Int J Interact Des Manuf (2009) 3:1323

DOI 10.1007/s12008-009-0057-9
ORIGINAL PAPER
Analysis and evaluation of product design through design aspects
using digraph and matrix approach
V. Paramasivam V. Senthil
Received: 17 November 2008 / Accepted: 15 January 2009 / Published online: 6 February 2009
Springer-Verlag 2009
Abstract Product design evaluation is essential for all
manufacturing industries to explore the soundness and effec-
tiveness of the product design. This paper presents a math-
ematical model for evaluating and analyzing the product
design alternatives using graph theory and matrix approach.
In this paper, various contributing factors are identied and
their relative importance has been considered. A digraph
model is constructed to represent the abstract information of
the product design which takes into account of all the factors
and is much useful for visual analysis. The digraph model
is converted into matrix form, which is used for computer
processing. An index is obtained from the product design
evaluation function, derived from the matrix for all prod-
uct design alternatives and it shows the effectiveness of the
product design. Indices are calculated for all product design
alternatives and they are ranked in ascending order and the
product design corresponding to the rst rank is selected as
best one. The proposed methodology is quite versatile from
the point of view that it integrates all factors of the prod-
uct design. An example has been included to illustrate the
methodology proposed in this paper.
Keywords Product design Digraph and matrix approach
Product design evaluation index
V. Paramasivam (B)
Department of Mechanical Engineering, PSNA College
of Engineering and Technology, Dindigul-5, India
e-mail: pudidha@yahoo.com
V. Senthil
Department of Mechanical Engineering, Coimbatore Institute
of Technology, Coimbatore, India
e-mail: vsenthil_cit@yahoo.com.sg
1 Introduction
A successful product design fulls the consumers needs. It
is very important for the designers to recognize these needs
go beyond the utilitarian and functional to include the inspi-
rational, emotional and cultural. Also new product develop-
ment is indeed very important for manufacturers. However,
developing newer products is a risky and uncertain process.
In order to reduce the risks and uncertainties, companies need
to evaluate their new product initiatives carefully and make
accurate decisions. This demonstrates the increasing impor-
tance of the role of design both for economic competitiveness
and for improvement of the quality of life and work. One of
the major directions during the design process is that the
products should manifest end users point of view, from ini-
tial concept to their distribution to the market place. Thus the
product design evaluation is must at all phases of the product
development process from concept phase to detailed design
phase. Design evaluation is time consuming and laborious,
since many factors have to be considered in relation to the
development and design which vary in character and com-
plexity. Design evaluation cannot be made without structured
decision making tools (aids) since the decision alternatives
are too many and simultaneous criteria impact on the decision
are too vast to consider at once by human decision makers.
Maddulapalli and Azarm [1] addressed both kinds of var-
iability, i.e., variability in the preferences of the decision
maker and variability in the attribute levels of the design
alternatives. They presented a method for product design
selection with variability in preferences for an implicit value
function and later extended it to account for variability in
attribute levels of design alternatives. Besharati et al. [2] pre-
sented an integrated design and marketing approach to facili-
tate the generation of an optimal robust set of product design
alternatives to carry forward to the prototyping stage. Their
1 3
14 V. Paramasivam, V. Senthil
approach evaluates performance and robustness of a design
alternative due to variations in its uncontrollable parameters.
Vanegas and Labib [3] presented several fuzzy approaches to
design evaluation. Weight of criteria and performance levels
are captured by fuzzy numbers, and the overall performance
of an alternative is calculated through the newfuzzy weighted
average. Li and Azarm [4] presented an approach wherein
product design is viewed as a selection process with two
main stages: design alternative generation and design alter-
native evaluation. The focus of their paper was mainly on a
design alternative evaluation model in that designers pref-
erences, customers preferences, and market competition are
accounted for the selection of the best possible design. See
et al. [5] discussed the problem of selecting a design from a
set of alternatives using multiple, potentially conicting cri-
teria. They also demonstrated the strengths and weaknesses
of the various decision-making approaches using an aircraft
selection problem and then presented a method based on the
concept of hypothetical equivalents andexpandedthe method
to include hypothetical in equivalents.
Ibusuki and Kaminski [6] suggested a methodology
for the product development process in an automotive com-
pany, aiming at the correct systematic approach of Value
Engineering (VE) and target-costing in cost management.
Their proposed approach was validated in a case study foc-
used on the engine-starter system of a vehicle, aiming at
improved product cost, functionality and quality accomplish-
ment, in accordance with customer needs and the company
strategy. AlbrittonandMcMullen[7] suggestedoptimal prod-
uct problem, where the best mix of product features was
formulated into an ideal offering and was optimized using
Ant Colony Optimization (ACO). Here, algorithm based on
the behavior of social insects was applied to a consumer deci-
sion model designed to guide new product decisions and to
allow planning and evaluation of product offering scenarios.
Chen and Weng [8] suggested Quality Function Deploy-
ment (QFD) toa product development process usedtoachieve
higher customer satisfaction. The engineering characteris-
tics affecting the product performance are designed to match
the customer requirements. Kim et al. [9] analyzed Automo-
tive Vehicle Design by using Target Cascading, a systematic
effort to propagate the desired top-level systemdesign targets
to appropriate specications for subsystems and components
in a consistent and efcient manner. Popovic [10] discussed
methods and techniques and used them for an ergonomic
evaluation of products, product interfaces and systems.
In order to assess and optimize the product design, an
objective evaluation framework is needed. Many methodol-
ogies have beendevelopedduringthe past decades bythe aca-
demia and industry, such as Concurrent Engineering, Robust
Design, Design for Manufacture and Assembly, Total Qual-
ity Development, etc. Some other methods and techniques
are being used and are given in the Table 1.
Table 1 Common evaluation methods and techniques
Evaluation
methods/Techniques
Purpose
CAD simulation models To evaluate design and its perceived use
during the different stages of design pro-
cess.
Checklists To dene operations of a product/system
and identify users needs.
Mock-up evaluation To evaluate product usage with users par-
ticipation
Motion studies To evaluate motion performances and
identify critical conditions
Protocol analysis To evaluate a design, users expertise level
and understand users concept of prod-
ucts.
Prototype evaluation To verify a design outcome under real con-
ditions.
Task-analysis To dene and evaluate operational proce-
dures of a human/product/system.
Interviewing users To identify users needs
Digraph and matrix approach has been used extensively
by a number of researchers in various engineering applica-
tions such as failure cause analysis [11], reliability evalua-
tion [12], proactive fault identication [13], TQMevaluation
[14, 15], risk mitigation [16] and in other areas also. Digraph
and matrix approach is a well systematic one which con-
siders all the inuencing factors (attributes) and its relative
importance of one factor on the other on the objective. This
consideration of relative importance is missing in all other
techniques discussed above.
This paper presents the analysis, evaluation and selection
of product design through design aspects using digraph and
matrix approach. There are many design aspects affect (fac-
tors) the design process and are called as design selection
attributes. They are given below:
1. Design for manufacturing
2. Design for assembly
3. Design for environment
4. Design for safety
5. Design for reliability
6. Design for maintenance
7. Design for aesthetic features
8. Design for economy
9. Design for ergonomics
The above said product design evaluation attributes are
listed based on the discussions in focus group (brain storm-
ing session) and from the literatures [1719]. A focus group
is a collection of individuals that has been brought together to
discuss a particular topic, issue or concern. This focus group
1 3
Analysis and evaluation of product design through matrix approach 15
Table 2 Value of product evaluation attributes (P
i
)
Quality measure of product evaluation Assigned value of product
attribute evaluation attribute
Exceptionally low 0
Extremely low 1
Very low 2
Below average 3
Average 4
Above average 5
Moderate 6
High 7
Very high 8
Extremely high 9
Exceptionally high 10
technique as one approach enables the designer to explore
user desires and needs.
In general these design aspects are innumerable and are
referred as design for X (DFX) where X represents a broad
variety of design considerations which inuence the design
selection and are referred as design selection attributes. All
these attributes and their importance on design selection are
explained in the forth coming section.
2 General methodology
A methodology for the evaluation and selection of product
design is suggested based on the digraph and matrix method.
The main steps of the methodology are:
1. Identication of product design evaluation attributes for
the given product design models and the relative impor-
tance among the attributes. Obtain the value for the
attributes (P
i
) by referring the Table 1 and/or question-
naire in the Appendix. Also obtain the value for relative
importance (r
i j
) by referring the Table 2.
2. Development of product designevaluationdigraphmodel
(PED) considering the identied attributes and their rel-
ative importance in step 1. The number of nodes is equal
to the number of attributes and direction will be decided
based on the relative importance (r
i j
).
3. Development of product design evaluation matrix (PEM)
for PED which will be a M M matrix with diagonal
elements as P
i
and off diagonal elements as r
i j
.
4. Obtaining product design evaluation function for the
matrix using Eq. (1).
5. Evaluation of product design evaluation index by substi-
tuting the value for P
i
and r
i j
in the product design eval-
uation function.
6. Arranging the product design in ascending order based
ontheir indexandselectingthe best product designwhose
PEI is highest.
3 Product design evaluation attributes
3.1 Design for manufacturing
DFM refers to the general engineering art of designing prod-
ucts in such a way that they are easy to manufacture. The
basic idea exists in almost all engineering disciplines, but
of course the details differ wildly depending on the manu-
facturing technology. Traditionally, DFM method evaluates
the feasibility and cost of manufacturing of the product at
the operation level. Bralla [20], Anderson [21], Corbett et al.
[22], and Boothroyd et al. [23] provide detailed discussions
on manufacturability and design. Design guidelines such as
those provided by Parmer and Laney [24], Singh [25], and
Fagade and Kazmer [26] are examples of DFM method. As
new DFX methods are explored, the definition of DFM has
expanded to become synonymous with DFX and concurrent
engineering. Various guide lines for the DFM are given in
the Appendix A.
3.2 Design for assembly
Design for Assembly (DFA) is a process by which products
are designed with ease of assembly in mind. If a product con-
tains fewer parts it will take less time to assemble, thereby
reducing assembly costs. In addition, if the parts are pro-
vided with features which make it easier to grasp, move,
orient and insert them, this will also reduce assembly time
and assembly costs. DFA guidelines adapted from several
sources such as Andreasen [27], Baldwin [28], DEC [29],
Huthwaite [30], Iredale [31] and Xerox [32].
DFM and DFA are most of the time considered simul-
taneously as DFMA. But they have conict nature between
them. DFMAis important for design because it has three ben-
ecial impacts. First and foremost, it reduces part count and
also reduces cost and time. For example Motorola introduced
DFMA and reduced its part count from 217 to 97 and there
by assembly time is reduced from 2,700 to 1,350s [17]. Sec-
ondly, DFMA increases reliability [33, 34]. Finally DFMA
increases the quality of the design. Various guide lines for
the DFA are given in the Appendix B.
3.3 Design for environment
The present problem of global warming is a challenging one
for the manufacturing industries. Design for environment
(DFE) is a product design approach for reducing the impact
of the products on the environment.
1 3
16 V. Paramasivam, V. Senthil
Products can have adverse impact on the environment dur-
ing manufacturing through the use of highly polluting pro-
cesses and consumption of large quantities of raw materials
and energy and disposal of waste. Because of these issues,
one must consider a products entire life cycle, from creation
todisposal. Inthis life cycle, there are manyevents of creating
pollution and many opportunities for recycling, remanufac-
turing, reuse, and reducing environmental impact. Designer
must bring all his ingenuity to bear on the challenging prob-
lem creating efcient products. Various guide lines for the
DFE are given in the Appendix C.
3.4 Design for safety
The goals of the design process are usually manifold. The
resulting system must not only satisfy its functional require-
ments but also have to fulll certain non-functional require-
ments. One of such requirements is safety. A safe product is
one that does not cause injury or property loss and does not
pollute the environment. There are three aspects to design for
safety.
1. Make the safe product.
2. If it is not possible to make the product inherently safe,
then design the proactive devices like guards, automatic
cutoff switches, pressure relief valves, to mitigate the
hazard.
3. If step 2 cannot remove all the hazards, then warn the
user of the product with appropriate warnings like labels,
ashing and loud sounds. Various guide lines for the
design for safety are given in the Appendix D.
3.5 Design for reliability
Reliability engineering is the discipline of ensuring that a
system will be reliable when operated in a specied manner.
Reliability engineering is performed throughout the entire
life cycle of a system, including development, test, produc-
tion and operation.
Reliability may be dened in several ways as:
Idea that something is t for purpose with respect to time.
Capacity of a device or system to perform as designed.
Resistance to failure of a device or system.
Ability of a device or system to perform a required func-
tion under stated conditions for a specied period of time.
Probability that a functional unit will performits required
function for a specied interval under stated conditions.
Reliability engineers rely heavily on statistics, probability
theory, and reliability theory. Many engineering techniques
are used in reliability engineering, such as reliability pre-
diction, Weibull analysis, thermal management, reliability
testing and accelerated life testing. Because of the large num-
ber of reliability techniques, their expense, and the varying
degrees of reliability required for different situations. Most
of the projects develop a reliability program plan to specify
the reliability tasks that will be performed for that specic
system. Various guide lines for the design for reliability are
given in the Appendix E.
3.6 Design for maintenance (Serviceability)
Serviceability is concerned with the ease with which main-
tenance can be performed on a product. Products often have
parts that are to be replaced at periodic intervals. It is impor-
tant to anticipate the required service operations during the
design of the product. Provision must be made for disas-
sembly and assembly. For example dont make an automo-
bile design that requires the removal of the panel to access
the oil lter. Also, remember that service usually will be
carried out in the eld where special tools and xtures
used in factory assembly are not available. The best way to
improve the serviceability is to reduce the need for service by
improving reliability of the components and systems. Vari-
ous guide lines for the design for maintenance are given in the
Appendix F.
3.7 Design for aesthetic features
Designers have many aesthetic qualities to improve the
marketability of manufactured products: smoothness,
shininess/reectivity, texture, pattern, curviness, color, sim-
plicity, usability, velocity, symmetry, naturalness, and mod-
ernism. The staff of the design aesthetics section focuses on
design, appearance and the way people perceive products.
Design aesthetics is interested in the appearance of products;
the meaningof appearance is studiedmainlyinterms of social
and cultural factors. The distinctive focus of the section is
research and education in the eld of sensory modalities in
relation to product. These elds of attention generate design
baggage that enables engineers to design products, systems,
and services, and match them to the correct eld of use. Var-
ious guide lines for the 7 design for aesthetic features are
given in the Appendix G.
3.8 Design for economy
It is important to note that cost is also a dimension of the
quality. Economically successful products are profitable; that
is, they generate more cumulative inows than the cumu-
lative outows.The main objective of all manufacturers is
1 3
Analysis and evaluation of product design through matrix approach 17
to make prot. So the producer should give a cost effec-
tive design. Based on the cost and other consideration, the
designer should select material or manufacturing processes.
Various guide lines for the design for economy are given in
the Appendix H.
3.9 Design for ergonomics
Ergonomics (or human factors) is the application of scien-
tic information concerning humans to the design of objects,
systems and environment for human use. (definition adopted
by the International Ergonomics Association in 2007). Ergo-
nomics is the study of the interaction of a person with a
machine. Information derived from Ergonomists contributes
to the design and evaluation of tasks, jobs, products, envi-
ronments and systems in order to make them compatible
with the needs, abilities and limitations of people (Interna-
tional Ergonomics Association in 2000). Ergonomics comes
into everything which involves people. Work systems, sports
and leisure, health and safety should all embody ergonomics
principles if well designed.(International Ergonomics Asso-
ciation in 2007). It is the applied science of equipment design
intended to maximize productivity by reducing operator
fatigue anddiscomfort. The eldis alsocalledbiotechnology,
human engineering, and human factors engineering. The role
of human factors in a product assumes importance in three
respects.
1. Man, as occupant of space that is to operate a machine,
the operator should have adequate space as dictated by
human body dimensions or anthropometry.
2. Man, as reader of display fromthe machine. That is based
on the display data, man processes the data and take
action.
3. Man, as one who takes action through operating controls
which form a part of the machine.
It will be obvious that human engineering in design should
consider application of forcess and study of displays and con-
trols. Various guide lines for the design for ergonomics are
given in the Appendix I.
Design for X (DFX), where X corresponds to one of doz-
ens of quality criteria which are listed earlier. This list is not
exhausted one and the same may be extended further dozens
of criteria. They may be specifically related to product or
processes. For example, design for handling , design for ow
ability (casting process), design for recycle ability, design for
remanufacturing, design for energy efciency and design for
regulations and standards, etc,.
1 4
3
2
9
8
7
6
5
Fig. 1 Product Evaluation digraph (PED) model
4 Product design evaluation digraph (PED) model
A digraph is used to represent the factors and their inter
dependences in terms of nodes and edges. In a digraph, direc-
tion is assigned to edges in the graph. Thus product design
evaluation attributes digraph consists of set of nodes, P =
{S
i
} with i = 1, 2. . ., M and set of edges, E = {r
i j
}. A node
S
i
represents i th product design evaluation attribute and the
r
i j
edge represents relative importance among the attributes.
The number of nodes M is equal to the number of product
design evaluation attributes considered for the selection of
design. If a node i has relative importance over another node
j in the selection then a directed edge (arrow) is drawn from
node i to j (i.e., r
i j
). If j has relative importance over i , then
the directed edge (arrow) is drawn fromnode j to i (i.e., r
j i
).
The most common representation of a graph is a diagram, in
which the vertices are represented as points and edges as line
segments joining the end vertices.
A universal product evaluation digraph model is
constructed (Fig. 1) by considering the attributes described
in the Sect. 2 as nodes and the relative importance between
the attributes as edges.
5 Product design evaluation matrix (PEM)
Since a digraph representation is a visual one, it helps to ana-
lyze the simple system having few attributes. When a sys-
tem becomes complex (i.e., when it has more attributes), the
digraph becomes clumsy and inconvenient for visualization.
To overcome these difculties, the digraph is represented in
matrix form.
Matrix representation of PED presents a one-to-one rep-
resentation. If a digraph has N factors, the matrix rep-
resentation will be a N N matrix, which considers all
1 3
18 V. Paramasivam, V. Senthil
the product design evaluation attributes and their relative
importance (r
i j
). The PEM is given in matrix 1., in which P
i
is the value of the i th attribute and r
i j
is the relative impor-
tance of i th attribute over the j th attribute. In general diago-
nal elements carry the value of attributes and the off diagonal
elements carry the value of the relative importance.
S =
_
_
_
_
_
_
_
_
_
_
_
_
_
_
P
01
r
12
r
13
r
14
r
15
r
16
r
17
r
18
r
19
r
21
P
2
r
23
r
24
r
25
r
26
r
27
r
28
r
29
r
31
r
32
P
3
r
34
r
35
r
36
r
37
r
38
r
39
r
41
r
42
r
43
P
4
r
45
r
46
r
47
r
48
r
49
r
51
r
52
r
53
r
54
P
5
r
56
r
57
r
58
r
59
r
61
r
62
r
63
r
64
r
65
P
6
r
67
r
68
r
69
r
71
r
72
r
73
r
74
r
75
r
76
P
7
r
78
r
79
r
81
r
82
r
83
r
84
r
85
r
86
r
87
P
8
r
89
r
91
r
92
r
93
r
94
r
95
r
96
r
97
r
98
P
9
_

_
Matrix 1. Universal product design evaluation matrix
6 Product design evaluation function (PEF)
It is a permanent function of standard matrix and is used
in combinatorial mathematics [35]. It is a desirable function
which interprets the design characteristics of the system. The
PEF is a function in which all system and their interdepen-
dence and structural complexity are accounted. Moreover,
PEF does not contain any negative sign and hence no infor-
mation is lost. The PEF is written in the form of equation as
per[S]. The PEF ensures a realistic estimate for the design
evaluation of product.
Per (S) =
m

i=1
Pi +

k
. . .

m
_
ri j r j i
_
Pk Pl . . . PM
+

k
. . .

m
_
ri j r j i rkl + ri k rkj r j i
_
Pm Pn . . . PM
+
_
_
_

k
. . .

m
_
ri j r j i
_
(rkl rlk ) Pm Pn . . . PM
+
_
_

k
. . . .

m
_
ri j r j k rkl rli + ril rlk rkj r j i
_
Pm Pn . . . PM
_
_
_
_
_
+
_
_

k
. . . .

m
_
ri j r j i
_
(rkl rlmrmk + rkmrml rlk ) Pn Po . . . PM
+

k
. . . .

m
_
ri j r j k rkl rlmrmi + ri mrlmrlk rkj r j i
_
Pn Po . . . PM
_
_
+ . . . .
(1)
The product design evaluation function Per(S) contains
terms arranged in (M + 1) groupings and these groupings
represent the presence of attributes and the relative impor-
tance loops. The rst grouping represents the presence of all
product design attributes. In general, the second grouping is
absent due to the absence self loop in product design attri-
butes digraph.The third grouping contains 2-attributes rela-
tive importance loops and the presence of (M 2) attributes.
Fourth grouping represents a set of 3-attributes relative
importance loops or its pairs and presence of (M 3) attri-
butes. The fth grouping contains two sub-groupings. The
terms of the rst sub-grouping are a set of two 2-attributes rel-
ative importance loops and the presence of (M4) attributes.
Each term of the second sub-grouping is a set of 4-attributes
relative importance loops or its pairs and the presence of
(M 4) attributes. The sixth groupings contain two sub-
groupings. The terms of the rst sub-grouping are a set of
2-attribues relative importance loops and 3 attributes relative
importance loops or its pairs and presence of (M 5) attri-
butes. The termof second sub-grouping is a 5-attributes rela-
tive importance loop or its pairs and the presence of (M 5)
attributes. Similarly other terms of the expression are dened.
Thus product design evaluation function characterizes the
product design as it contains all attributes and their relative
importances. This function will be useful for designers and
practicing engineers.
7 Product design evaluation index
Product design evaluation index (PEI) is a numerical value
denes the overall effectiveness of the product design with
respect to design aspects and provides the information of the
product design with respect to customer satisfaction. PEI is
obtained from the product design evaluation function (Eq. 1)
by substituting the values of attributes P
i
and the values of
relative importance between the attributes r
i j
. The values
for attributes P
i
are obtained from the data provided by the
industries and expertise. The value of many attributes cannot
be expressed as numerical values. If a quantitative value is
not available, then a ranked value judgment on a scale, e.g.,
from 0 to 10scales, is adopted. Table 2 represents the PEA
on a qualitative scale. It is seen that many of the attributes
are not easy to measure in terms of qualitative scale hence
a questionnaire has been designed to measure each attribute
in terms of weightage (questionnaire enclosed in appendix).
The numerical values obtainedthroughquestionnaire are nor-
malized on the same scale, i.e., 010. If P
i
has a range P
il
and P
i u
the value 0 is assigned to the lowest range value
(P
il
) and 10 is assigned to the highest range value (P
i u
). The
other intermediate values P
i i
of the product design attribute
are assigned values between 0 and 10 as follows:
P
i
= {10/P
i u
} P
ii
for P
il
= 0
P
i
= {10/(P
i u
P
il
)} (P
i i
P
il
) for P
il
> 0 (2)
Equation (2) is applicable for general benecial attributes
only. A benecial attribute i.e., design for safety means its
higher attribute values are more desirable for the given prod-
uct design, whereas a nonbenecial attribute i.e., design for
1 3
Analysis and evaluation of product design through matrix approach 19
economyis one for whichlower attribute values are desirable.
Therefore, in case of nonbenecial product design attributes,
the value 0 and 10 is assigned to the highest range value (P
i u
)
and the lowest range value (P
il
) respectively. The other inter-
mediate values P
i i
of the product designattribute are assigned
between 0 and 10 as follows:
P
i
= 10{1 (P
i i
/P
i u
)} for P
il
= 0
P
i
= {10/(P
i u
P
il
)} (P
i u
P
i i
) for P
il
> 0 (3)
For example, the attribute design for safety having the higher
value is much safer to use. So the value for the attribute
design for safety having higher value is benecial to cus-
tomers as well as manufacturers. So the quantitative value
obtained from the appendix for DFS is substituted in Eq. (2)
for normalising. On the other hand, the attribute design for
economy means total cost of the product involved in product
development. If the total cost is higher, both customer and
manufacturer will not be beneted. So the attribute DFE is
considered as non-benecial attribute and it should have min-
imum value. So the value abtained from appendix for DFE
is substituted in Eq. (3) for normalising.
The relative importance between two attributes (i.e., r
i j
)
for a given product design is also assigned a value on the scale
010 and is arranged into six classes. The relative impor-
tance implies that an attribute i is compared with another
attribute j in terms of its relative importance for the given
product design. The relative importance between i, j and j, i
is distributed on the scale 010 and is dened as
r
j i
= 10 r
i j
(4)
This means, a scale is adopted from 0 to 10 on which the
relative importance values are compared. If r
i j
represents
the relative importance of the i th attribute over the j th attri-
bute, then the relative importance of the j th attribute over
the i th attribute is evaluated using Eq. (4). For example, if
j th attribute is slightly more important than the i th attribute
then r
j i
= 6 and r
i j
= 4. Table 3 is used for assigning
r
i j
values using Eq. 4. The product design evaluation index
value for all design alternatives is evaluated using Eq. (1) by
substituting the values of Pis and r
i j
s. The product designs
can be arranged in the descending or ascending order of PEI,
to rank them for their performance. The product design, for
which the value of PEI is highest, is considered to be the best.
The value of the relative importance between the attributes
obtained using Table 3 is given in Table 4.
Example In order to select the optimum product design
model, it is necessary to evaluate all the models of certain
product. An example of kicker models of motor bike (prod-
uct) is considered for illustrating the procedure for obtaining
the PEI and is shown in Fig. 2.
For the example considered the above said steps are fol-
lowed for analyzing and for obtaining evaluation index.
Table 3 Relative importance of Product selection attributes
Class description Relative importance
of attributes
r
ij
r
ji=10
r
ij
Two attributes are of equal importance 5 5
One attributes is slightly more important
than other
6 4
One attributes is more important than other 7 3
One attributes is much more important
than other
8 2
One attributes is extremely more important
than other
9 1
One attributes is exceptionally more
important than other
10 0
Table 4 Value of relative importance
DFM DFA DFE DFS DFR DFMn DFAF DFEc DFEr
DFM *** 5 7 6 6 7 8 6 7
DFA 5 *** 7 6 6 7 7 6 7
DFE 3 3 *** 4 5 6 6 4 6
DFS 4 4 6 *** 5 7 7 5 6
DFR 4 4 5 5 *** 6 7 5 6
DFMn 3 3 4 3 4 *** 6 5 6
DFAF 2 3 4 3 3 4 *** 3 4
DFEc 4 4 6 5 5 5 7 *** 7
DFEr 3 3 4 4 4 4 6 3 ***
Step 1: Identication of the attributes
To simplify the illustration the following main attributes
are considered. They are
1. Design for manufacturing-P
1
2. Design for assembly-P
2
3. Design for maintenance-P
3
4. Design for economy-P
4
Step 2: Construction of digraph model
By considering the mentioned attributes a digraph model
is constructed and is shown in Fig. 3.
Step 3: Development of product design evaluation matrix
In this step, the attributes [P
1
, P
2
, P
3
, P
4
] are taken as
diagonal elements and the relative importance between the
attributes is taken as off diagonal elements.
S =
_
_
_
_
P
1
r
12
r
13
r
14
r
21
P
2
r
23
r
24
r
31
r
32
P
3
r
34
r
41
r
42
r
43
P
4
_

_
S
1
=
_
_
_
_
4.4 5 6 6
5 5.4 6 6
4 4 1.6 5
4 4 5 1.6
_

_
1 3
20 V. Paramasivam, V. Senthil
(a) (b)
Fig. 2 Various models of the product (kicker)
1 2
3 4
Fig. 3 Digraph model for the given example
S
2
=
_
_
_
_
3.8 5 6 6
5 4.2 6 6
4 4 1.2 5
4 4 5 1.2
_

_
Matrix 2. Product design evaluation matrix for the given
example.
Note: The values for the attributes are arrived by lling the
questionnaires given in the appendix.
Example to obtain the value for attribute is as follows:
1. Considering the attribute DFM and assigning the value
in the appendix against each guideline for the considered
design alternative.
2. Assigning the minimumpossible value for the each quide
lines of DFM and summing up all the values and assum-
ing as the P
il
(lower value of the i th attribute). For this
example P
il
= 25.
3. Assigningthe maximumpossible value for the eachquide
lines of DFM and summing up all the values and assum-
ing as the P
i u
(higher value of the i th attribute). For this
example P
i u
= 115
4. Assigning the value for the each quide lines of DFM of
the considered product design and summing up all the
values and assuming as the P
i i
(value of the attribute of
considered product design) For the product design given
in Fig. 3 P
i i
= 65
5. Substituting these values in the Eq. 2 to get the normal-
ized value on 010 scale.
P
i
= {10/(P
i u
P
il
)} (P
i i
P
il
) for P
il
> 0
= {10/(115 25)} (65 25)
= 0.111 40 = 4.4
Similarly the other attributes values are normalized.
Step 4: Obtaining product design evaluation function
In this step, product design evaluation function is obtained
which is nothing but the determinant of the given matrix. The
PEF is written by following the Eq. 1.
Per(S) = P
1
P
2
P
3
P
4
+ {(r
12
r
21
P
3
P
4
+ r
13
r
31
P
2
P
4
+ r
14
r
41
P
2
P
3
+ r
23
r
32
P
1
P
4
+ r
24
r
42
P
1
P
3
+ r
34
r
43
P
1
P
2
)} + {(r
12
r
23
r
31
P
4
+ r
13
r
32
r
21
P
4
+ r
12
r
24
r
41
P
3
+ r
14
r
42
r
21
P
3
+ r
13
r
34
r
41
P
2
+ r
14
r
43
r
31
P
2
+ r
23
r
34
r
42
P
1
+ r
24
r
43
r
32
P
1
)}
+ {(r
12
r
21
r
34
r
43
+ r
13
r
31
r
24
r
42
+ r
14
r
41
r
23
r
32
+ r
12
r
23
r
34
r
41
+ r
14
r
43
r
32
r
21
+ r
13
r
34
r
42
r
21
+ r
12
r
24
r
43
r
31
+ r
14
r
42
r
23
r
31
+ r
13
r
32
r
24
r
41
)}
(5)
Step 5: Evaluation of index
In this step, the value of the attributes and relative impor-
tance of one attribute over the other is substituted in Eq. (5)
for both the models.
1 3
Analysis and evaluation of product design through matrix approach 21
PEI(S
1
) = 60.82 + 1410.64 + 3120 + 5329
(Values are calculated for each subgroup)
= 9920.46
PEI(S
2
) = 22.92 + 895.8 + 2136 + 5329
= 8381.72
Step 6: Arranging the PEI in ascending order:
Hence from the indices (PEI) of the product models, it is
understood that the product in Fig. 3 is the best (i.e., whose
PEI is 9920.46).These values calculated by considering all
the design aspects and their relative importance.
8 Conclusion
In this paper, the methodology for the evaluation of the prod-
uct design has been developed using digraph and matrix
approach and the necessity of the product design evaluation is
emphasized. The digraph and matrix approach, which con-
sists of product design evaluation digraph, product design
evaluation matrix and product design evaluation function, is
applied for the motor bike kicker. The product design evalua-
tion digraph is found to be a visual representation of the char-
acteristics and their interdependences of the product design.
The methodology proposed for the analysis and evaluation
of a product design concludes the following:
Graph theory and matrix approach is applicable to any
product design consisting of any number of factors.
Application of matrix algebra is found to be convenient
both for visual and computer analysis.
Product design evaluation index represents the product
design characteristics and is useful in ranking the various
product models based on the design aspects.
Graph theory and matrix approach can be employed to
various problems of conicting nature and to the prob-
lems, where the quantitative data are not available i.e.,
material handling system evaluation, machine cell lay-
out analysis, supplier selection problem, vehicle routing
optimization etc.
Appendix I:
Procedure for lling the questionnaire: 1. All positive con-
tributing criteria are given values from 0 to 5 in ascend-
ing order. For example more simplicity the value 5 will be
assigned and for least simplicity the value 0 will be assigned.
2. All least contributing criteria are given values from 5
to 0 in descending order. For example closer tolerance the
value 0 is given and for wider tolerance the value 5 will be
assigned.
A: Design for Manufacturing
0 1 2 3 4 5
Simplicity
Standard components
Process ability of material
Common parts and materials
Parts ease for fabrication
Liberal/tight Tolerance
Utilizing special process
Manufacturing sequence
Minimize number of parts
Shape of the parts
Easy for inspection
Robustness of design
Use common parts
Parts to be multifunctional
Avoid tight tolerance
Avoid secondary operation
Use common datum
Minimum weight
Use general purpose tooling
Many operation in one setting
Compact design
Foldable and stackable design
Use of available stock
B: Design for Assembly
Number of part count
Modularize multiple parts
Assemble space
parts identify oriented for insertion
Prefer self-locating parts
Standardize to reduce part variety
Maximize part symmetry
Design in geometric or weight polar
properties if nonsymmetrical
Eliminate tingly parts
Color-code for parts having similarity
Prevent nesting of parts
Mating features for easy insertion
Provide alignment features
Insert new parts into an assembly from above
Eliminate re-orientation of parts
Eliminate fasteners
Provide flats for uniform fastening
Space between fasteners& features
Prefer easily handled parts
Weight of the components
Foolproof the assembly design (poke-yoke)
Prefer self-locating parts
Avoid flexible and fragile parts.
C. Design for environment
Material recovery and reuse
Design for disassembly
Product waste minimization
Waste recovery and reuse
Packaging recovery
Use of hazardous or undesirable material
Hazard to production workers
Noise level(80 decibels accepted limit)
Disposal of waste
Recyclable material
Restricted material
D. Design for safety (DFS)
Robustness
Error feed back signal
Multiple trigger for safe
Nonflammable materials
1 3
22 V. Paramasivam, V. Senthil
Condition monitoring system
Avoid sharp corner in parts
Avoid projection in parts
Rotational parts to be covered
Alarm to toggle from the danger
Sign post in dangerous operations
Handling slippery parts
Provision for handling the par
Optimal feedback
Failure feedback
Safety lacks
E.Design for reliability (DFR)
Fail safe approach(weak spots monitoring)
Replace the components in life period
Increase the reliability of the components
Design with higher factor of safety
Test under maximum load condition
Test for real condition
Improve design for failed components
Define the definite working condition
F. Design for maintenance (DFMn)
Easily visible and accessible
Easy removal of Cover, housing, panels
Avoid pres fits, bonding, riveting,
Modular design is great boon.
On-line testing possibilities
Indicator for maintenance
Standard spares
Compatibility with maintenance tools
G. Design for aesthetic features (DFAF)
Smoothness
Shininess/reflectivity
texture,
curviness
color
simplicity
usability
symmetry
modernism.
structural integrity
Appearance
Handy
Pride
H. Design for economy (DFEc)
Design and development cost
Material cost
Production cost (for all operations)
Assembly cost
Cost of the fasteners
Other over head expenses
I. Design for ergonomics (DFEr)
? Space for assembly
? Orientation
? Position of operator during assembly
? Fatigue of the operator
? Position of the control panel
? Surrounding of the operator
? Operator work volume
? Operator safety
? Automation when possible
References
1. Maddulapalli, A.K., Azarm, S.: Product selection with preference
and attribute variability for an implicit value function. J. Mech.
Des. 128, 10271037 (2006)
2. Besharati, B., Luo, L., Azarm, S., Kannan, P.K.: Multi-objective
single product robust optimization: an integrated design and mar-
keting approach. Trans. ASME 128, 884892 (2006)
3. Vanegas, V.L., Labib, A.W.: Fuzzy approaches to evaluation in
engineering design. Trans. ASME 127, 2433 (2005)
4. Li, H., Azarm, S.: Product selection under uncertainty and with
competitive advantage. J. Mech. Des. 122, 411417 (2000)
5. See, T.-K., Gurnani, A., Lewis1, K.: Multi-attribute decision
making using hypothetical equivalents and inequivalents. Trans.
ASME 126, 950958 (2004)
6. Ibusuki, U., Kaminski, P.C.: Product development process with
focus on value engineering and target-costing: A case study in an
automotive company. Int. J. Prod. Econ. 105(2), 459474 (2007)
7. Albritton, M.D., McMullen, P.R.: Optimal product using a colony
of virtual ants. Eur. J. Oper. Res. 176(1), 498520 (2007)
8. Chen L.-H., Weng M.-C.: An evaluation approach to engineering
design in QFD processes using fuzzy goal programming models.
Eur. J. Oper. Res. 172(1):230248 (2006)
9. Kim, H.M., Rideout, D.G., Papalambros, P.Y., Stein, J.L.: Analyt-
ical target cascading in automotive vehicle design. J. Mech. Des.
125, 481489 (2003)
10. Popovic, V.: Product evaluation methods and their applications.
School of Architecture, Interior and Industrial Design, Queens-
land University of Technology, GPO Box 2434, Brisbane 4001,
Australia
11. Roa, R.V., Gandhi, O.P.: Failure cause analysis of machine tools
using digraph and matrix methods. Int. J. Mach. Tools Manuf. 42,
521528 (2002)
12. Sehgal, R., Gandhi, O.P., Angra, S.: Reliability evaluation and
selection of rolling element bearings. Int. J. Reliab. Eng. Syst. Saf.
68, 3952 (2000)
13. Yu, Y., Liu, Q., Tan, L.: A graph-based proactive fault identica-
tion approach in computernetworks. J. Comput. Commun. 28, 366
378 (2005)
14. Grover, S., Agrawal, V.P., Khan, I.A.: A digraph approach to TQM
evaluation of an industry. Int. J. Prod. Res. 42(19), 40314053
(2004)
15. Kulkarni, S.: Graph theory and matrix approach for performance
evaluation of TQM in Indian industries. TQM Mag. 17(6), 509
526 (2005)
16. Faisal, M.N., Banwet, D.K. Ravishankar: Quantication of risk
mitigation environment of supply chains using graph theory and
matrix methods. Eur. J. Ind. Eng. 1, 122139 (2007)
17. Otto, K., Wood, K.: Product. Pearson Education (2004)
18. Ulrich, K.T., Eppinger, S.D.: Product and Development. Mc-Graw
Hill Higher Education, New York (2000)
19. Chitale, A.K., Gupta, R.C.: Product and Manufacturing. Prentice
Hall of India(P) Ltd, New Delhi (2001)
20. Bralla, J.G.: Design for Manufacturability Handbook. McGraw-
Hill Professional, New York (1986)
21. Anderson, D.M.: Design for Manufacturability. CIMPress, Lafay-
ette (1990)
22. Corbett, J., Dooner, M., Meleka, J., Pym, C.: Design for Manu-
facture, Strategies, Principles and Techniques. Addison-Wesley,
Workingham (1991)
23. Boothroyd, G., Peter, D., Winston, K.: Product Design for Manu-
facture and Assembly, 2nd edn. Marcel Dekker, New York (2002)
1 3
Analysis and evaluation of product design through matrix approach 23
24. Parmer, C., Steve, L.: DFM&T guidelines for complex PCBs. Surf.
Mt. Technol. 7(7), 2931 (1993)
25. Singh, K.: Mechanical Design Principles: Applications, Tech-
niques and Guidelines for Manufacture. Nantel Publications,
Melbourne (1996)
26. Fagade, A., Kazmer, D.: Economic design of injection molded
parts using DFM guidelinesa review of two methods for tool-
ing cost estimation. In: Annual Technical ConferenceANTEC.
Conference Proceedings, pp. 869873 (1998)
27. Andreasen, M.: Design for Assembly. IFS Publications Ltd/
Springer, UK (1983)
28. Baldwin, S.: How to make sure of easy assembly. Tool Manufac-
turing Engineering (1966)
29. Digital Equipment Corporation (DEC): Design for Assembly:
General Principles Applied to Power Design/Manufacture. Power
System Business Unit, Digital Equipment Corporation, USA
(1993)
30. Huthwaite, B.: How to reduce part counts. Des. News 154 (1990)
31. Iredale, R.: Automatic assembly-components and products. Metal
Working Production, 8 April (1964)
32. Xerox Corporation: Generic Tool Scoping Annual. RMO/QA/
PQE/AME, 2nd edn. Xerox Corporation, USA (1986)
33. Bralla, J.G.: Design for Manufacturability Handbook. McGraw-
Hill Professional, New York (1991)
34. Barkan, P., Hinckley, C.: The benets and limitations of structured
design methodologies. Manuf. Rev. 6(3), 211220 (1993)
35. Jarkat, W.B., Rayser, K.J.: Matrix factorization of determinants and
permanents. J. Algebra 3, 127 (1996)
1 3