
Stage
Stage
Stage
Stage
1:
2:
3:
4:
5:
6:
7:
8:
9:
Identify the decision-makers

Identify the issue of the evaluation
Identify the alternatives
Identify the criteria
Determine the alternative-ratings under each criterion
Rank the criteria in order of importance
Determine the weight of each of the criteria
Calculate the overall values and make a provisional decision
Perform sensitivity analysis
In Stage 7 the least important criterion from Stage 6 is assigned the value 10.
The second-least important criterion is given a value that indicates its value
relatively to the least important criterion and so on until all the criteria are assigned
values. These are then normalised into weights summing up to 1.
In SMART, the ratings of the alternatives in Stage 5 are assigned directly by
using appropriate, natural scales of the criteria. For instance, when assessing the
criterion cost for the choice between different road layouts, a natural scale
would be a range between the most expensive and the cheapest road layout. In
order to keep the weighting of the criteria and the rating of the alternatives as
separate as possible, the different scales of the criteria need to be converted into a
common internal scale. In SMART, this is done mathematically by the decisionmakers by means of a value function. The simplest and most widely used value
function is a scale going from 0 to 100.
SMART Exploiting Ranks
The assessment of value functions and swing weights in SMART can sometimes
be a difficult task, and decision-makers may not always feel confident about it.
Because of this, Edwards and Barron have suggested a simplified form of SMART
named SMARTER (Roberts and Goodwin 2002). Using the SMARTER technique
the decision-makers place the criteria into an importance order, for example
Criterion 1 is more important than Criterion 2, which is more important than
Criterion 3, which is more important Criterion 4 and so on leading to C1 C C2 C
C3 C C4 . Afterwards SMARTER assigns weights by using the rank order
distribution (ROD) method.
142
Appendix A: COSIMA
ROD is based on a weight approximation that assumes that valid weights can be
elicited by the ranking of criteria; this is very convenient in a decision conference
as ranking is more easily negotiated than are direct weights.
The approximated ROD weights for n = 2 to 10 have been found
mathematically. For further information about the underlying calculations, see
(Roberts and Goodwin 2002). Weight sets are indicated below for a number of
criteria between two and eight with rank order of criteria from left to right. Note
that Roberts and Goodwin originally indicate the weights with four decimals,
which may, however, indicate a kind of precision that is not realistic. They also
indicate weights for nine and ten criteria but as can be seen from the values below
more than eight criteria will mean that practically no discriminatory power is given
to the criteria ranked as numbers nine and ten. This information is relevant for the
participants in a decision conference when engaged in reducing an initially long
list of criteria, see the case examples in Sect. 8.1.
Two criteria
Three criteria
Four criteria
Five criteria
Six criteria
Seven criteria
Eight criteria
(0.69;
(0.52;
(0.42;
(0.34;
(0.30;
(0.26;
(0.23;
0.31)
0.33; 0.15)
0.30; 0.19; 0.09)
0.27; 0.20; 0.13; 0.06)
0.24; 0.19; 0.14; 0.09; 0.04)
0.22; 0.18; 0.14; 0.10; 0.07; 0.03)
0.20; 0.17; 0.14; 0.11; 0.08; 0.05; 0.02)
The Multi-Criteria Method AHP

The analytic hierarchy process (AHP) method was developed by Saaty in the
1970s and is based on utilising pairwise comparisons as a way of assessing a set of
alternatives (Hwang and Yoon 1995; Saaty 2001). The method is applicable for a
hierarchical structure set out as a decision tree with several levels. In most cases,
however, three levels are applied, namely the alternatives level and the criteria
level and the goal level, with the latter expressing the overall purpose to be
achieved. The idea is then to compare all alternatives under each criterion and
afterwards all the criteria under the goal, with the latter expressing the overall
rationale of implementing one of the alternatives.
The scale applied is shown below:
Equal importance
In-between grading
Moderate importance
In-between grading
Strong importance
In-between grading
(1)
(2)
(3)
(4)
(5)
(6)
Appendix A: COSIMA

In-between grading
Extreme importance
143
(7)
(8)
(9)
With two alternatives (or criteria) A and B compared, the DM are asked to
indicate the preference intensity. If, for example, A is preferred to B with strong
importance grade 5 is chosen (if B is preferred to A with the same intensity the
reciprocal value 1/5 is indicated). In this way a positive comparison matrix aij is
set out with the reciprocal property that aij = 1/aji and the diagonal-elements equal
to 1. Furthermore preference transitivity leads to aij = aik/ajk.
In the original approach by Saaty the PerronFrobenius theory is used to
determine the preference weights (relative performance indicators) for the
alternatives with respect to each criterion and on the next level for the criteria
with respect to the goal. Saaty found that the problem of determining these
preference weights can be formulated as an eigenvector problem (Hwang and
Yoon 1995). By multiplicating upwards in the hierarchy the overall relative
performance of each alternative with respect to the goal can then be determined.
Due to its intuitive appeal and the development of software such as Expert
Choice, AHP gained widespread use (Expert Choice Inc. 2004). This success is no
doubt due to the fact that decisions produced in an AHP session have generally
appeared to be in accordance with the preference-views expressed by the DM
participating in the AHP process (Vaidya and Kumar 2006).
It should be noted, however, that criticism has been expressed concentrating on
the following three issues (Olson et al. 1995; Lootsma 1999):
The 19 ratio scale applied to quantify the preference judgments
The application of the PerronFrobenius eigenvector method
The arithmetic mean aggregation rule
It can be noted that these criticisms have been dealt with by introducing a
multiplicative version of AHP called the REMBRANDT method, which is
based on (1) applying a difference-based scale judging the preference intensity for
the individual pairwise comparisons, (2) replacing the eigenvector method by
geometric mean aggregation (therebyin case of adding a new alternativerank
reversal of presently examined alternatives can be avoided) and (3) by replacing
the weighted arithmetic mean aggregation of scores with aggregation by the
product of relative scores lifted into the power of the criteria weights (Ibid.).
It has been concluded that both the ratio-based AHP and the difference-based
REMBRANDT methods are useful, practical multi-criteria decision making tools
(Barfod and Leleur 2011). The REMBRANDT method is explained as part of the
Technical Notes accompanying Appendix B below.
A thorough treatment of AHP for strategic decision making is given in (Saaty
2001; Bhushan and Rai 2004). As mentioned in Chap. 8 the latter reference has
144
Appendix A: COSIMA
served as a general background reference for the approach to strategic decision

making set out with the systemic planning (SP) approach.
Calibration of COSIMA with a Numerical Example

Calibration Principles
It should be observed that in the software system the COSIMA calculations use a
calibration parameter UPj that functions as a kind of unit or shadow price per index
value for each of the J MCA criteria to produce the bjk values. These benefit
values are determined by bjk = VMCA(Xjk)UPj with UPj[a(MCA%), w(j),
RiRj[K(bij), RjVMCA(Xjj)], see Eqs. A.1 and A.2, and with j indicating the
alternatives Aj that have been included in the set of alternatives serving as a base
for the model calibration.
In the procedure a(MCA%) and w(j) determine a fraction of RiRj[K(bij) that by
unit scaling, see below, leads to the J unit prices that are applied then to calculate
TRR(Ak). Note that TRR(Ak) values are calculated also for the alternatives not
included in the calibration set. Changes in the set of alternatives Aj behind the
calibration will influence the UPj values and thereby the total rate of return (TRR)
values. This pool dependence, of course, will be of interest for the decision
analysts working to formulate the model set up. The alternatives in this respect
ought to be scrutinised as either serious contenders or maybe as wildcards. The
latter type of alternatives ought not to be included in the calibration pool. The
analysts are supposed to inform the decision-makers about the principally open
calibration questions that could be of interest.
In the COSIMA software both the ratio-based original eigenvector method and
the more recent difference-based geometric (REMBRANDT) pairwise comparison
method have been implemented. In overview the following options are available
for intra-criterion rating: AHP, REMBRANDT, direct rating based on either
judgement or measurement, while the following options are available for intercriteria weighting: ROD weight technique or swing weights. There is a procedure
to ascertain the discriminatory power of each MCA criterion included in the
examination. The swing weight (SW) method is based on setting and comparing
the swing in one criterion with the swing of another criterion (ratio of relative
importance with criteria typically chosen in order of importance). SW is seen as
more precise but also more demanding than the ROD weight technique, which
determines more general importance weights. The latter aim at allocating or
distributing means, whereas SW weights are based on trade-off considerations.
With ROD weights the scaling of the unit price UPj is based on the individual
j-sums RjVMCA(Xjj), while, with swing weights applied for setting the criteria
weights, the individual j-sums are replaced by an average of the summed ratings
over the J criteria being equal to 1/J(RjRjVMCA(Xjj).
Appendix A: COSIMA
Table A.8 Benefit-cost data
for the calculation example
145
Alternatives
A1
A2
A3
A4
B1..B4
C1..C4
BCR
110
70
1.57
160
80
2.00
165
120
1.38
120
65
1.85
Numerical Example
The COSIMA calculations can be illustrated with a simple calculation example
based on (Hiselius et al. 2010).
In the present alternative survey four alternatives A1, A2, A3 and A4 are
available. Using a national cost-benefit manual and its fixed unit price values the
total benefits are calculated: B1, B2, B3 and B4, which by dividing them with the
observed total expenditure C1 C2, C3 and C4 leads to benefit-cost rates (BCR) for
the four alternatives, see Table A.8.
If the DM agreeafter the content of the CBA is reviewedthat the decision
making is complete, a decision to choose A2 can be taken, since this alternative
has the highest BCR value = 2.
If the CBA is insufficient, new criteria can be added to evaluate them by a MCA
and finally perform a composite analysis according to the COSIMA principles. The
procedure is as follows: first a number of criteria are described, which in this
example leads to the criteria k1, k2, k3 and k4 that are determined in such a way
that an overlap with the components of the CBA is avoided.
Next, the four criteria are rated and weighted. The rating means that each
alternative for each criterion is assigned a value (score), which lies between 0 and
100. The value 0 is given to the alternative that is performing worst under the
given criterion and 100 to the alternative which is performing the best. The two
remaining alternatives will have values between 0 and 100. The approach is based
on pairwise comparison of all four alternatives under each of the four criteria k1,
k2, k3 and k4. For each of these criteria with four alternatives examined altogether
(4 9 3)/2 = 6 pairwise comparisons are needed. Based on the MCA method
REMBRANDT (see Appendix B) the following scores are obtained from a
transformation of the REMBRANDT results into a value function, see Table A.9.
Since the criteria are usually not assigned equal importance by the decisionmakers, the criteria are assigned the weights K1, K2, K3 and K4. This can be done
directly or by using the ranking criteria method ROD. The result, where the
weights are set directly and summarise to 1, is for this example: (K1, K2, K3,
K4) = (0.20; 0.55; 0.10; 0.15).
In the last part of the calculation the CBA and MCA are linked together, which
is done by decision-makers providing the MCA%. At a high MCA% the MCA will
dominate the final result, while a low MCA% means that it will be the CBA and
the BCR values that dominate.
146
Appendix A: COSIMA
Table A.9 Value function

scores for the calculation
example
Criteria/Alternative
A1
A2
A3
A4
k1
k2
k3
k4
25
0
0
100
100
75
26
68
0
60
100
35
45
100
35
0
The decision-makers are asked about the MCA% and they decide, for example,
to set this CBA/MCA trade-off parameter to 50%. Thus MCA and CBA count the
same in the overall analysis. Based on the choice of A2 with the highest BCR as
calibration basis, the MCA part should now count the same. Benefit value B2
was found to be 160 which means that the MCA part of A2 should also sum up to
160. Adding up the MCA-components of A2 using the scores in Table A.9 p1 can
be determined in the following manner, with p2, p3 and p4 expressed by p1 and the
criteria weights:
100 p1 75 p2 26 p3 68 p4 160 >
100
0:20
0:55
0:10
0:15
p1 75
p1 26
p1 68
p1 160
0:20
0:20
0:20
0:20
Hereby the set of prices is determined:

p1 0:43
0:55
0:43 1:19
p2
0:20
0:10
p3
0:43 0:22
0:20
0:15
p4
0:43 0:32
0:20
With this set of prices the following values of the total rate (total rate of return
TRR) are given, which expresses the overall attractiveness of an alternative from
CBA and MCA:
110 25 0:43 0 1:19 0 0:22 100 0:32
2:18
70
TRRA2 4:00
TRRA1
TRRA3 2:25
TRRA4 4:09
From this it is seen that A4 is the most attractive alternative. In Table A.10 the
results are shown in overview. The example is based on the use of cost-benefit
analysis (CBA) carried out by use of a national manual (NM) and a multi-criteria
analysis (MCA). CBA + NM produce a monetary result, which is validated from
socio-economic thinking and common use. MCA produces a result which is based
on preferences indicated in the decision process and the result is in principle only
valid from this point of view.
Appendix A: COSIMA
147
Table A.10 The results of the calculation example

COSIMA example A1 A2 A3 A4 Method
Costs
Benefits
BCR
k1
k2
k3
k4
Total MCA
Total value
Total rate
70
110
1.57
11
0
0
32
43
153
2.19
80
160
2.00
43
89
6
22
160
320
4.00
120
165
1.38
0
71
22
11
104
269
2.24
65
120
1.85
19
119
8
0
146
266
4.09
CBA +
CBA +
CBA +
MCA
MCA
MCA
MCA
MCA
CBA +
Unit
NM
NM
NM
m DKK
m DKK
eval m DKK
eval m DKK
eval m DKK
eval m DKK
eval m DKK
NM + MCA m DKK and eval m DKK
With the CBA benefits expressed in million DKK (m DKK) the MCA results
(MCA benefits) will also be expressed in million DKK, here to be expressed as
evaluation DKK (eval m DKK) to indicate their way of determination. With the
TRR based on both types of benefits (m DKK and eval m DKK) the TRR benefits
are therefore a mix of m DKK and eval m DKK. The total rate for an examined
alternative is found by adding the benefits from CBA and MCA and dividing them
by the costs in m DKK. Table A.10 presents the results of the numerical example.
The final result is determined by having A2 as the basis for the calibration and
the MCA% set to 50.
By means of the COSIMA calculations it is possible to base the decision of the
choice of alternative on the socio-economic BCR core contribution in combination
with its MCA-based wider performance. This combined result expresses the
overall attractiveness of a given alternative. As mentioned, the CBA result is valid
in the light of being provided by a socio-economic evaluation, while the MCA
result in principle only is valid on the basis of the conducted decision conference
and the actual deliberations that have taken place. A better background can be
obtained if a log book is worked out with user inputs and background comments.
This allows a second-opinion to be set out by inspecting these.
As stated, the values in Table A.10 are developed on the basis of the phrase
influence of the MCA must be 50%. How can this be interpreted further? The
basis for choosing among A1, A2, A3 and A4 is a CBA, which shows that the A2
due to the highest BCR (= 2.00) is the best choice. This BCR value for A2 is given
by C2 = 80 and B2 = 160. A balance between CBA and MCA must be arranged
so that the MCA criteria indirectly priced also contribute with 160, which has just
been illustrated in the calculation example. Keeping the scores of the alternatives
and the criteria weights unchanged this determines unique total rates (TRRs) for
all four alternatives, and A4 stands as the most attractive alternative. In brief the
COSIMA analysis replaces A2 by A4 as the most attractive choice.
Appendix B
SIMDEC
Overview
This appendix demonstrates the SIMDEC approach that is one of the hard methods
in the seven hard and seven soft (2 9 7) methodologies in the SP toolbox, see
Table 5.5. The idea behind risk simulation and multi-criteria analysis in combination for decision making (SIMDEC) is to incorporate risk analysis (RA) by
using Monte Carlo simulation (MCS) as one specific criterion within a multicriteria analysis (MCA). Absolute assessment of each examined alternatives core
performance is provided by RA as a probability-based interval result, whereas
MCA provides a relative, context-dependent assessment related to each alternatives wider performance as set against the other alternatives under examination.
SIMDEC can be adapted to very different types of selection problems, where
criteria are available both in monetary and non-monetary terms. The Decision
Modelling Group (DMG) at DTU Transport at the Technical University of
Denmark has developed software that makes it possible to customise evaluation
models for a specific study.
Appendix B is an updated version of (Leleur et al. 2010) supplemented with
technical notes based on DMG (2010).
B.1 Purpose, Background and Outline

Providing suitable decision support for strategic transport decision making is a
topic of growing concern. For large infrastructure investments, to exemplify with
one important transport topic area, comprehensive assessments are needed
(Banister and Berechman 2000). Typically such investments have many-sided
consequences which all ought to be taken into consideration to seek out the best
alternative from a set of candidates that have come forward from the preparatory
planning and design phases.
The traditional cost-benefit analysis (CBA) as prescribed in various national
and international appraisal manuals is insufficient for comprehensive assessments

149
150
Appendix B: SIMDEC
as often important decision factors such as environment, regional development,

etc. are not possible to cover by CBA (Leleur 2000). Furthermore, uncertainties
play a major role in connection with large-scale projects, where factors such as
construction costs and demand prognoses are uncertain for a number of reasons but
clearly of very high importance for the long-term feasibility of the investment.
Many cases have been documented where uncertainty of these factors have led to
investments that later on turned out to be less than satisfactory (Priemus et al.
2008).
This case concerns a new approach to strategic transport decision making,
SIMDEC, based on using risk simulation and multi-criteria analysis in combination
for decision support. First the SIMDEC modelling framework and the theories
behind are described, and afterwards SIMDEC is illustrated by an example
concerning examination of four alternatives between Helsingr (Elsinore)Helsingborg. These alternatives need to be assessed to facilitate decision making
about a second northern fixed link between Denmark and Sweden to supplement the
already established fixed link in the southern part of resund between Copenhagen
and Malm. Finally, after a discussion of results, a conclusion is given together with
a perspective on the further application of SIMDEC.
B.2 The SIMDEC Modelling Framework

As mentioned in the introduction, SIMDEC is based on applying risk analysis
(RA) in combination with multi-criteria analysis. With the focus of the RA on the
feasibility of each of the alternatives, this concerns feasibility risk assessment
(FRA) where the focus is on the risk that the investment could turn out not to be
socio-economically feasible. SIMDEC proceeds by first examining FRA for the
alternatives one by one, and afterwards the FRA results are used as input as one of
the criteria within a set of decision criteria to a multi-criteria analysis (MCA) that
aims at ranking the alternatives. The two major components of the SIMDEC
modelling framework are described below.
B.2.1
Feasibility Risk Assessment
The FRA is carried out by using Monte Carlo simulation (Vose 2002; Salling
2008) on the results stemming from a conventional cost-benefit analysis, which is
assumed to be prescribed by and conducted in accordance with a manual that can
generally be accepted in the study context. For a large transport infrastructure
investment the impacts to be covered will consist of: construction and maintenance
costs, time savings, operation costs, accident savings, noise emissions, local air
pollution and climate effects based on change in CO2 emissions. With the
exception of construction costs and time savings these effects can be determined in
Appendix B: SIMDEC
151
a relatively precise manner on the basis of current transport engineering modelling

knowledge represented first and foremost by traffic and impact models (Leleur
2000; Trafikministeriet 2003).
Recently a methodology to handle construction costs and time savings that relate
to demand prognoses has been set out with the reference scenario forecasting (RSF)
technique presented in (Salling and Leleur 2009; 2010). RSF concerns Monte Carlo
simulation and is based on prospect theory developed by Kahneman and Tversky in
1979 (Daniel Kahneman received the Nobel Prize in Economics in 2002 for his work
in collaboration with Amos Tversky (19371996)). The prospect theory gave rise to
reference class forecasting (RCF), which was used by Flyvbjerg and others to set out
various uplift principles (Priemus et al. 2008). This again laid the foundation for a
scenario-based simulation procedure that made it possible to formulate and
introduce reference scenarion forecasting (RSF) (Salling and Leleur 2009, 2010).
This appendix about SIMDEC concentrates on applying RSF to provide input to
MCA as this methodology (based on scenarios related to the actual study) for each of
the examined alternatives can produce what has been termed a certainty graph (CG).
This graph is made up of the probability estimates of achieving at least the benefitcost rate (BCR) indicated as argument; thus CG(x) = Prob(BCR C x). Examples
are given in the following section about case illustration.
B.2.2 Multi-Criteria Analysis
The MCA is carried out by using the REMBRANDT technique (Olson et al.
1995; Olson 1996) based on pairwise comparisons for rating of the alternatives and
determination of the criteria weights. The REMBRANDT technique is recognised
as a both valid and practical framework (Lootsma 1999; Barfod et al. 2011; Barfod
and Leleur 2011).
In SIMDEC a set of relevant decision criteria for the decision problem at hand
is laid down. Generally such a set will consist of both monetary and non-monetary
criteria. As already indicated the monetary criteria are taken into account by using
a CBA, which again is used as an input to conducting a feasibility risk assessment
(FRA) of each of the alternatives. Specifically a certainty graph, CG(x), is
produced for each of the alternatives. For each alternative this graph represents its
FRA-performance (based on the monetary criteria and the conducted MCS). This
FRA-performance is added as a criterion to the formulated non-economic criteria
that typically represent strategic issues and impacts relating to the decision
problem. Thereby the total criteria set is established for the examination of the
decision problem. This set should be scrutinised to reduce possible overlapping
with regard to criteria definitions, while at the same time it should be ensured that
no valuable information for the decision making has been left out.
The multi-criteria analysis proceeds by making pairwise comparisons (either by
the decision-makers (DM) themselves or facilitated by analysts interpreting
152
Appendix B: SIMDEC
information revealed by the decision-makers about their preferences). In the rating,

alternatives are successively compared two by two demanding a preference
statement of the following type: very strong preference for.., strong preference
for.., definite preference for.., weak preference for.., and indifference.. (Olson et al.
1995). A numerical REMBRANDT scale value associated with each statement is
fed into the model and afterwards the same procedure is conducted for the other
criteria to determine all the ratings and for the set of criteria to determine weights.
Weights can also be obtained as ROD weights, see Appendix A. Based on ratings
and criteria weights REMBRANDT finally produces a total score for each
alternative, which makes it possible to rank the alternatives in accordance with
their attractiveness.
One of the new features of SIMDEC is the mixing of monetary and non-monetary
decision criteria with the first type of concern represented by the calculated FRAperformance. In SIMDEC the FRA-performance ratings of the alternatives are
based on a set of pairwise comparisons of the previously determined certainty
graphs with each of these representing one of the alternatives that candidate for a
decision about being selected and implemented. The final information presented to
the decision-makers to base their decision on consists of the overall ranking based
on the wider criteria set, where also the non-monetary aspects have been rated based
on pairwise comparisons. Below it is illustrated how SIMDEC manages to reduce
very complex decision-related information to a set of criteria with rated alternatives
that are used to produce a final ranking of the examined alternatives and thereby to
indicate which alternative ought to be preferred.
B.3 Case Illustration

The case illustration concerns an examination of four alternatives for a new fixed
link between Helsingr (Elsinore) and Helsingborg, see Fig. B.1.
The case clearly concerns a complex planning problem where both core and
wider performance of the candidate alternatives need to influence the choice of
alternative. Due to lack of rail capacity on the southern fixed link between
Copenhagen and Malm (opened in 2000) there is a need to provide more crosssound capacity for railway transport. A special concern will be to relieve the
current fixed link for goods traffic. At the same time a new northern link will be an
important piece of the northsouth EU transport corridor Sweden-DenmarkGermany for which reason also person traffic need to be considered. At the same
time a new fixed link will influence the resund Region transport, among other
things by finalising a regional Danish-Swedish public transport circle line
connecting the major urban centres.
Altogether four alternatives were assessed in the SIMDEC examination (Larsen
and Skougaard 2010). As can be seen from the results all four alternatives are
relevant and sound candidates as each of these have qualities that are successively
demonstrated in the examination.
Appendix B: SIMDEC
153
Fig. B.1 HH-fixed link location at Helsingr (Elsinore)-Helsingborg (from Google Maps)
Table B.1 The four alternatives incl. cost in bn DKK for the HH-fixed link
HH-fixed link (alternatives)
Description (type of construction)
Alternative
Alternative
Alternative
Alternative
A1
A2
A3
A4
Tunnel for rail (2 tracks) passenger traffic only

Tunnel for rail (1 track) goods traffic only
Bridge for road and rail (2 9 2 lanes & 2 tracks)
Bridge for road (2 9 2 lanes)
Total cost
(bn DKK)
7.7
5.5
11.5
6.0
The four alternatives are listed in Table B.1 with indication of type of
construction and total cost (1 US$ equals around 5 DKK) (Ibid.)
In general the pure person transport alternatives seem to perform best in a
conventional CBA whereas alternatives with rail improvement gain in the wider
SIMDEC analysis. Due to the fixed links high influence on nearby towns, the
impact on the ecology conditions in the sound (tunnels better than bridge solutions
with regard to flow of water), on regional development (trade, work and educationrelated benefits) and the improvement in network (locally, nationally and in an
EU-perspective) the wider set of decision criteria shown below has been adopted.
Criterion 1: Robustness of feasibility (FRA-performance)
Criterion 2: Impact on towns
154
Appendix B: SIMDEC
Table B.2 The four alternatives with conventional cost-benefit rates

Cost-benefit rates for the four alternatives related to scenarios
The four alternatives for the link
Economic growth expressed by three scenarios
A1:
A2:
A3:
A4:
Tunnel, rail passenger service

Tunnel, rail goods transport
Bridge, road & rail
Bridge, road
High
Middle
Low
1.34
0.43
2.40
3.01
1.15
0.39
2.17
2.63
1.01
0.34
1.94
2.41
Fig. B.2 Certainty graphs

for the four alternatives in the
middle scenario
Criterion 3: Impact on ecology in sound

Criterion 4: Impact on regional economics
Criterion 5: Impact on transport network and accessibility
Based on the Danish national manual for socio-economic assessment
(Trafikministeriet 2003) the benefit-cost rate (BCR) values shown in Table B.2
have been determined applying transport modelling for road and rail traffic set in a
context of three economic development scenarios spanning high, middle (continuation
of established trend) and low economic growth (Larsen and Skougaard 2010).
Afterwards this conventional socio-economic calculation reference scenario
forecasting (RSF) is applied to produce three sets of certainty graphs (CGs)
consisting of CG(A1), CG(A2), CG(A3) and CG(A4), with one set for each of the three
scenarios mentioned above, see Fig. B.2 for the four CGs in the middle scenario.
The RSF-calculations behind the CGs are based on cost-benefit analysis and
Monte Carlo simulation using estimated RSF distributions (Erlang and Beta Pert
distributions for construction costs and time savings respectively with the latter
influenced by the actual scenario), see (Salling and Leleur 2009; 2010).
Each CG can be interpreted as follows: for x = 1 the probability or certainty of
BCR C 1 (with 1 indicating the ordinary socio-economic cut-off value with regard
to feasibility) can be seen from the y-axis value. For the alternatives in the middle
scenario in Fig. B.2 their certainty values (CVs) are: CV(A1) = 22%,
CV(A2) = 0%, CV(A3) = 83% and CV(A4) = 97%. In the scenario runs, the
Appendix B: SIMDEC
155
Table B.3 The four alternatives also expressed by certainty values

Cost-benefit rates and certainty values for the four alternatives related to the scenarios
The four alternatives for the link
Economic growth expressed by three scenarios

High
A1:
A2:
A3:
A4:
Tunnel, rail passenger service

Tunnel, rail goods transport
Bridge, road & rail
Bridge, road
1.34;
0.43;
2.40;
3.01;
37%
0%
86%
100%
Middle
Low
1.15;
0.39;
2.17;
2,63;
1.01;
0.34;
1,94;
2.41;
22%
0%
83%
97%
11%
0%
78%
96%
Table B.4 REMBRANDT rating of criterion 1: robustness of feasibility

Pairwise comparison of alternatives under criterion 1: robustness of feasibility
Scale value (j, k)
A1
A2
A3
A4
A1
A2
A3
A4
0
Strong (-6)
Definite (+4)
Strong (+6)
Strong (+6)
0
Strong (+6)
Very strong (+8)
Definite (-4)
Strong(-6)
0
Weak (2)
Strong (-6)
Very strong(-8)
Weak (-2)
0
Note relating to j compared to k: indifference 0, weak +2, definite +4, strong +6 and very strong +8.
Observe that a reversal of j to k is indicated by -. Elements in the diagonal are all necessarily 0.
expected economic growth associated with the actual scenario will affect the
obtained certainty values, see Table B.3.
Where the conventional cost-benefit rate gives a deterministic point estimate of
the feasibility, the RSF-based certainty values give a probability-based interval
estimate of how the two most important uncertainty factors could affect such a
point estimate. Specifically, construction costs and time savings are simulated
using historical reference class knowledge made operational by using the Erlang
and Beta Pert distributions respectively with the latter embedded in a scenario
context (Ibid.). In the simulation the uncertainty due to the estimation of
construction costs is considered generally, i.e. across the scenarios and not related
to a specific scenario.
In SIMDEC the certainty graphs and certainty values are used as the basis of the
final REMBRANDT procedure with regard to criterion 1 about robustness of
feasibility. The four alternatives are compared two by two resulting in altogether
(4 9 3)/2 = 6 pairwise comparisons as shown in Table B.4.
For the remaining four criteria, information has been gathered to serve as
sufficient background for the criteria rating, which leads to additional 4 9 6
comparisons. As an example the pairwise comparisons for criterion 5 about impact
on transport network and accessibility are indicated in Table B.5.
The rating values for all five criteria are shown in Table B.6 together with the
criteria weights, which have been determined by the ROD technique. In case
pairwise comparison had also been applied for the five criteria, this would have
demanded another (5 9 4)/2 = 10 comparisons.
156
Appendix B: SIMDEC
Table B.5 REMBRANDT rating of criterion 5: impact on transport network and accessibility
Pairwise comparison of alternatives under criterion 5: network and accessibility
Scale value
(j, k)
A1
A2
A3
A4
A1
A2
A3
A4
0
Strong (-6)
Definite (+4)
Weak (+2)
Strong (+6)
0
Very strong (+8)
Strong (+6)
Definite(-4)
Very strong (-8)
0
Definite (-4)
Weak(-2)
Strong (-6)
Definite (+4)
0
Table B.6 Rates and weights for the five criteria

Rates and weights determined by using altogether 30 pairwise comparisons and importance
ranking of the five criteria
Five criteria/four alternatives
A1 rates
A2 rates
A3 rates
A4 rates
Weights
Robustness of feasibility
Impact on towns
Impact on ecology
Impact on regional economics
Impact on network and accessibility
0.03
0.20
0.47
0.05
0.05
0.00
0.79
0.47
0.00
0.00
0.19
0.00
0.03
0.76
0.84
0.78
0.01
0.03
0.19
0.11
0.35
0.13
0.06
0.19
0.27
Table B.7 The four alternatives with total scores and rank order indicated
HH-fixed link Description (type of construction)
Cost
(alternatives)
(bn DKK)
Score
Rank
A1
A2
A3
A4
0.09
0.13
0.44
0.34
4
3
1
2
Tunnel for rail (2 tracks) passenger traffic only

Tunnel for rail (1 track) goods traffic only
Bridge for road and rail (2 9 2 lanes & 2 tracks)
Bridge for road (2 9 2 lanes)
7.7
5.5
11.5
6.0
As concerns the theoretical set up of REMBRANDT it should be noted that the

processing of scale values for the determination of ratings and criteria weights
differs with regard to the so-called progression factors, see (Olson et al. 1995;
Barfod and Leleur 2011). This, however, does not influence the easy and
straightforward application of this multi-criteria methodology. The DMG at the
Technical University of Denmark has developed SIMDEC software that can be
applied for instantaneous use in the context of a decision conference (DMG 2010).
It should be observed that a documentation report, also referred to as a log
book, is worked out as part of doing the pairwise comparisons. Afterwards the
considerations behind each comparison can be studied/inspected to judge the
overall validity of the model outcome. In case of disagreement and debate among
the decision-makers this also makes it possible to make adjustments as a basis for
model reruns.
Based on the values in Table B.6 the total score for each alternative is
determined using all five criteria by R(weight 9 rate) leading to the prioritising of
the four alternatives shown in Table B.7.
Appendix B: SIMDEC
157
B.4 Discussion
Conventional decision support for deciding upon the four HH-alternatives would
consist of a calculation of benefit-cost rates (BCRs) supplemented by various
information not accounted for in the BCRs. SIMDEC offers an approach where
simulation and MCA are applied to deal with the complex decision problem. The
BCR information contained in Table B.2 indicates that three (A1, A3 and A4) out
of the four alternatives are socio-economically sound with BCR values also in the
low growth scenario being around or above 1. Conducting reference scenario
forecasting (RSF), however, indicates, by inspecting the produced certainty graphs
(CGs) and certainty values (CVs) that really only two alternatives (A3 and A4) are
sound when including RA in the assessment (CV C 7580%). By accounting for
estimation bias in the cost estimates and prognosis bias in the forecasting of traffic
and exploring this by simulation embedded in scenarios, the alternative A1 is
shown to have a feasibility that cannot be considered robust.
With alternatives A3 and A4 remaining as candidates for implementation, these
are explored in a wider context where strategic, non-monetary issues are
introduced together with the criterion about robustness of feasibility based on
the described risk analysis. With CGs and CVs of the alternatives as input to this
criterion a REMBRANDT multi-criteria analysis is carried out which comprises
also the criteria about the impacts on towns, on ecology, on regional economics
and on transport network and accessibility. The result is that even with the highest
criterion weight on robustness of feasibility the order of importance with regard to
A4 and A3 is now reversed as alternative 3 now becomes the most attractive
alternative.
For the lower ranking alternatives A1 and A2 it should be observed that A2 is
now better than A1. The wider assessment based on the multi-criteria analysis has
thus revealed and indicated some qualities contained in A2, which were not
captured by the BCR values.
B.5 Conclusions and Perspective

The SIMDEC approach is seen as promising since relatively complex decision
problems of a strategic nature can be based on both explicit risk precaution and
influence from a set of wider, non-monetary issues. As the approach with its
successive building up of assessment information is easy to grasp it can be applied
in decision sessions where a high involvement of decision-makers is possible.
SIMDEC software has been worked out that can facilitate such decision
conferences. Hereby the SIMDEC methodology becomes embedded in a process
that includes also criteria formulation by use of soft operations research methods
such as, for example, brainstorming and futures workshops; based on the software
the participants can ask various what-if-questions to test the robustness of the
158
Appendix B: SIMDEC
priority-ordering of the alternatives. One important issue that can be treated is how
different stakeholder strategiesdefined by the chosen set of decision criteria and
the stated value inputs to the pairwise comparisonswill affect the result (Leleur
2008; Jeppesen 2010).
The SIMDEC approach has so far been tested on transport planning problems
but the perspective is to explore its potential also for complex decision problems
outside the transport sector. It is expected that alternatives for construction projects
in general can be examined by SIMDEC in a way that satisfies both theoretical
validity and practical userfriendliness.
Technical Notes About REMBRANDT, Monte Carlo

Simulation and Certainty Graphs
These notes first describe the principles of the REMBRANDT method used in
SIMDEC for the multi-criteria analysis. REMBRANDT has been shown to have
some advantages compared to AHP. After this presentation the notes describe how
risk analysis (RA) by the use of Monte Carlo simulation can produce the certainty
graphs made use of in SIMDEC as one of the criteria that enter the REMBRANDT
examination of alternatives. The notes are based on (DMG 2010).
The Multi-Criteria Method REMBRANDT

There are various methods for the assessment of alternatives based on pairwise
comparisons. The most known and used method is the AHP method developed by
Saaty over the past 30 years. Saatys method has been criticised because of
weaknesses in the theoretical basis, see (Belton and Stewart 2002). An improved
theoretical model has been formulated by Lootsma in the beginning of the 1990s
with the REMBRANDT method (Ratio Estimation in Magnitudes or deci-Bells to
Rate Alternatives which are Non-DominaTed), see (Olson et al. 1995; Lootsma
1999).
The REMBRANDT system is intended to remedy three contended flaws in
AHP. First, direct rating is now on a geometric scale (using logarithmic
transformation) instead of on a ratio scale. Second, the scores and weights are now
calculated by the geometric mean. Third, aggregation of scores by the arithmetic
mean is replaced by the product of the individual scores weighted by the power of
the normalised weights obtained from analysis of hierarchical elements above the
alternatives. The AHP and REMBRANDT methods are based on the preference
scales in Table B.8.
In connection with the specific pairwise comparison, the user should only
concentrate on the verbal scale, while the two numerical scales only are of
technical interest as input into the mathematical model.
Appendix B: SIMDEC
159
Table B.8 AHP and REMBRANDT scales

Preference
Explanation
intensity
AHP
REMBRANDT
Indifference
2, 4, 6, 8
1, 3, 5, 7
Weak
Definite
Strong
Very strong
Compromise
Neither of the two alternatives is preferable

over the other
One of the alternatives is preferred slightly
over the other
One of the alternatives is preferred definitely
over the other
One of the alternatives is preferred strongly
over the other
One alternative is preferred very strongly
over the other
Values for graduation between two
of the preferences above
There are as mentioned three main criticisms of the AHP method which the
REMBRANDT method tries to correct. The first is related to the scale in AHP
where 1 represents two objects being equal in value, 3 means that the first object is
slightly better than the second object, 5 indicates clear preference in this respect, 7
a strong preference and 9 a very strong preference. Based on a number of examples
and reflections, Lootsma (1999) has adjusted the numerical scale for
REMBRANDT, so it is more convenient for subsequent calculations.
The second point that REMBRANDT tries to improve is the calculation of
scores. AHP uses a method which has the disadvantage that if a new alternative is
added later in the process, it may reverse the existing ranking of alternatives
(known as rank reversal of alternatives). REMBRANDT uses logarithmic
regression or geometric mean, whereby the potential problem of rank reversal is
overcome. For a more detailed technical analysis refer to Olson et al. (1995).
The third and last point which the REMBRANDT method tries to improve
compared with AHP is the way the individual scores are aggregated. The AHP
uses a method based on calculation of eigenvectors leading to scores and
arithmetic mean aggregation by summation of the scores multiplied by the criteria
weights, while REMBRANDT calculates the value of an alternative by using the
geometric mean scores and multiplying these scores after they have been uplifted
with the criteria weights.
To illustrate the principles of REMBRANDT a small calculation example is
described below (Ibid.).
There is a decision problem involving three alternatives (A, B and C) and four
criteria (W, X, Y and Z). The criteria weights are already set to:
0:493; 0:246; 0:174; 0:087
Scores for each alternative under each criterion is calculated using the following
transformation: eln(2) d(jk). It is noted that when REMBRANDT is used to
determine criteria weights the transformation eln(H2) d(jk) is used (Lootsma 1999;
160
Appendix B: SIMDEC
Barfod and Leleur 2011). The pairwise comparisons of the three alternatives under
each of the four criteria are shown below:
Pairwise comparison:
Transformation:
Criterion W:
A B
A 0
4
B -4 0
C -6 -4
C
6
4
0
Criterion X:
A B
A 0
-2
B 2
0
C -1 -4
C
1
4
0
A
B
C
Criterion Y:
A
B
A 0
0
B 0
0
C 4
3
C
-4
-3
0
Criterion Z:
A
B
A 0
1
B -1 0
C 1
2
C
-1
-2
0
B
16
1
0.0625
C
64
16
1
10.08
1
0.0992
A
1
4
0.5
B
0.25
1
0.0625
C
2
16
1
0.7937
4.0
0.3150
A
B
C
A
1
1
16
B
1
1
8
C
0.0625
0.125
1
A
B
C
A
1
0.5
2
B
2
1
4
C
0.5
0.25
1
A
B
C
A
1
0.0625
0.015625
Geometric mean:
0.3969
0.5
5.0397
1
0.5
2
Afterwards the total score for each alternative A, B and C is found by using the
determined values above and the criteria weights of the four criteria W, X, Y and Z
as indicated below; the obtained total scores are as a matter of convention (like in
AHP) transformed into a normalised set of numbers.
A:
B:
C:
10.080.493
10.493
0.09920.493
* 0.79370.246
* 40.246
* 0.3150.246
* 0.39690.174
* 0.50.174
* 5.03970.174
* 1 0.087
* 0.50.087
* 20.087
=
=
=
2.513
1.174
0.339
0.624
0.292
0.084
It should be noted that in the SIMDEC example criteria weights were

determined by the ROD technique.
Monte Carlo Simulation and Certainty Graphs

Monte Carlo simulation (MCS) is a commonly used technique for risk analysis
(RA) as concerns project appraisal. The purpose of RA is to calculate the
combined impact of various uncertainties in model variables to determine an
Appendix B: SIMDEC
161
overall uncertainty influence (Vose 2002). In this context, a risk or uncertainty

assessment is prepared by inserting different continuous or discrete probability
distributions. In SIMDEC the Erlang distribution has been used for the
construction costs and the Beta Pert distribution for the traffic demand forecasts
on the new transport infrastructure alternative (Salling 2008). Other application
cases would make it relevant to examine other types of probability distributions to
be applied for the model variables expected to dominate the overall uncertainty. In
SIMDEC the model software @RISK has been applied, which can be linked to the
Excel-based SIMDEC calculations (Palisade 2007; DMG 2010).
The principle behind MCS is, based on the applied probability distribution
functions, to make (simulate) a large number of individual model results or
events, typically around 2,000 runs or more. In case of simulating the benefitcost rate (BCR) the individual model results are pieced together to represent the
overall simulation result. The following explains the process of an MCS within the
frame of SIMDEC as four main steps:
1.
2.
3.
4.
Determine the uncertain variables

Add an appropriate probability distribution to selected variables
Simulate a set of benefit-cost rates (BCR) by making a number of runs
Plot and interpret the probability distribution of the benefit-cost rates (BCR)
values
The use of MCS for transport evaluation is described in (Salling 2008) with a
special focus on feasibility risk assessment (FRA) concerned with the examination
of the certainty that a given transport investment project is feasible from a socioeconomic viewpoint. This can be illustrated as follows.
The CBA result is typically presented by one or more of the following index
values: the net present value (NPV), the internal rate of return (IRR), the benefitcost rate (BCR) or a combination of these. Setting focus on the BCR value the
MCS is used to determine the robustness of feasibility when a CBA-based point
estimate is transformed into an interval result provided by a certainty graph for the
investment showing non-feasible and feasible outcomes by use of probability
estimates of the BCR values. Specifically, this graph is made up of the probability
estimates of achieving at least the BCR indicated as argument; thus
CG(x) = Prob(BCR C x). An example is shown in Fig. B.3 indicating a 90%
confidence interval between BCR-rate values 0.80 and 2.44 and a certainty value
(CV) equal to 83%, which means that the BCR-rate value has a 83% probability of
being equal to or higher than the cut-off value equal to 1.
In recent research of feasibility of transport investment projects it has been
verified that especially construction costs and traffic demand forecasts are
important (Priemus et al. 2008; Salling and Leleur 2009; 2010).
In an ongoing research project Uncertainty in Transport Project Evaluation,
UNITE (20092012), funded by the Danish Strategic Research Council a decision
support system (DSS) has been developed containing both a deterministic CBA
module, a deterministic MCA module (AHP and REMBRANDT) and a stochastic
MCS module. Thereby the DSS software can support both the COSIMA
162
Appendix B: SIMDEC
Fig. B.3 The concept of certainty graph
methodology described in Appendix A and the SIMDEC methodology described

in this Appendix B.
The DSS software has as part of the research project EcoMobility (20102012),
funded by the EU Regional Development Fund Interreg IV-A, been used to
implement the EcoMobility assessment model. This model applies the SIMDEC
approach to make it possible to include an impact studied and labelled greening
of goods logistics as a decision factor. The model has been tested in a recent
decision conference (October 2011), and it was verified that the certainty graphs
can easily be understood and interpreted by the participants to judge the influence
of socio-economic robustness being one of the important criteria in the SIMDEC
analysis.
References for Appendices A and B

Andersen M, Petersen NB (2006) Bedmmelse af omfartsveje i Allerd. M.Sc. thesis, Centre for
Traffic and Transport (CTT), Technical University of Denmark (DTU)
Banister D, Berechman J (2000) Transport investment and economic development. UCL Press,
London
Barfod MB, Leleur S (2011) Scaling transformations in the REMBRANDT technique: a
sensitivity examination of the progression factors. Paper presented at the 21st international
conference on multiple criteria decision making, Finland
Barfod MB, Vestergaard AV, Leleur S (2011) Examination of decision support systems for
composite CBA and MCA assessments of transport infrastructure projects. In: Shi Y, Wang S,
Kou G, Wallenius J (eds) New state of MCDM in the 21st century. Lecture notes in
economics and mathematical systems, vol 648. Springer, pp 167176
Appendix B: SIMDEC
163
Belton V, Stewart TJ (2002) Multiple criteria decison analysisan integrated approach. Kluwer
Academic Publishers, Dordrecht
Bhushan N, Rai K (2004) Strategic decision making: applying the analytic hierarchy process,
springer series in decision engineering
DMG (2010) Decision modelling group compendium decision support: theory and practice,
DMG, DTU Transport. Technical University of Denmark, Denmark
DTU (2007) Dokumentationsmodel for TGB Vurderingsmodel (Trafikplan for Grnland:
Beslutningsredskab), Center for Trafik og Transport (CTT). Danmarks Tekniske Universitet
ECMTEuropean Conference of Ministers of Transport, Group of Experts (1981) Exchange of
information criteria applied to transport infrastructure projects, ECMT Publication
Expert Choice Inc (2004) Expert Choice, Arlington
Goodwin P, Wright G (2010) Decision analysis for management judgement, 4th edn. Wiley, New
York
Hiselius LW, Barfod MB, Leleur S, Jeppesen SL, Jensen AV, Hjalte K (2010) Helhetsorienterad
utvrdering av kollektivtrafiktgrder, Bulletin 246, Institutionen fr Teknik och Samhlle,
LTH, Lunds Universitet
Hwang CL, Yoon K (1995) Multi attribute decision makingan introduction, Sage University
Series No. 07-104, Sage Publications, London
Jensen AV, Barfod MB, Leleur S (2007) Strategic planning and decision analysis: presentation of
the COSIMA software system. Paper presented at the INFORMS 2007 Congress, Puerto Rico,
July 2007
Jeppesen, SL (2010) Sustainable transport planning. Ph.D. dissertation, Department of Transport,
Kronbak J (1998) Trafikplanlgning og GIS-baserede konsekvensberegninger. Ph.D. Dissertation, Institut for Planlgning. Danmarks Tekniske Universitet
Larsen LA, Skougaard BZ (2010) Vurdering af alternativer for en fast forbindelse HelsingrHelsingborg. M.Sc. thesis, DTU Transport, Danmarks Tekniske Universitet
Polyteknisk Forlag, Lyngby
Leleur S (2008) At navigere mod fremtiden: Systemisk planlgning som ide og metode.
Polyteknisk Forlag, Lyngby
Leleur S, Holvad T (2004) New appraisal methodology for large European transport investments.
In: Proceedings of the European transport conference 2004. Association for European
Transport (AET), Strasbourg
Leleur S, Jensen AV, Salling KB (2004a) COSIMAsoftware manual, Version June 2004,
Centre for Traffic and Transport (CTT). Technical University of Denmark
Leleur S, Jensen AV, Salling KB (2004b) Modelling decision support and uncertainty for large
transport infrastructure projects: the CLG-DSS model of the resund fixed link. In: Antunes
CH, Dias LC (eds) Proceedings of the 15th mini-EURO conference in managing uncertainty
in decision support modelsMUDSM 2004, INESC-Coimbra, Coimbra, 2004
Leleur S, Petersen NB, Barfod MB (2007) The COSIMA approach to transport decision making:
combining cost-benefit and multi-criteria analysis for comprehensive project appraisal. In:
Kim K-S (ed) Proceedings of the Korean Development Institute and World Bank conference,
Seoul, May 2007. Improving public investment management for large-scale government
projects: focusing on the feasibility studies. KDI, pp 100122
Leleur S, Larsen AL, Skougaard BZ (2010) Strategic transport decision making: the SIMDEC
approach based on risk simulation and multi-criteria analysis. In: EUROSIS Proceedings of
the Asian simulation technology conference ASTEC 2010 conference, Shanghai, March 2010
Lootsma FA (1999) Multi-criteria decision analysis via ratio and difference judgement. Kluwer
Academic, Dordrecht
Olson DL (1996) Decision aids for selection problems, springer series in operations research
Olson DL, Fliedner G, Currie K (1995) Theory and methodology: comparison of the
REMBRANDT system with analytic hierarchy process. Eur J Oper Res 82:522539 (Elsevier)
164
Appendix B: SIMDEC
Palisade (2007) @RISK version 5.0Advanced risk analysis for spreadsheets. Palisade
Corporation, New York
Priemus H, Flyvbjerg B, van Wee B (2008) Decision making on mega-projects: cost-benefit
analysis, planning and innovation. Edward Elgar Publishing, Cheltenham
Roberts RJ, Goodwin P (2002) Weight approximations in multi-attribute decision models.
J Multi-Criteria Decis Anal 11:291303 (Wiley)
Saaty TL (2001) Decision making for leaders: the analytic hierarchy process for decisions in a
complex world. RWS Publications, Pittsburgh
Salling KB (2008) Assessment of transport projects: risk analysis and decision support. Ph.D.
Dissertation, Department of Transport, Technical University of Denmark
Salling KB, Leleur S (2009) Modelling of transport project uncertainties: risk assessment and
scenario analysis. In: Proceedings of the modelling and applied simulation09 conference,
Puerto de la Cruz, Tenerife, Sept 2009
Salling KB, Leleur S (2010) Reference scenario forecasting: a new approach to transport project
assessment. Paper presented at the 12th world conference on transport research in Lisbon, July
2010
Steffensen M, Testmann L (2000) Vurdering af omfartsveje belyst gennem alternative projekter
for Hng. M.Sc. thesis, Institut for Planlgning, Danmarks Tekniske Universitet
Trafikministeriet (2003) Manual for samfundskonomisk analyse, Copenhagen, juni 2003
Tsamboulas D, Beuthe M, Grant-Muller S, Leleur S, Nellthorp J, Panou K, Pearman A, Rehfeld C
(1998) Innovations in decision analysis, deliverable D10, EUNET research project, EU 4th
Framework Programme
Vaidya SO, Kumar S (2006) Analytic hierarchy process: an overview of applications. Eur J Oper
Res 169:129 (Elsevier)
Vose D (2002) Risk analysis: a quantitative guide, 2nd edn. Wiley, New York
von Winterfeldt D, Edwards W (1986) Decision analysis and behavioural research. Cambridge
University Press, Cambridge
Index
@ RISK, 161
A
Ackoff, 43
AHP, 65
Airport in Greenland case, 100
Analyst, 96
Analytic hierarchy process (AHP), 52, 62,
80, 108
Appraisal study, 125, 140
Arationality, 38
Aristotle, 23
Artificial intelligence, 37
Ashby, 19
Assessment, 55
Autopoeisis, 15
B
Benefit-cost rate (BCR), 64
Bhushan, 109
Black Swan theory, 113
Bohr, 2930
Boundary setting, 7, 57
Brainstorming (BS), 52, 72, 90, 103
Buchanan, 6
Business innovation, 117
Business problems, 108
Butterfly Effect, 8
C
Capra, 2930
Causality, 23
CBA information, 106

Chaos management, 22
Checkland, 43, 5960
Choice intelligence, 66
Churchman, 48
Closed change, 21
Cognition, 27
Cognitive billboard, 44, 46, 48
Cognitive endeavours, 115
Cognitive pluralism, 67
Complementarity, 29
Complex planning problem, 12, 27, 55, 152
Complex strategic choices, 3, 13, 66, 109
Complex world, 1
Complexity, 18, 23, 28, 112
Complexity mode of enquiry, 46
Complexity paradigm, 28
Complexity theory, 6, 147
Composite methodology, 62
Composite methodology for assessment
(COSIMA, SIMDEC), 52
Computable general
equilibrium (CGE), 137
Consequences, 62
Constructive circularity, 31
Constructivist perspective, 16
Contained change, 21
Contingency, 15, 1819
Conventional planning, 20
Core performance, 57, 75, 88
Corporate social
responsibility (CSR), 117
COSIMA, 85, 106
COSIMA approach, 123
COSIMA principles, 145

165
166
C (cont.)
COSIMA software, 136, 144
Cost-benefit analysis (CBA), 52, 6263, 92
Criteria weights, 84
Critical systems heuristics (CSH), 44, 52,
5758, 76, 90, 98
Critical systems thinking (CST), 44
Customised decision model, 136137
D
Danish bypasses case, 97
Danish Strategic Research Council, vii
Decisions, 20
Decision analysis, 7677, 140
Decision analyst, 77
Decision analysts, 13, 22
Decision awareness, 115
Decision conference, 73, 76, 95, 135
Decision conference participants, 92
Decision criteria, 124
Decision engineering, 109
Decision ownership, 115
Decision space, 57, 73, 115
Decision support, 2, 66, 77, 106, 124, 150
Decision tree, 80, 142
Detached understanding, 96
Detail complexity, 5
Deterministic point estimate, 155
Direct pricing, 64
Dreyfus and Dreyfus, 31, 3538, 68
Dynamic complexity, 8
E
EcoMobility (20102012), vii
EcoMobility assesment model, 162
Economics, 64
Economic assessment, 63
Economic consequences, 64
Eigenvector problem, 143
Ellis, 117
Emancipatory mode of enquiry, 46
Emancipatory paradigm, 46
Emergency planning, 108
Energy planning, 106
Epistemic lenses, 30
Epistemology, 27
EU Regional Development Fund, 162
European strategic transport
research, 106
EU transport studies, 124
Expert Choice, 143
Index
F
Facilitator, 90, 96
Financial analysis (FA), 63
Five-stage learning model, 37, 96
Fixed link between Elsinore
and Helsingborg case, 107
Flood, 7
Flyvbjerg, 151
Foucault, 44
Functionalist mode of enquiry, 45
Functionalist paradigm, 45
Futures workshop (FW), 52, 57, 90
G
Generic decision making
problem, 108
Geographic information
system (GIS), 76
Gilboa, 68
Globalised world, 1
Glocalisation, 1
Goodwin, 141
Governance problems, 108
Group learning, 76
Group processes, 68, 7677, 95
H
Habermas, 10, 44
Hard methods, 52, 62, 107, 116
Hardin, 9
Heisenberg, 2930
High speed rail case, 102
Holistic, 2, 29, 39
Horizon of possibilities, 112
I
Information technology, 76
Intelligent transport systems
(ITS) case, 105
Interactions, 17
Interpretation, 95
Interpretive mode
of enquiry, 45
Interpretive paradigm, 45
Investment criteria, 64
Involved understanding, 96
J
Jackson, 4348
Index
K
Kahneman, 151
Khisty, 12
L
Law of Requisite Variety, 19
Learning, 95
Leleur, 47, 52
Light rail service case, 103
Linstone, 48
Litmus test, 97
Log book, 83, 106, 147, 156
Long-term consequences, 90
Lootsma, 158159
Lorenz, 8
Luhmann, 15, 1720, 31, 68, 115
Lyotard, 44
M
Management, 20, 23
Management thinking, 67
Mandelbrot, 5
Market pricing, 64
Martin, 117118
Maturana, 16
MCA information, 106
McCarthy, 10
Means-ends configuration, 12
Midgley, 115
Mind mapping (MM), 52, 57, 104105
Mitroff, 48
Modes of enquiry (MOEs), 27, 4547, 95
Mohammadi, 12
Monte Carlo simulation, 66, 92
Morin, 29, 31, 115
Multi-attribute utility theory, 65
Multi-criteria analysis (MCA), 64, 92
Multi-methodology approach, 109
Multiple perspectives, 48
N
Narayanan, 118
Net present value (NPV), 64
Newtonian physics, 28
Non-economic
consequences, 64
O
Olson, 141, 158159
Open-ended change, 21, 27
167
Operations research, 39, 64
Optimisation, 39, 64
Organisation, 20
Organisations, 17
Organisational decision making, 63
Organised complexity, 10
Ownership of the problem, 97
P
Pairwise comparison, 65, 84, 103, 157
Paradigms, 27
Perception of learning, 35
PerronFrobenius theory, 143
Planners, 13, 22
Planning, 1819, 22, 49, 56
Planning team, 48, 50
Poincar, 8
Postmodern mode of enquiry, 46
Postmodern paradigm, 46
Preference analysis (PA), 52, 62, 65
Preference complexity, 10, 112
Preference information, 81
Prigogine, 23
Proactive effort, 22
Probability-based interval, 155
Problem-solving, 49
Project appraisal, 123
Project feasibility, 124
Promoting biking in
Denmark case, 104
Q
Quantitative risk analysis, 137
Qvortrup, 1
R
Rai, 109
Rank order distribution (ROD)
weights, 84, 126
Rationality, 38
Recasting of systemic perceptions, 48
Reframing/back-talk schema, 32
REMBRANDT, 65, 143, 151, 158159
Risk, 65
Risks, 62
Risk analysis, 66, 92
Risk analysis based on Monte Carlo
simulation (RA), 52, 62
Robustness, 116, 153, 157, 161
Rosenhead, 23
Rumsfeld, 112
168
S
Saaty, 142
Scanning, 55
Scenarios, 154
Scenario analysis (SA), 52, 62
Schn, 3132
Scoping, 5556
Search-learn-debate process, 39
Self-organisation phenomena, 24
Self-organising multi-causality, 28
Senge, 7, 9, 11
Shannon, 10
SIMDEC, 106, 151, 156
SIMDEC approach, 149
Simon, 10, 36
Simple multi-attribute ranking technique
(SMART), 52, 62, 141
Simplicity paradigm, 28
SMART technique, 141
Social return of investment (SROI), 117
Social systems, 1617, 19
Societal complexity, 1
Socio-technical system, 5, 7, 112
Soft methods, 52, 57, 72, 90, 116
Soft systems methodology (SSM), 52, 57, 59,
76, 90, 101
SP framework, 2, 69, 109, 118
Stacey, 8, 2123, 68
Stakeholder analysis (STA), 52, 57, 65, 72
Stakeholder viewpoint, 90
Step-by-step approach, 49
Strategic cognition (SC), 118
Strategic decision making, 56
Strategic decisions, 2
Strengths, weaknesses, opportunities and
threats (SWOT), 52, 57
Subnuclear physics, 28
Suboptimisation, 57
Subworld, 38, 76, 95, 115
Sustainability, 117
SWOT analysis, 90
Swing weight (SW) method, 144
Sympoietic, 24
System, 16, 47, 49
System complexity, 19
System demarcation, 7
System environment, 17
System/environment, 15
Systematic approach, 2, 4950
Systematic assessment, 99
Systematic method-elements, 2
Systematic planning, 2, 24, 72
Systematic thinking, 30
Index
Systemic perception, 48
Systemic planning (SP), 95
Systemic scanning, 99
Systemic thinking, 30
Systemic toolbox, 50, 52, 76, 9091
Systems analysis, 49
Systems science, 28, 43, 49
Systems techniques, 39
Systems theory, 16
Systems thinking, 16, 44
T
Taleb, 113114, 117
Theory of communication, 10
Thyssen, 17, 20
Total rate of return (TRR), 87
Total systems intervention (TSI), 44
Trade-off analysis, 85
Tragedy of the commons, 9
Transformative teleology, 23
TRANS-IT Consult, 71, 92
Transport decision making, 124, 137
Transport infrastructure planning, 124
Transport modelling, 154
Transport planning, 106, 116
Tversky, 151
Types of change, 21
U
Ulrich, 44, 58
Uncertainty, 1, 29, 65, 117
Uncertainty principle, 30
UNITE (20092012), vii
Unknown unknowns, 2, 112
V
Value function (VF), 79
Varela, 16
Vessel traffic service case, 101
Von Koch, 5
W
Weight sets, 142
Weights, 64
Wider performance, 57, 75, 88, 91, 97
World complexity, 39
Wright, 141

Complex Strategic Choices PDF

Diunggah oleh

Informasi Dokumen

Judul Asli

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Complex Strategic Choices PDF

Diunggah oleh

Hak Cipta:

Format Tersedia

Decision Engineering

For further volumes:

Complex Strategic Choices

Applying Systemic Planning for Strategic

Springer London Heidelberg New York Dordrecht

to office clerk at DTU Transport, translator Ulla Salado-Jimena for valuable

Introduction: Complexity as a Challenge . . . . . . . . . .

The Condition of Complexity . . . . . . . . . . . . . .

Linking Complexity and Simplicity. . . . . . . . . . . . . . . . . . . . . .

The Systemic Process . . . . . . . . . . . . . . . . . . . . . . . . . .

The Systemic Toolbox . . . . . . . . . . . . . . . . . . . . . . . . . . .

Setting up the Decision Support . . . . . . . . . . .

Company Relocation as Demo-Case . . . . . . . . . . . . . . . . . . . . .

The Relocation Decision Conference: Scenario

A Summing up: The Challenge of Strategic Decision Making

Analytic hierarchy process

Introduction: Complexity as a Challenge

1.1 Complex Strategic Choice and Systemic Planning

S. Leleur, Complex Strategic Choices, Decision Engineering,

1 Introduction: Complexity as a Challenge

1.1.1 The Idea of Systemic Planning

1.1.2 The Framework of Systemic Planning

1.1 Complex Strategic Choice and Systemic Planning

1.2 An Overview of the Contents

1 Introduction: Complexity as a Challenge

1.2 An Overview of the Contents

1.3 A First Look at Complexity and Foresight Problems

1.3.1 Detail Complexity

1 Introduction: Complexity as a Challenge

Fig. 1.1 The Koch curve as

1.3 A First Look at Complexity and Foresight Problems

There is no doubt that system demarcation or boundary setting becomes

1.3.2 Dynamic Complexity

1 Introduction: Complexity as a Challenge

1.3 A First Look at Complexity and Foresight Problems

1.3.3 Preference Complexity

1 Introduction: Complexity as a Challenge

1.3 A First Look at Complexity and Foresight Problems

Fig. 1.3 The shaping of an

1 Introduction: Complexity as a Challenge

Adapted from Khisty and Mohammadi (2001, p. 22)

1.3.4 Categories of Problems

1.3 A First Look at Complexity and Foresight Problems

optimising, scheduling etc., whereas type D are the utmost challenges as it

1 Introduction: Complexity as a Challenge

Khisty CJ, Mohammadi J (2001) Fundamentals of systems engineering with economics,

The Condition of Complexity

2.1 The Systems World of Luhmann

System/environment distinctions are a main focus in the part of Luhmanns

S. Leleur, Complex Strategic Choices, Decision Engineering,

The Condition of Complexity

2.1.1 Systems According to Luhmann

2.1 The Systems World of Luhmann

The basic features of autopoietic systems, briefly introduced above, make it

2.1.2 Systems and Complexity

How can operationally closed systems interact?This apparent difficulty is

The Condition of Complexity

non-members of the organisation to get an understanding of it and act accordingly

2.2 System and Environment

Recognising contingency as a condition of planning means recognising that

2.2 System and Environment

2.2.1 The Midas Touch of Contingency

There seems to be some promise here for planning, if planning as a discipline