International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056

Volume: 04 Issue: 05 | May -2017 www.irjet.net p-ISSN: 2395-0072

“A Simplified Description of FUZZY TOPSIS Method for

Multi Criteria Decision Making”
Miss. Neelima B. Kore1, Prof. Dr. K. Ravi2, Asst. Prof. Mr. S. B. Patil3
1PGStudent, Civil Engineering Department, Ashokrao Mane College of Engineering,
Vathar tarf Vadgaon, Maharashtra, India.
2Professor, Civil Engineering Department, Ashokrao Mane College of Engineering,

Vathar tarf Vadgaon, Maharashtra, India.

3Assistant Professor, Civil Engineering Department, Ashokrao Mane College of Engineering,

Vathar tarf Vadgaon, Maharashtra, India.

---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract – To achieve success in any project, a quality of 2. FUZZY TOPSIS THEORY DETAILS
work and various decision modes plays an important role.
Usually, many factors affect the project work and with few The Technique for Order of Preference by
similar options for work done creates complications in Similarity to Ideal Solution (TOPSIS) is a multi-criteria
decision making. Hence, cause the rising of Optimization decision analysis method, which was originally developed by
Techniques to have a full proof decisions and conclusions. Hwang and Yoon in 1981 with further developments by
Fuzzy TOPSIS (Technique for Order Preference by Similarities Yoon in 1987 and Hwang, Lai and Liu in 1993. TOPSIS is
to Ideal Solution) is one of the best method to get ideal based on the concept that the chosen alternative should have
solution among the similar options. Also it can be used to the shortest geometric distance from the positive ideal
automate the process and overcome ambiguity, uncertainty in solution (PIS) and the longest geometric distance from the
selection process. So, we have described method of Fuzzy negative ideal solution (NIS). It is a method of compensatory
TOPSIS for a multi-criteria group decision making scenario. aggregation that compares a set of alternatives by
identifying weights for each criterion.
Key Words: Optimization Techniques, Similar Options, Fuzzy
TOPSIS, multi-criteria decision making. As the parameters or criteria are often of
incongruous dimensions in multi-criteria problems it may
1. INTRODUCTION create problems in evaluation. So, to avoid this problem a
need of Fuzzy system is necessary. Using Fuzzy numbers in
In general, Multi Criteria Group Decision Making TOPSIS for criteria analysis make it simple for evaluation.
(MCGDM) problems are frequently evaluated. To solve Hence, Fuzzy TOPSIS is simple, realistic form of modelling
problems related to decision making several optimization and Compensatory method which include or exclude
methods are used in practice. But, in case where decision alternative solutions based on hard cut-offs.
activity is based on similar options it becomes critical to
analyse various factors, alternatives with similar category. 2.1 Definitions

One simple example, a group of three person (say A, The definitions of fuzzy TOPSIS have been adapted
B and C) intends to determine which mobile phone is to buy from sources. These definitions are presented as follows.
based on certain criteria. Let say they have various criteria
like price, model quality, screen size, battery life and Definition 1: A fuzzy set ā in a universe of discourse
memory etc. But each person among A, B and C may give X is characterized by a membership function μā(x) that maps
different importance to different criteria. So, now it becomes each element x in X to a real number in the interval [0, 1].
challenge to decision makers (i.e. A, B and C) to find which The function value μā(x) is termed the grade of membership
alternative best meets the group’s criteria. of x in ā. The nearer the value of μā(x) to unity, the higher
the grade of membership of x in ā.
Fuzzy TOPSIS is a method that can help in objective
and systematic evaluation of alternatives on multiple Definition 2: A triangular fuzzy number is
criteria. Here, we provide a simplified description of represented as a triplet ā = (a1, a2, a3). The membership
required Fuzzy theory details and an example scenario has function μā(x) of triangular fuzzy number ā.
been worked out to illustrate the Fuzzy TOPSIS steps.

© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 05 | May -2017 www.irjet.net p-ISSN: 2395-0072

A fuzzy set ā, membership function 𝜇ā that maps each 3. FUZZY TOPSIS METHOD
element x in X to a real number in the interval [0, 1]. A
triangular fuzzy number is represented ā = (𝑎1, 𝑎2, 𝑎3) The technique called fuzzy TOPSIS (Technique for
Order Preference by Similarity to Ideal Situation) can be
used to evaluate multiple alternatives against the selected
criteria. In the TOPSIS approach an alternative that is
nearest to the Fuzzy Positive Ideal Solution (FPIS) and
farthest from the Fuzzy Negative Ideal Solution (FNIS) is
chosen as optimal. An FPIS is composed of the best
performance values for each alternative whereas the FNIS
consists of the worst performance values. Here, we have
presented relevant steps of fuzzy TOPSIS as below.
Fig -1: Triangular Fuzzy number system
3.1 Steps to illustrate Fuzzy TOPSIS method using
𝑎2gives the maximal grade of 𝜇 𝑎that 𝜇 𝑎=1 an example scenario
𝑎1gives the minimal grade of 𝜇 𝑎that 𝜇 𝑎=0
𝑎1and 𝑎3are the lower and upper bounds of the available Let’s say, the decision group has K members and the
area for the evaluation data. ith alternative on jth criterion. The fuzzy rating and
x – a1/ a2 – a1 if a1 ≤ x ≤ a2 importance weight of the kth decision maker, about the ith
𝜇ā(x)= a3 – x / a3 – a2 if a2 ≤ x ≤ a3 alternative on jth criterion.
0 Otherwise
Step 1: ALTERNATIVES RATINGS B Y DECISION MAKERS

0, x < a1 Here, we have two alternatives such as A1 and A2

x – a1/ a2 – a1 if a2 ≥ x ≥ a1 for comparison with four criteria such as C1, C2, C3 and C4
𝜇ā(x)= x – a2 / a3 – a2 if a2 ≥ x ≥ a3 also we have three decision makers namely DM1 and DM2.
0 x > a3 …..(1) Now, decision makers rate the alternatives as shown in
table-2.
2.2 The distance between fuzzy triangular Table -2
numbers ALTERNATIVE RATING
A1 A2
Criteria
Let ā = (a1, a2, a3) and b = (b1, b2, b3) be two DM1 DM2 DM1 DM2
triangular Fuzzy numbers. The distance between them is C1 F F G G
given using the vertex method by: C2 VG VG G VG
d (ā, b) = √ 1/3 [(a1 − b1)2 + (a2 − b2)2 + (a3 − b3)2] …..(2) C3 P F P P
C4 F F P P
2.3 Fuzzy Set Theory
Step 2: CRITERIA WEIGHTAGE BY DECISION MAKERS
Conversion scales are applied to transform the Table -3
linguistic terms into fuzzy numbers. Usually apply a scale of CRITERIA WEIGHTAGE
1 to 9 for rating the criteria and the alternatives. The
intervals are chosen so as to have a uniform representation Criteria DM1 DM2
from 1 to 9 for the fuzzy triangular numbers used for the five C1 H M
linguistic ratings. C2 VH H
C3 VH H
Table -1 C4 M L
FUZZY RATINGS FOR LINGUISTIC VARIABLES
Fuzzy Alternative QA Weights Step 3: APPLY FUZZY NUMBERS (REFER TABLE NO. 1)
number Assessment Table -4
(1,1,3) Very Poor (VP) Very Low (VL) FUZZY NUMBERS FOR ALTERNATIVE RATING
(1,3,5) Poor (P) Low (L) A1 A2
Criteria
(3,5,7) Fair (F) Medium (M) DM1 DM2 DM1 DM2
(5,7,9) Good (G) High (H) C1 F(3,5,7) F(3,5,7) G(5,7,9) G(5,7,9)
(7,9,9) Very Good (VG) Very High (VH) C2 VG(7,9,9) VG(7,9,9) G(5,7,9) VG(7,9,9)
C3 P(1,3,5) F(3,5,7) P(1,3,5) P(1,3,5)
C4 F(3,5,7) F(3,5,7) P(1,3,5) P(1,3,5)

© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 2
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 05 | May -2017 www.irjet.net p-ISSN: 2395-0072

Table -5 Be cost criteria. Aim is to maximize benefit and minimize the

FUZZY NUMBERS FOR CRITERIA WEIGHTAGE cost. A fuzzy multi criteria Group Decision Making (GDM)
problem which can be concisely expressed in matrix format
Criteria DM1 DM2
as:
C1 H(5,7,9) M(3,5,7) C1 C2 Cn
C2 VH(7,9,9) H(5,7,9) A1 x11 x12 … x1n
C3 VH(7,9,9) H(5,7,9) D = A2 x21 x22 … x2n
Am xm1 xm2 … xmn …..(5)
C4 M(3,5,7) L(1,3,5)

W = (w1, w2, …. , wn) …..(6)

Step 4: AGGREGATED ALTERNATIVE AND CRITERIA
WEIGHTAGE FUZZY DECISION MATIX NORMALIZING
xki j = (aki j , bki j , cki j) R = [rij]m×n, i = 1, 2, . . . ,m; j = 1, 2, . . . , n …..(7)
wkj = (wkj 1, wkj 2, wkj 3)
K
rij = (aij / c*j, bij / c*j, cij / c*j) and
ai j = min{aki j}, bi j =1/K ∑ bki j , ci j = max{cki j} …..(3) c*j = max cij (benefit criteria) …..(8)
k k=1 k
i
K
rij = (aj / cij , aj / bij , aj / aij) and
wj1 = min{wj k1}, wj2 =1/K ∑ wj k2 , wj3 = max{wj k3} …..(4) aj = min aij (cost criteria) …..(9)
k k=1 k
i
Table -8
Table -6 NORMALIZED AGGREGATED FUZZY DECISION MATRIX
AGGREGATED FUZZY DECISION MATRIX FOR FOR ALTERNATIVE
ALTERNATIVE Criteria A1 A2
Criteria A1 A2
C1 (0.429,0.600,1.000) (0.333,0.429,0.600)
C1 (3.000,5.000,7.000) (5.000,7.000,9.000)
C2 (0.778,1.000,1.000) (0.556,0.889,1.000)
C2 (7.000,9.000,9.000) (5.000,8.000,9.000)
C3 (0.143,0.571,1.000) (0.143,0.429,0.714)
C3 (1.000,4.000,7.000) (1.000,3.000,5.000)
C4 (0.429,0.714,1.000) (0.143,0.429,0.714)
C4 (3.000,5.000,7.000) (1.000,3.000,5.000)
[The ranges of normalized triangular fuzzy numbers belong
to (0, 1)]
Example – C1 A1 – from Table -4 (use Eq.3)
1. ai j = min{aki j} = 3.000 [i.e. minimum value of first place Weighted Normalized Fuzzy Decision Matrix,
k (3,5,7 & 3,5,7)]
K
𝑃=[ 𝑝𝑖𝑗] where 𝑝𝑖𝑗= 𝑟𝑖𝑗× 𝑤𝑗 …..(10)
2. bi j =1/K ∑ bki j = 5.000 [i.e. average of values at middle
k=1 place (3,5,7 & 3,5,7)] Table -9
WEIGHTED NORMALIZED FUZZY DECISION MATRIX
3. ci j = max{cki j} = 7.000 [i.e. maximum value of last place Criteria A1 A2
k (3,5,7 & 3,5,7)] C1 (1.287,3.600,9.000) (0.999,2.574,5.400)
C2 (3.890,8.000,9.000) (2.780,7.112,9.000)
Table -7
C3 (0.715,4.568,9.000) (0.715,3.432,6.426)
AGGREGATED FUZZY DECISION MATRIX FOR CRITERIA
WEIGHTAGE C4 (0.429,2.856,7.000) (0.143,1.716,4.998)
Criteria Agg. Weightage [Refer Table- 7, Table- 8 and Eq.10]
C1 (3.000,6.000,9.000)
C2 (5.000,8.000,9.000) Step 6: FPIS AND FNIS
C3 (5.000,8.000,9.000)
A+ = (𝑝+1, 𝑝+2, . . . , 𝑝+n) where
C4 (1.000,4.000,7.000)
𝑝+j = max {𝑝ij3}, i = 1, 2 . . . ,m; j = 1, 2, . . . , n …..(11)
[Same procedure of work as did for table -6 (use Eq.4)] i
A- = (𝑝-1, 𝑝-2, . . . , 𝑝-n ) where
Step 5: FUZZY MULTI CRITERIA GROUP DECISION 𝑝-j = min {𝑝ij1}, i = 1, 2 . . . ,m; j = 1, 2, . . . , n …..(12)
MAKING (GDM) AND PROCESS OF NORMALIZING i
Select the maximum value from each row as p+ and select
As we are working on various criteria for decision the minimum value from each row as p-.
making, some might be benefit criteria and some might

© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 3
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 05 | May -2017 www.irjet.net p-ISSN: 2395-0072

A+ = [𝑝+1(9,9,9,), 𝑝+2(9,9,9), 𝑝+3(9,9,9), 𝑝+4(7,7,7)] …..(13) Step 10: RANKING OF EACH ALTERNATIVE

A- = [𝑝-1(0.999,0.999,0.999), 𝑝-2(2.780,2.780,2.780), Hence, the ranking order of A1 > A2,

𝑝-3(0.715,0.715,0.715), 𝑝-4(0.143,0.143,0.143)] …..(14) A1 is the Best Choice considering the given criteria.

Step 7: FPIS AND FNIS for each criteria 3.2 Summary

FPIS (A1) = d (pij, p1+) and Above we describe the application of FUZZY TOPSIS for a
FNIS (A1) = d (pij, p1-) scenario where there are 2 decision makers, 4 evaluation
Here, we have to use equation no. (2) To find distance of criteria C1 – C4, and rating scale is as shown in Table 1. Key
each criteria from FPIS AND FNIS for both alternatives. input from decision makers is typically alternative rating and
So, using equation 2, 13, 14 and table no. 9 we get, to identify the proper weightage to various criteria. As result
of above Fuzzy TOPSIS steps, Closeness coefficients CCi, of
Table -10 the two A1 and A2 come out to be 0.510 and 0.393
DISTANCE OF CRITERIA OF EACH ALTERNATIVE FROM respectively. Hence the ranking order for the alternatives is
FPIS AND FNIS A1 > A2, that is, A1 is the best choice considering the given
Criteria FPIS(A1) FPIS(A2) FNIS(A1) FNIS(A2) criteria. The closeness coefficient scores for alternatives are
C1 5.436 6.279 4.860 2.698 numeric values and can be further utilized to indicate the
C2 3.006 3.753 4.732 4.376 degree of inferiority or superiority of the alternatives w.r.t
C3 5.425 5.952 5.275 3.651 each other.
C4 4.485 5.129 4.261 2.946
4. CONCLUSION
Example - from table 9 Criteria A1
C1 (0.999,2.574,5.400) As a result, use of optimization techniques in work
helps in reducing human efforts. Fuzzy TOPSIS is a method
which uses data in any form like numerical and linguistic etc.
From eq. 13, 𝑝+1(9.000,9.000,9.000) After data collection and converting it into Fuzzy system
From eq. 2 mode it reduces chances of errors caused by units of
d (ā, b) = √ 1/3 [(a1 − b1)2 + (a2 − b2)2 + (a3 − b3)2] parameters which also creates problem for mathematical
calculations. To overcome these challenges Fuzzy TOPSIS is
Now, for (C1-A1+) an ideal method. Also using software like MATLAB and its
Fuzzy tools we can create a model for Fuzzy TOPSIS. So,
d = √ 1/3 [(0.999-9.000)2 + (2.574-9.000)2 + (5.400– 9.000)2] Fuzzy TOPSIS is a simple, superior and full proof solution to
multi criteria decision making.
= √ 1/3 [(64.02) + (41.29) + (12.96)]

= √ [39.423] REFERENCES
= 6.279 (likewise use same method to find remaining values
of table-10)
[1] S. Saghafian and S. Hejazi, “Multi-criteria group decision
making using a modified fuzzy topsis procedure,” vol. 2,
pp. 215 –221, nov. 2005.
Step 8: THE DISTANCE OF EACH WEIGHTED [2] J. Buckley, “Ranking alternatives using fuzzy numbers,”
ALTERNATIVE Fuzzy sets and systems, vol. 15, no. 1, pp. 21–31, 1985.
[3] Y. Jiang et al., “A method for group decision making with
𝑑𝑖+= ∑𝑛𝑗=1𝑑( 𝑝𝑖𝑗,𝑝𝑗+) …..(15) multigranularity linguistic assessment information,”
Information Sciences, vol. 178, no. 4, pp. 1098–1109,
𝑑𝑖−= ∑𝑛𝑗=1𝑑( 𝑝𝑖𝑗,𝑝𝑗−) …..(16) 2008.
𝑑1+= 18.352 𝑑2+= 21.113 [4] FUZZY TOPSIS by Mohammad Khalily Dermany
𝑑1−= 19.128 𝑑2−= 13.671
BIOGRAPHIES
Step 9: CLOSENESS COEFFICIENT OF EACH ALTERNATIVE
Miss. Neelima Balawant Kore
CCi = d−i / (d−i + d+i), i = 1, 2 . . . ,m …..(17) PG student, Civil Engineering
Department, Ashokrao Mane
𝐶𝐶1=19.128 /(18.352+19.128)= 0.510
College of Engineering, Vathar tarf
𝐶𝐶2=13.671 /(13.671+ 21.113)= 0.393 Vadgaon, Maharashtra, India.

© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 4