Anda di halaman 1dari 19

Decision Support Mechanism

Using DEA, GAs Benefit-Cost Analysis and TOPSIS


ISSN 2319-9725

Prof. Yogesh D. Bagul(Author)


Mechanical Engineering Department
Sanghavi College of Engineering,
Nashik, India

Abstract: Small Scale Industry (SSI) contributes to as good as 40% of export of India and contribute
50% to Indias GDP. National Industry Policy and National Manufacturing Policy of India,
therefore, lays emphasis on SSI promotion through such means as tax exemption, sanction of loans,
providing infrastructure, etc. In order to excel in export promotion and to capture a large chunk of
Indian market. SSI needs to go for advanced manufacturing technology like FMS, Robotics,
automation, etc. CNCs are, therefore, the right machine tool for SSI. CNCs are capital intensive,
however. SSI being small organizations having no expertise in procurement of such capital intensive
units needs some way out to overcome this roadblock.
The present paper is an attempt in this direction namely to develop a Decision Making Support
Mechanism for SSIs to procure capital intensive machine tools. The data of 21 CNC is is taken from
a paper Comparing CNC Lathes Using Data Envelopment Analysis (DEA)[1]. The data is subjected
to Data Envelopment Analysis and Genetic Algorithms obtaining a set of solutions. The set thus
obtained is subjected to Benefit-Cost analysis (BCA). The weighting for 3 major and 13 sub-factors
are finalized in consultation with a committee set up by management for this purpose. TOPSIS
analysis is also applied to the set of selected candidates (CNC machines). The work done by us
provides a good take off for the industry, a fruitful Decision Support System (DSS) approach. The
approach adopted in this paper i.e. subjecting data to DEA, GAs, Benefit-Cost analysis and
validation through TOPSIS is found to be satisfactory and acceptable to the industry. The approach
or methodology developed in this paper will be useful to SSI manufacturing units and can be
extended to the service sector as well.

December, 2013

www.ijirs.com

Vol 2 Issue 12

1. Introduction:
The small-scale industries (SSI) sector plays a vital role in the growth of the country. It
contributes almost 40% of the gross industrial value added in the Indian economy. It has been
estimated that a million Rs. of investment in fixed assets in the small scale sector produces
4.62 million worth of goods or services with an approximate value addition of ten percentage
points [2]. By its less capital intensive and high labor absorption nature, SSI sector has made
significant contributions to employment generation and also to rural industrialization.
Government of India under her NMP has introduced some promotional measures for SSI like
reserving over 294 items for SSI, investment Rs 100 lakhs, no duty on clearances of excisable
goods up to Rs. 100 lakhs [3], etc. Such and similar provisions have been encouraging SSIs to
move towards WCM, encouraging them to go for capital intensive tools like CNC,
CAD/CAM, robotics, application soft-wares, etc. Cost efficient purchase of such tools is one
of the most important aspects of manufacturing management leading towards more
productivity and sustainability. The present paper deals with how to arrive at the optimal
purchase model on similar lines of DEA philosophy, in case of CNC lathes using one input
(unit cost) with five outputs: spindle max speed, spindle speed range, number of tool position,
range of rapid traverse X-axis, range of rapid traverse Z-axis). The efficiencies obtained with
this input-output set will be referred to as Cost efficient . The present paper is organized as
follows, Chapter 2 deals with the problem solving using DEA, Chapter 3 gives the solution
obtained by Genetic Algorithms, Chapter 4 deals with result analysis and discussion chapter 5
gives the conclusion and then references.

2. Data analysis using dea approach:


The data of 21 CNC lathes named as Decision making units (DMU) which shows input and
output parameters for each CNC lathes is shown in TABLE I. The data of TABLE I is
subjected to Data Envelopment Analysis, the result of analysis using DEA yields the result in
TABLE II. The TABLE II shows that only 4 DMUs (Decision making units) namely DMU1,
DMU4, DMU8 and DMU15 has 100% efficiency i.e. they are all efficient and all the other
DMUs have efficiency score less than 100% so are not efficient [1]. TABLE II gives a
decision maker 4 alternatives with same efficiency score i.e. DEA reduces the 21 alternatives
to 4 alternative beyond which selecting the best alternative has to be subjected to further
analysis. The final set of solution presented by DEA is given in TABLE III

International Journal of Innovative Research and Studies

Page 647

www.ijirs.com

December, 2013

Vol 2 Issue 12

3. Data Analysis Using Gas Approach:


The data of 21 DMUs [1] is subjected to further analysis using Genetic Algorithms (GAs).
Genetic Algorithms are the heuristic search and optimization techniques that mimic the
process of natural evolution. The principles of natural selection are Select the best and
discard the rest.GA is very appealing for single and multi-objective optimization problems.
GA uses the selection, crossover & mutation as operators for solving the optimization problem
[4].
3.1. Decision variables:
The decision variables or process parameters chosen for the optimization of the process are:
i.

Money X1 Input

ii.

Spindle max. speed Y1 Output

iii.

Spindle speed range Y2 Output

iv.

No. of tool positions Y3 Output

v.

Rapid traverse in X direction Y4 Output

vi.

Rapid traverse in Z direction Y5 Output


INPUT

OUTPUTS

SPINDL
CNC LATHE

( DMU'S )
MONEY

E MAX.
SPEED
(RPM)

SPINDL NO.
E

OF

TOOL

SPEED POSITIO
RANGE NS

RAPID

RAPID

TRAVERS TRAVERS
E X- AXIS E Z- AXIS
(m/min)

(m/min)

(Y3)

(Y4)

(Y5)

5950

24

24

15,50,000 3500

3465

20

20

YCM-TC-15 (DMU3)

14,00,000 6000

5950

12

15

20

VTURN 16 (DMU4)

11,00,000 6000

5940

12

12

15

FEMCO HL-15 (DMU5)

12,00,000 6000

5940

12

12

16

FEMCO WNCL - 20 (DMU6) 15,00,000 3500

3465

12

12

(RPM)

(NOS.)

(Y1)

(Y2)

YANG ML - 15A (DMU1)

12,00,000 6000

YANG ML - 25A (DMU2)

(X1)

International Journal of Innovative Research and Studies

Page 648

www.ijirs.com

December, 2013

Vol 2 Issue 12

FEMCO WNCL -30 (DMU7) 26,00,000 4000

3960

12

12

16

EX-106 (DMU8)

13,20,000 5000

4950

12

24

30

ECOCA SJ20 (DMU9)

11,80,000 4500

4480

24

24

ECOCA SJ 25 (DMU10)

15,50,000 4000

3950

12

15

20

ECOCA SJ30 (DMU11)

16,00,000 3500

3450

12

15

20

TOPPER TNL -85A (DMU12) 12,00,000 3500

3465

20

24

14,50,000 3000

2970

20

24

14,00,000 3000

2970

12

24

30

TOPPER TNL -85T ( DMU15) 13,50,000 3500

3465

12

30

30

14,50,000 3000

2970

12

20

24

(DMU17)

15,20,000 2500

2475

12

20

24

ATECH MT - 52S (DMU18)

13,76,000 4800

4752

12

20

24

ATECH MT - 52L (DMU19)

14,40,000 4800

4752

12

20

24

ATECH MT - 75S (DMU20)

18,24,000 3800

3970

10

12

20

ATECH MT - 75L ( DMU21) 19,20,000 3800

3970

10

12

20

TOPPER

TNL

100A

(DMU13)
TOPPER

NL

100AL

DMU14)

TOPPER TNL - 100T (DMU


16)
TOPPER

TNL

120T

Table 1: The DATA [1]

International Journal of Innovative Research and Studies

Page 649

www.ijirs.com

December, 2013

Vol 2 Issue 12

Table 2: Efficiency Analysis Using Dea[1]

Sr.

DMU

No

no.

Mone

Efficiencies

Spindle Speed
speed

Range

(RPM)

(RPM)

Rapid traverse
Tool
X

Positions

Direction

Y
Direction

595
1

100%

12,00,000

6000 0

24

24

12

12

15

12

24

30

12

30

30

594
2

100%

11,00,000

6000 0
495

100%

13,20,000

5000 0
346

15

100%

13,50,000

3500 5

Table 3: Final solution using DEA


3.2. Objective function
The objective is to minimize the money. Thus the objective function (worked out using
Minitab 15 software) is to Minimize Z1 =1972678 - 8624 Y1 + 8573 Y2 + 51451 Y3 - 10493
Y4 - 14776 Y5
3.3. GAs Methodology [5]
The procedure adapted involves the following steps
International Journal of Innovative Research and Studies

Page 650

December, 2013

www.ijirs.com

Vol 2 Issue 12

Step 1: Population size selected is 5. The initial solution is binary coded using random
numbers. The string length chosen is 6.
Step 2: The objective function is then calculated from the model developed for each solution
of population.
Step 3: The new fitness function is then defined by the incorporating adaptive penalty
parameter as shown in the equation
F(x) = f(x) + kgi(x)
Step 4: The GA operators reproduction, crossover and mutations are now applied to these
newly defined fitness functions and all the above mentioned steps are repeated till optimum
solution is obtained.
Reproduction is performed by using random numbers in such a way that fittest strings are
selected with largest probability as in roulette wheel selection. Then crossover is performed
using scattered method with the help of random numbers. Mutation is performed with
probability of 0.03. The crossover operation is performed with fraction of 0.33.
The optimum result obtained after 101 generation is
Y1 = 6000.000,
Y2 = 5950.000,
Y3 = 12.000,
Y4 = 16.085,
Y5 = 19.95,
X1 = 1366804.100
This corresponds to the DMU 5.
The results obtained by GAs and DEA do not match hence the analysis has to be continued
further. The next chapter presents the result analysis and discussion using Cost-Benefit
analysis and TOPSIS.

4. Result analysis and discussion:


The above solutions of DEA and GAs are subjected to Benefit Cost Analysis to arrive at the
final acceptable solution. The next section deals with the aspects of Benefit Cost analysis.
4.1.Benefit-Cost Analysis:
In engineering economics, often cost-benefit analysis is used. However, a few researchers
prefer benefit-cost analysis for the higher the ratio higher the benefits and higher ranking.
Groover and Zimmers have presented a lucid explanation how to go about selecting software
International Journal of Innovative Research and Studies

Page 651

December, 2013

www.ijirs.com

Vol 2 Issue 12

[6]. They have introduced some major factors contributing towards benefits of a unit to be
procured with appropriate weightage to each factor. They suggests that rather than a single
person attempting to develop criteria for selection a purchase committee should be setup for
this purpose. The committee shall evaluate the alternative systems and make recommendation
for vendor selection. The purchase committee selected for this purpose should include the
members from design engineering, production department, purchase department and accounts
department. The committee should develop a comprehensive list of applicable criteria for
purchase. The TABLEIVmust be compiled on the basis of reference of their decisions. The
analysis is divided in to 3 major factors namely, General consideration, Applications
Electricals and Applications Mechanical all the considerations are allotted 1000 points each.
Then each factor is further divided into sub factors as General consideration in 5 sub factors as
Cost, Services, Quality, Delivery and logistics support and programming etc. and the 1000
points allotted to each factor is subdivided into these sub factors as decided by the purchase
committee. For example cost is allotted 200 points, services 100 points, quality 300 points
delivery and logistic support 100 points and programming 300 points. The sub factors are
further divided into sub-sub factors as the sub factor cost is divided into hardware, software,
maintenance, spares, transportation, field support etc. and the sub-sub factors are also allotted
points. This allocation may be varied from situations to situations. Similarly all the factors,
their sub factors and their points allocation are completed as shown in Table IV. All the
alternatives given by DEA and GAs are subjected to Benefit-Cost analysis. All the DMUs are
given points.

A. General Consideration

Max.

DMU DMU

DMU DMU

Marks

15

DMU 3

COST (200)
A.

Hard ware

50

20

20

30

20

30

B.

Software

30

10

20

20

30

30

C.

Maintenance

30

30

20

30

30

30

International Journal of Innovative Research and Studies

Page 652

December, 2013

www.ijirs.com

Vol 2 Issue 12

D.

Spares

30

20

20

20

30

10

E.

Transportation

20

10

20

20

10

10

F.

Field Support

40

10

10

10

200
2

SERVICES (100)
A.

Contract - 1yr, 2 yr, 5 yr

30

10

10

20

20

10

B.

Parts Location

20

20

20

20

20

20

10

10

30

20

20

20

20

10

C.

D.

Warrantee and
discontinuance clause
Software service

100
3

QUALITY (300)
A.

Reliability

40

20

40

40

20

20

B.

Simultaneous operation

40

40

40

20

40

40

C.

Crash recovery

40

40

20

40

20

D.

Power loss recovery

40

40

40

20

20

40

20

20

20

20

40

20

20

40

20

E.

F.

Environmental
sustainability
Human factor
consideration

G.

Output devices speed

20

20

20

20

20

20

H.

Product documentation

20

20

20

20

20

20

I.

Training

40

40

40

40

20

40

International Journal of Innovative Research and Studies

Page 653

www.ijirs.com

December, 2013

Vol 2 Issue 12

300

DELIVERY AND LOGISTICS


SUPPORT (100)
A.

Packaging

10

10

10

10

10

B.

Installation

20

20

10

20

10

20

C.

Pre-delivery inspection

20

20

20

20

20

20

D.

Installation guide

10

10

10

10

10

10

E.

Revisions

10

10

10

10

10

10

F.

Billing

10

10

10

10

10

10

G.

Proprietary agreement

20

20

20

20

20

100

100

100

100

100

100
5

PROGRAMMING (300)

A.

High level vendor


language

B.

Standard language

50

50

30

50

50

40

C.

Assembly language

50

50

50

50

50

D.

User protection

50

50

30

50

50

50

E.

Diagnostic aids

50

50

30

50

50

50

610

850

750

700

300
TOTAL 790
B. Application - Electricals (15 Marks )
Max

DMU DMU

DMU DMU

Marks

International Journal of Innovative Research and Studies

15

DMU3

Page 654

December, 2013

www.ijirs.com

Vol 2 Issue 12

ELECTRICAL DESIGN
APPLICATIONS (500)
A. Logic diagram

100

50

50

100

50

100

B. Circuit diagram

100

100

100

100

100

100

C. Electrical wirings

200

200

100

50

D. Bill of materials

100

100

100

100

100

100

500
2 NC PACKAGES (300)
A. Auto placement

50

50

50

50

50

B. Cable viewing

200

200

100

200

150

150

C. Wire Wrap

50

50

50

50

300
3 DOCUMENTATION (200)
A. Handbooks

80

60

50

70

80

80

B. Organization chart

40

40

40

C. Flow diagrams

40

20

20

40

40

D. PERT charts and scheduling

40

30

30

40

550

930

580

800

200
TOTAL 630
C. Applications - Mechanicals ( 35 Marks )
Max.

DMU DMU

DMU DMU

Marks

15

DMU3

1 MECHANICAL DESIGN (400)


International Journal of Innovative Research and Studies

Page 655

December, 2013

www.ijirs.com

Vol 2 Issue 12

A.

Process Diagrams

80

60

80

80

80

60

B.

Architecherial Drawing

20

20

20

20

C.

Plant layout

30

D.

Pipe layout

30

15

E.

Sheet metal Design

10

10

F.

2D drawings

10

10

10

10

10

10

G.

3D drawings

10

10

10

10

10

H.

Tools selection charts

80

80

80

80

80

80

I.

Gigs and Fixtures

80

40

80

80

80

80

J.

Design Data

50

50

50

50

50

50

400

MECHANICAL INTERFACES
(250)
A.

Stress analysis

50

50

50

B.

Finite element modeling

50

C.

Flow analysis

50

50

50

50

50

50

D.

Mechanical analysis

50

50

50

50

E.

Material Properties analysis

50

250

MFG PLANNING PACKAGES


(350)
A.

Machining NC programming

50

50

20

40

40

40

B.

Automatic NC programming

150

150

100

150

150

100

International Journal of Innovative Research and Studies

Page 656

www.ijirs.com

December, 2013

Vol 2 Issue 12

Functions
C.

Other NC Functions

50

D.

NC default values

50

50

50

50

50

E.

NC interfaces

50

30

20

50

20

30

670

695

690

610

350
TOTAL 575

Table 4: Questioner of Cost- Benefit Analysis


The total of all the factors and sub factors for each DMU is calculated which is named as total
term shown in TABLE V.The truncated price is calculated for each DMU so as to make the
calculation of Benefit-Cost Analysis more simplified. Then the ratio of Benefit to Cost is
calculated. Highest the ratio greater is the ranking of the DMU.
DMU'S
1

15

Total Term Score

1995

1830

2475

2020

2110

Price

1200000 1100000 1320000 1350000 1400000

Trunketed Price

120

110

132

135

140

Benefit/Cost Ratio

16.625

16.636

18.75

14.962

15.071

Rank

Table 5: Questioner of Cost- Benefit Analysis

4.2. TOPSIS approach [7]:


TABLE V shows a set of five machine tools from which a single machine needs to be
recommended these 5 machines are further subjected to TOPSIS (Technique for order
performance by similarity to ideal solution) approach. TOPSIS is one of the latest
approachesintroduced,essentially it includes such considerations as formation of virtual DMU.
Following is the TOPSIS methodology,
International Journal of Innovative Research and Studies

Page 657

www.ijirs.com

December, 2013

Vol 2 Issue 12

Ideal point is called DMU+ and anti-ideal point is called DMU-. Suppose { x1+ , x2+ , ,..xm+}
and { y1+ , y2+ , ,..ys+ } represents the input and output vectors of DMU+

and { x1- , x2- , ,..xm-

} and { y1- , y2- , ,..ys- } represents the input and output vectors of DMU-

respectively.

Consider the following relations


xi+ =

, i=1,.m

xij

yr+ =

ij

, r=1,.s

xi- =

xij

, i=1,.m

yr- =

yij

, r=1,.s

.. (1)

DMUjs distance to the ideal point is calculated as follows

.(2)

DMUjs distance to the anti-ideal point is calculated as follows

.. (3)

Now DMUj s general distance to the ideal and anti-ideal point is calculated as
.. (4)
The larger value of Dj represents the better value of DMUj compared to other DMUs
In this method, first the ideal point and the anti-ideal point are obtained using relation (1), all
the efficient DMUs are identified, and then obtained results are put in E = {1, . . . , e}.
Continuing with all the efficient DMUs, we produced a virtual DMU. The acquired inputs and
Outputs of the virtual DMU are the average of the corresponding inputs and outputs of all the
efficient DMUs, respectively and it is called DMUM(DMUMis the center of gravity of all the
efficient DMUs). After that, general distance of DMUMto the ideal point and the antiIdeal point are obtained through applying relation (9), this distance is called DM. In the next
step DMUq, q E, is deleted from the efficient frontier and the new virtual DMU is created
which is the average of all corresponding inputs and outputs of all the efficient DMUs
respectively expect DMUq. This is called

. After that general distance to

to the

ideal and anti-ideal point is obtained using relation (4). Then the true difference between Dm
and Dq/will be computed and named as
the distance

. The method is repeated for every DMUq, q E,

is the criteria for ranking efficient DMUs larger the

better is the ranking

DMUqwill have.
The sample calculation for DMU5 of the chosen data is shown below
International Journal of Innovative Research and Studies

Page 658

www.ijirs.com

December, 2013

Vol 2 Issue 12
Rapid

Sr.

DMU

No

Efficienc

Money

ies

Spindle

Speed

Tool

traverse

speed

Range

Positi

(RPM)

(RPM)

ons

Directi

Directi

on

on

100%

12,00,000

6000

5950

24

24

100%

11,00,000

6000

5940

12

12

15

100%

13,20,000

5000

4950

12

24

30

15

100%

13,50,000

3500

3465

12

30

30

14,00,000

6000

5950

12

15

20

6370000

26500

26255

56

105

119

1274000

5300

5251

21

23.8

Input Xi

Output Yi

1100000

6000

5950

12

30

30

1400000

3500

3465

12

15

-126000

1800

1786

3.2

8.8

158760000

324000

31897
10.24

81

77.44

-0.8

-9

-6.2

0.64

81

38.44

85.98
5

%
TOTA
L
Avera
ge

Ideal

11.2

AntiIdeal
DMU
D+

Square 00

0
158760000

Total

00

Subtra
ct

96
642996

5
158695700

35
125974.48
SqRt

17

D-

174000

-700

-699

302760000
Square 00

48860
490000

International Journal of Innovative Research and Studies

Page 659

www.ijirs.com

December, 2013

Vol 2 Issue 12

302760000
Total

00

Subtra
ct

978721
302750212

79
173997.18
SqRt

Dm

76

0.580045
Rapid

Sr.

DMU

Efficienc

No

no.

ies

Money

Spindle

Speed

Tool

traverse

speed

Range

Positi

(RPM)

(RPM)

ons

Directi

Directi

on

on

100%

12,00,000

6000

5950

24

24

100%

11,00,000

6000

5940

12

12

15

100%

13,20,000

5000

4950

12

24

30

15

100%

13,50,000

3500

3465

12

30

30

20500

20305

44

90

99

( DMU5 )
TOTA
L

4970000
Avera

5076.2

ge

Ideal

1242500

5125

Input Xi

Output Yi

1100000

6000

1350000

3500

24.7
11

22.5

5950

12

30

30

3465

12

15

-1

-7.5

-5.25

AntiIdeal

DMUQ1

D+

142500

-875

873.75

203062500

76343

Square 00

765625
203062500

Total

00

56.25

27.5625

152914
9

International Journal of Innovative Research and Studies

Page 660

www.ijirs.com

December, 2013
Subtra

Vol 2 Issue 12

203047208

ct

51
142494.63
SqRt

45
1611.2

D-

-107500

1625

115562500

264062

Square 00

5
115562500

Total

00

Substr

10.
3

110.25

9.75

25961
27

95.0625

523696
6

115510130

act

34
107475.63
SqRt

93

Dq1

0.429954

Dq1-

0.150092
Table 6: Sample calculation for TOPSIS

The solution of TOPSIS is presented below


Sr.
No.

DMU
Dq-

no.

Rank

0.2217
1

1
0.1676

2
0.0967

4
0.0967

15

5
0.1500

Table 7: Final Ranking by Topsis


International Journal of Innovative Research and Studies

Page 661

December, 2013

www.ijirs.com

Vol 2 Issue 12

Thus we have got 3 solutions to be considered finally namely;


i.

DMU8 through DEA

ii.

DMU3 through GAs

iii.

DMU1 through TOPSIS

4.3. Discussions:
In the preceding section three candidates are the winners namely DMU1, DMU3 and DMU8.
The following considerations are taken into account for final selection .DMU 3 has got the
second lowest benefit cost ratio and the highest cost. But it rank is 3rd in TOPSIS analysis.
However, TOPSIS does not consider some vital factors like warrantee period, after sales
services, vendor rating etc. hence DMU 3 is dropped for further consideration.
TOPSIS and BCA based DEA generates DMU1 and DMU8 as elements of final set of
solutions. DMU8 has got higher Benefit-Cost ratio of 18.75 (See TABLE V). Moreover
TOPSIS in case of DMU1 does not take care about such factors as warrantee period, vendor
rating, quality of product, no. of tool positions, rapid traverse etc. Therefor even though the
DMU8 is costlier by 1, 20,000.00 than DMU1, the DMU 1 is dropped and DMU 8 has
emerged as the final solution recommended by the committee. Thus the winner candidate is
DMU 8 i.e EX 106.

5. Conclusion:
Based on the data for 21 CNC lathes the DEA approach is used to obtain a set of 4 competitive
machine tools with 100% efficiency. Further the same data of 21 CNC lathes is used for GAs
approach. Thus a set of 5 competitive candidates are worked out and subjected to benefit cost
analysis. The benefit analysis 3 major factors with 13 sub factors are identified and assigned
weightages to various sub factor the total weightages being 1000 and the BCA analysis and
one candidate out of 5 is recognized as winner candidate. The 5 competitive candidates are
again subjected to TOPSIS analysis using Microsoft Excel and one winning candidate is
worked out.The output of cost benefit analysis and TOPSIS was analyzed and a single winner
from this set emerges.
Thus DEA analysis identifies 4 competitive candidates as DMU1, DMU4, DMU8 and
DMU15. And GAs identifies DMU3 this is subject to BCA which identifies DMU8 as winner
candidate. Whereas, TOPSIS identifies DMU1 as winner that has already appeared in above
solutions.Finally DMU8 emerges as a winning candidate. It is believed that the approach
generated in the present paper will work as a vital tool of decision support mechanism not only
International Journal of Innovative Research and Studies

Page 662

December, 2013

www.ijirs.com

Vol 2 Issue 12

for SSI units but also for the other units of industry and service sector, especially for
procurement of units.

International Journal of Innovative Research and Studies

Page 663

December, 2013

www.ijirs.com

Vol 2 Issue 12

References:
1. Nan Fang, Shinn Sun, Comparing CNC Lathes Using Data Envelopment Analysis
APDSI 2000
2. Website: http://dcmsme.gov.in/ssiindia/performance.htm
3. MadhuBala, Lesson 20: Structure, Policy s Support to Small Scale Industries
4. Bhattacharya Rajiv K, 2012, Introduction to Genetics Algorithm, Department of Civil
Engineering, IIT, Guwahati.
5. Pawar PJ, and Bhusnar, 2010, Genetic Algorithm using adaptive selection method,,
Pune University.
6. GrooverMikell P and Zimmers Emory W Jr., 2011, CAD CAM: Computer-Aided
Design and Manufacturing, Tenth Impression, New Delhi: Dorling Kindersley (India)
Pvt. Ltd.
7. F. HosseinzadehLotfi, M. Rostamy mal khalifeh, M. HeydariAlvar (2012), A New
Method for Ranking Efficient DMUs Based on TOPSIS and Virtual DMUs Int. J.
Research in Industrial Engineering, pp. 1- 9

International Journal of Innovative Research and Studies

Page 664

Anda mungkin juga menyukai