Applied Soft Computing 21 (2014) 194209

Contents lists available at ScienceDirect

Applied Soft Computing
j our nal homepage: www. el sevi er . com/ l ocat e/ asoc
A comparison between Fuzzy AHP and Fuzzy TOPSIS methods to
supplier selection
Francisco Rodrigues Lima Junior
a
, Lauro Osiro
b
, Luiz Cesar Ribeiro Carpinetti
a,
a
School of Engineering of So Carlos, University of So Paulo, Production Engineering Department, So Carlos, SP, Brazil
b
Federal University of Triangulo Mineiro (UFTM), Production Engineering Department, Uberaba, MG, Brazil
a r t i c l e i n f o
Article history:
Received 8 July 2013
Received in revised form 13 January 2014
Accepted 20 March 2014
Available online 13 April 2014
Keywords:
Supplier selection
Fuzzy TOPSIS
Fuzzy AHP
Multicriteria decision making
a b s t r a c t
Supplier selection has become a very critical activity to the performance of organizations and supply
chains. Studies presented in the literature propose the use of the methods Fuzzy TOPSIS (Fuzzy Tech-
nique for Order of Preference by Similarity to Ideal Solution) and Fuzzy AHP (Fuzzy Analytic Hierarchy
Process) toaidthe supplier selectiondecisionprocess. However, there are nocomparative studies of these
two methods when applied to the problemof supplier selection. Thus, this paper presents a comparative
analysis of these two methods in the context of supplier selection decision making. The comparison was
made based on the factors: adequacy to changes of alternatives or criteria; agility in the decision process;
computational complexity; adequacy to support group decision making; the number of alternative sup-
pliers and criteria; and modeling of uncertainty. As an illustrative example, both methods were applied
to the selection of suppliers of a company in the automotive production chain. In addition, computational
tests were performed considering several scenarios of supplier selection. The results have shown that
both methods are suitable for the problemof supplier selection, particularly to supporting group decision
making and modeling of uncertainty. However, the comparative analysis has shown that the Fuzzy TOP-
SIS method is better suited to the problem of supplier selection in regard to changes of alternatives and
criteria, agility and number of criteria and alternative suppliers. Thus, this comparative study contributes
to helping researchers and practitioners to choose more effective approaches for supplier selection. Sug-
gestions of further work are also proposed so as to make these methods more adequate to the problem
of supplier selection.
2014 Elsevier B.V. All rights reserved.
1. Introduction
Supplier selection (SS) is one of the most important activities of
acquisition as its results have a great impact on the quality of goods
andperformanceof organizations andsupplychains [13]. Through
SS it is also possible to anticipate evaluation of the potential of
suppliers to establish a collaborative relationship [4].
Essentially, supplier selection is a decision process with the aim
of reducing the initial set of potential suppliers to the nal choices
[5,6]. Decisions are based on evaluation of suppliers on multiple
quantitative as well as qualitative criteria. Depending on the sit-
uation at hand, selecting suppliers may require searching for new

Corresponding author at: School of Engineering of So Carlos, University of

So Paulo, Production Engineering Department, Av. Trabalhador Sancarlense 400,
13566-590 So Carlos, SP, Brazil. Tel.: +55 1633739421.
E-mail addresses: eng.franciscojunior@gmail.com (F.R. Lima Junior),
lauro.osiro@gmail.com (L. Osiro), carpinet@sc.usp.br,
carpinetti.luizcesar46@gmail.com (L.C.R. Carpinetti).
suppliers or choosing suppliers fromthe existing pool of suppliers.
In any case there is a degree of uncertainty in the decision process,
which is caused by subjective evaluation of qualitative or quanti-
tative criteria, by multiple decision makers, with no previous data
to rely on [2,5,7,8].
Fuzzy set theory combined with multicriteria decision making
(MCDM) methods has been extensively used to deal with uncer-
taintyinthesupplier selectiondecisionprocess [9], sinceit provides
a suitable language to handle imprecise criteria, being able to inte-
grate the analysis of qualitative and quantitative factors. This is
the case of Fuzzy AHP Fuzzy Analytic Hierarchy Process [1016],
Fuzzy TOPSIS Fuzzy Technique for Order Preference by Similarity
to Ideal Solution [1,1723], among others.
Despite the large number of articles proposing the use of Fuzzy
AHP and Fuzzy TOPSIS, there are no comparative studies of these
two methods when applied to the problem of supplier selection.
Ertugrul and Karakasoglu [24] report a comparison of Fuzzy AHP
and Fuzzy TOPSIS methods applied to facility location decision
making. However, as the authors point out, there is still a need
for a comparative evaluation of both methods in the context of
http://dx.doi.org/10.1016/j.asoc.2014.03.014
1568-4946/ 2014 Elsevier B.V. All rights reserved.
F.R. Lima Junior et al. / Applied Soft Computing 21 (2014) 194209 195
Fig. 1. The supplier selection framework [5].
supplier selection, since the relative advantages of both methods
also depend on the characteristics of the problem domain. To ll
this gap, this paper presents a comparative analysis of the methods
Fuzzy TOPSIS and Fuzzy AHP applied to the problem of supplier
selection. Comparison of the methods was made considering the
factors: adequacy to changes of alternatives or criteria; agility
in the decision process; computational complexity; adequacy
to support group decision making; the amount of alternative
suppliers and criteria; and modeling of uncertainty.
Adescriptivequantitativeapproachwas adoptedas theresearch
method [25]. Algorithms of the methods Fuzzy TOPSIS and Fuzzy
AHP were developed in Matlab

and applied to the selection of

suppliers of a company. Comparison of both methods was made
based on the analysis of mathematical procedures considering the
structure of the problem depicted by the illustrative application
case.
The paper is organized as follows: Section 2 briey revises the
subject of supplier selection and the main requirements of multi-
criteria decision making methods used in this context. Section 3
presents some fundamental concepts regarding fuzzy set theory
and the methods Fuzzy TOPSIS and Fuzzy AHP. Section 4 presents
the results of using both methods in a real case application. Sec-
tion 5 presents the comparative analyses of both methods. Finally,
conclusions about this research work and suggestions for further
research are made in Section 6.
2. Supplier selection process
Supplier selection is a decision-making process comprising sev-
eral steps. Based on the studies of Faris et al. [26] and Kraljic [27],
De Boer et al. [5] propose a framework for supplier selection that
consists of four steps: problem denition, formulation of criteria,
qualication and nal choice, as illustrated in Fig. 1. The rst step
aims at clearly dening the problem at hand, which may mean
searching for new suppliers for a completely new product, replac-
ing current suppliers, or choosing suppliers for new products from
the existing pool of suppliers. Especially in the case of selecting
newsuppliers, depending on the itemto be purchased, the number
of alternative suppliers may be very large. This situation demands
decision making techniques that are able to simultaneously evalu-
ate several alternatives.
In the next step, the buyer should convert its requirements into
decisioncriteria so as to guide the choices. There are several criteria
that must be considered in the selection process, both quantitative
and qualitative. Table 1 lists some important criteria for supplier
selection. On top of traditional quantitative measures of perfor-
mance, such as quality of conformance, delivery time or cost, other
measures of subjective evaluation, for example supplier prole and
relationship, aregainingimportance[28]. Therefore, thetechniques T
a
b
l
e
1
S
u
p
p
l
i
e
r
p
e
r
f
o
r
m
a
n
c
e
c
r
i
t
e
r
i
a
a
c
c
o
r
d
i
n
g
t
o
s
e
l
e
c
t
e
d
a
u
t
h
o
r
s
.
C
r
i
t
e
r
i
a
P
r
o
p
o
s
e
d
b
y
K
a
n
n
a
n
a
n
d
T
a
n
[
2
9
]
K
a
h
r
a
m
a
n
e
t
a
l
.
[
1
1
]
K
a
t
s
i
k
e
a
s
e
t
a
l
.
[
3
0
]
C
h
a
n
a
n
d
K
u
m
a
r
[
1
2
]
K
i
r
y
t
o
p
o
u
l
o
s
e
t
a
l
.
[
3
1
]
G
u
n
e
r
i
e
t
a
l
.
[
3
2
]
S
h
e
n
a
n
d
Y
u
[
3
3
]
O
r
d
o
o
b
a
d
i
[
3
4
]
B
o
r
a
n
e
t
a
l
.
[
3
5
]
A
m
i
n
a
n
d
R
a
z
m
i
[
3
6
]
K
u
e
t
a
l
.
[
3
7
]
W
a
n
g
[
3
8
]
L
i
n
e
t
a
l
.
[
3
9
]
B

z
k
a
n
a
n
d
C

i
f
c

i
[
4
0
]
T
e
c
h
n
i
c
a
l
c
a
p
a
b
i
l
i
t
y
X
X
X
X
X
X
X
X
X
X
C
o
m
m
i
t
m
e
n
t
t
o
q
u
a
l
i
t
y
X
X
X
X
X
X
X
X
Q
u
a
l
i
t
y
o
f
c
o
n
f
o
r
m
a
n
c
e
X
X
X
X
X
X
X
X
F
l
e
x
i
b
i
l
i
t
y
(
r
e
s
p
o
n
s
e
t
o
c
h
a
n
g
e
)
X
X
X
X
X
X
X
C
o
s
t
/
p
r
i
c
e
X
X
X
X
X
X
X
X
X
F
i
n
a
n
c
i
a
l
s
i
t
u
a
t
i
o
n
X
X
X
X
X
X
X
X
E
a
s
y
o
f
c
o
m
m
u
n
i
c
a
t
i
o
n
X
X
X
X
X
X
X
O
n
-
t
i
m
e
d
e
l
i
v
e
r
y
X
X
X
X
X
X
X
X
R
e
p
u
t
a
t
i
o
n
X
X
X
X
X
X
R
e
l
a
t
i
o
n
s
h
i
p
X
X
X
X
X
X
P
r
o
d
u
c
t
p
e
r
f
o
r
m
a
n
c
e
X
X
X
X
X
D
e
l
i
v
e
r
y
r
e
l
i
a
b
i
l
i
t
y
X
X
X
X
X
A
f
t
e
r
s
a
l
e
/
w
a
r
r
a
n
t
y
X
X
X
X
X
X
G
e
o
g
r
a
p
h
i
c
l
o
c
a
t
i
o
n
X
X
X
X
E
n
d
u
s
e
X
X
X
S
o
c
i
a
l
f
a
c
t
o
r
s
X
E
n
v
i
r
o
n
m
e
n
t
a
l
f
a
c
t
o
r
X
L
o
g
i
s
t
i
c
c
o
s
t
s
X
I
n
n
o
v
a
t
i
o
n
X
196 F.R. Lima Junior et al. / Applied Soft Computing 21 (2014) 194209
Table 2
Decision making approaches applied to supplier selection.
Approach Method(s) Proposed by
Single method Analytic Hierarchy Process
(AHP)
Tam and Tummala [43]
Hudym cov et al. [44]
Analytic Network Process
(ANP)
Gencer and Grpinar [45]
Kirytopoulos et al. [31]
Data Envelopment Analysis
(DEA)
Saen [46]
Fuzzy Inference Carrera and Mayorga [47]
Fuzzy Preference Relation Hsu et al. [48]
Genetic Algorithm Liao and Rittscher [49]
Combined
method
Fuzzy Adaptive Resonance
Theory (Fuzzy ART)
Keskin et al. [50]
Fuzzy AHP Kahraman et al. [11]
Chan and Kumar [12]
Lee [13]
Chamodrakas et al. [14]
Zeydan et al. [15]
Kilinci and Onal [16]
Fuzzy ANP nut et al. [51]
Vinodh et al. [52]
Fuzzy c-means and Rough Set
Theory
Omurca [53]
Fuzzy Deployment Quality
Function (Fuzzy QFD)
Bevilacqua et al. [54]
Amin and Razmi [36]
Dursun and Karsak [55]
Fuzzy DEMATEL Bykzkan and C ifc i [40]
Fuzzy Inference and Fuzzy
Algebraic Operations
Amindoust et al. [8]
Fuzzy Multiobjective Linear Amid et al. [56]
Programming Arikan [57]
Fuzzy TOPSIS Chen et al. [1]
Chen [17]
Dagdeviren et al. [18]
Bykzkan and Ersoy [19]
Awasthi et al. [20]
Liao and Kao [21]
Jolai et al. [22]
Zouggari and Benyoucef [23]
Shahanaghi and Yazdian [58]
Bottani and RIZZI [59]
Fuzzy Two-Tupple Wang [38]
Fuzzy VIKOR Shemshadi et al. [60]
used in the decision process must be able to process several criteria
of both qualitative and quantitative nature [7].
In the qualication step, the main objective is to reduce the ini-
tial set of suppliers by sorting potential suppliers from the initial
set of suppliers based on qualifying criteria. The last step aims to
rank the potential suppliers so as to make the nal choice. Based
on this framework, Wu and Barnes [6] proposed a further step
with the purpose of giving potential suppliers feedback on their
performance in the selection process.
2.1. Multicriteria decision making methods in supplier selection
Multicriteria decision making (MCDM) methods for supplier
selection include multi-attribute techniques, mathematical pro-
gramming, stochastic programming and articial intelligence
techniques [6,9,41]. There are several different MCDM methods
used mostly for outranking.
As shown in Table 2, the combination between techniques is
usually adopted to deal with the problem of supplier selection.
Fuzzy set theory (FST) [42] has been extensively used for modeling
decision making processes based on imprecise and vague informa-
tion such as judgment of decision makers. The use of appropriate
techniques can bring effectiveness and efciency to the selection
process [7]. To decide which techniques to use one must take into
account the alignment of the particularities of the problemat hand
with the characteristics of the techniques [24]. For instance, when
selecting a new supplier of a routine item with many potential
suppliers [5], techniques that do not limit analysis to only a few
alternatives are more adequate than others.
Other aspects to be considered to align techniques to particu-
larities of supplier selection are as follows:

Adequacy to support group decision making: purchasing deci-

sions are inuenced by several requirements from different
functional areas within an organization. This implies that mul-
tiple actors from different functional areas are involved in the
decision making process [7]. Therefore, it is desirable that the
techniques used in supplier selection be adequate to combine
different judgments of multiple decision makers.

Adequacy to changes of alternatives or criteria: in the case of

modied rebuy, one may be interested in purchasing existing
products from new suppliers. This may lead to inclusion or
exclusion of supply alternatives in the evaluation process. Alter-
natively, whenmodiedrebuy refers to purchasing newproducts
from current suppliers, this may imply inclusion or exclusion
of decision criteria [5,26]. In both cases, the outranking tech-
niques should be robust enough not to cause inconsistencies in
the ordering of alternative suppliers.

Agility in the decision process: this factor relates mainly to the

required amount of judgments of the decision makers in data
collection. Depending on the MCDM technique and the number
of criteria and alternatives, the quantity of judgments needed to
collect all the data can make the supplier selection process very
time consuming [61].

Computational complexity: this factor may be related to either

time or space complexity. The mainconcerninthe supplier selec-
tion decision process is related to time complexity, which refers
to the time in which the algorithm is accomplished [10]. Time
complexity varies from technique to technique as a function of
the number of input variables, whichinthe case of supplier selec-
tion refers to the number of alternative suppliers and criteria.

Uncertainty: in supplier selection, the uncertainty in decision

making may refer to the lack of precision of the scores of the
alternatives as well as the relative importance of different crite-
ria. This imprecision may be due to: subjective evaluation by
multiple decision makers; inexistence of previous data on the
performance of potential suppliers and; difculty of assessing
intangible aspects of supplier performance [7,62].
3. Fuzzy set theory
Fuzzy set theory [42] has been used for modeling decision mak-
ing processes based on imprecise and vague information such as
judgment of decision makers. Qualitative aspects are represented
by means of linguistic variables, which are expressed qualitatively
by linguistic terms and quantitatively by a fuzzy set in the universe
of discourse and respective membership function [63]. Operations
between linguistic variables involve the concepts presented next.
3.1. Fundamental denitions
3.1.1. Denition 1: fuzzy set
A fuzzy set

A in X is dened by:

A = {x,
A
(x)}, x X (1)
in which
A
(x) : X [0, 1] is the membership function of

A and

A
(x) is the degree of pertinence of x in

AIf
A
(x) equals zero, x does
not belong to the fuzzy set

A. If
A
(x) equals 1, x completely belongs
to the fuzzy set

A. However, unlike the classical set theory, if
A
(x)
F.R. Lima Junior et al. / Applied Soft Computing 21 (2014) 194209 197
has a value between zero and 1, x partially belongs to the fuzzy set

A. That is, the pertinence of x is true with degree of membership

given by
A
(x) [42,62].
3.1.2. Denition 2: fuzzy numbers
Afuzzy number is a fuzzy set in which the membership function
satises the conditions of normality
sup

A(x)
x X
= 1 (2)
and of convexity
C
1
C
2
C
j
C
m

D =
A
1
A
i
A
n
_

_
x
11
x
12
x
1j
x
1m
.
.
.
.
.
.
.
.
.
.
.
.
x
n1
x
n2
x
nj
x
nm
_

_
(3)
for all x
1
, x
1
X and all [0,1]. The triangular fuzzy number is
commonly usedindecisionmaking due to its intuitive membership
function,

W = [ w
1
+ w
2
+ + w
m
], given by:

A
(x) =
_

_
0 for x < l,
x l
ml
for l x m,
u x
u m
for m x u,
0 for x > u,
(4)
inwhichl, manduare real numbers withl <m<u. Outside the inter-
val [l, u], the pertinence degree is null, andmrepresents the point in
which the pertinence degree is maximum. Trapezoidal fuzzy num-
bers are also frequently used in decision making processes [62,64].
3.1.3. Algebraic operations with fuzzy numbers
Given any real number K and two fuzzy triangular numbers

A =
(l
1
, m
1
, u
1
) and

B = (l
2
, m
2
, u
2
), the main algebraic operations are
expressed as follows [62,64]:
(1) Addition of two triangular fuzzy numbers

A(+)

B = (l
1
+l
2
, m
1
+m
2
, u
1
+u
2
) l
1
0, l
2
0 (5)
(2) Multiplication of two triangular fuzzy numbers

A()

B = (l
1
l
2
, m
1
m
2
, u
1
u
2
) l
1
0, l
2
0 (6)
(3) Subtraction of two triangular fuzzy numbers

A()

B = (l
1
l
2
, m
1
m
2
, u
1
u
2
) l
1
0, l
2
0 (7)
(4) Division of two triangular fuzzy numbers

A()

B = (l
1
l
2
, m
1
m
2
, u
1
u
2
) l
1
0, l
2
0 (8)
(5) Inverse of a triangular fuzzy number

A
1
=
_
1
u
1
,
1
m
1
,
1
l
1
_
0 (9)
(6) Multiplication of a triangular fuzzy number by a constant
k

A = (k l
1
, k m
1
, k u
1
) l
1
0, k0 (10)
(7) Division of a triangular fuzzy number by a constant

A
k
=
_
l
1
k
,
m
1
k
,
u
1
k
_
l
1
0, k0 (11)
3.2. Fuzzy TOPSIS
The Fuzzy TOPSIS method was proposed by Chen [17] to solve
multicriteriadecisionmakingproblems under uncertainty. Linguis-
tic variables are used by the decision makers, D
r
(r =1, . . ., k), to
assess the weights of the criteria and the ratings of the alternatives.
Thus,

W
j
r
describes the weight of the jth criterion, C
j
(j =1, . . ., m),
given by the rth decision maker. Similarly, x
r
ij
describes the rating of
the ith alternative, A
i
(i =1, . . ., n), with respect to criterion j, given
by the rth decision maker. Given that, the method comprises the
following steps:
(i) Aggregate the weights of criteria and ratings of alternatives
given by k decision makers, as expressed in Eqs. (12) and (13)
respectively:
w
j
=
1
k
[ w
1
j
+ w
2
j
+ + w
k
j
] (12)
x
ij
=
1
k
[ x
1
ij
+ x
r
ij
+ + x
k
j
] (13)
(ii) Assemble the fuzzydecisionmatrixof the alternatives (

D) and
the criteria (

W), according to Eqs. (14) and (15):
C
1
C
2
C
j
C
m

D =
A
1
A
i
A
n
_

_
x
11
x
12
x
1j
x
1m
.
.
.
.
.
.
.
.
.
.
.
.
x
n1
x
n2
x
nj
x
nm
_

_
(14)

W = [ w
1
+ w
2
+ + w
m
] (15)
(iii) Normalize the fuzzy decision matrix of the alternatives (

D)
using linear scale transformation. The normalized fuzzy deci-
sion matrix

R is given by:

R = [ r
ij
]
mn
(16)
r
ij
=
_
l
ij
u
+
j
,
m
ij
u
+
j
,
u
ij
u
+
j
_
andu
+
j
= max
i
u
ij
(benet criteria) (17)
r
ij
=
_
l

j
u
ij
,
l

j
m
ij
,
l

j
l
ij
_
andl

j
= max
i
l
ij
(cost criteria) (18)
(iv) Compute the weighted normalized decision matrix,

V, by
multiplying the weights of the evaluation criteria, w
j
, by the
elements r
ij
of the normalized fuzzy decision matrix

V = [ v
ij
]
mn
(19)
where v
ij
is given by Eq. (20).
v
ij
= x
ij
w
j
(20)
(v) Dene the Fuzzy Positive Ideal Solution (FPIS, A
+
) and the
Fuzzy Negative Ideal Solution (FNIS, A

), according to Eqs.
(21) and (22).
A
+
= { v
+
1
, v
+
j
, . . ., v
+
m
} (21)
A

= { v

1
, v

j
, . . ., v

m
} (22)
where v
+
j
= (1, 1, 1) and v

j
= (0, 0, 0).
(vi) Compute the distances d
+
j
and d

i
of each alternative from
respectively v
+
j
and v

j
according to Eqs. (23) and (24)
d
+
i
=

n
j=1
d
v
( v
ij
, v
+
j
) (23)
198 F.R. Lima Junior et al. / Applied Soft Computing 21 (2014) 194209
d

i
=

n
j=1
d
v
( v
ij
, v

j
) (24)
where d(.,.) represents the distance between two fuzzy num-
bers according to the vertex method. For triangular fuzzy
numbers, this is expressed as in Eq. (25).
d( x, z) =
_
1
3
[(l
x
l
z
)
2
+(m
x
m
z
)
2
+(u
x
u
z
)
2
] (25)
(vii) Compute the closeness coefcient, CC
i
, according to Eq. (26).
CC
i
=
d

i
d
+
i
+d

i
(26)
(viii) Dene the ranking of the alternatives according to the close-
ness coefcient, CC
i
, in decreasing order. The best alternative
is closest to the FPIS and farthest to the FNIS.
3.3. Fuzzy AHP
Chang [10] proposed a Fuzzy AHP approach based on the
extent analysis method, which is widely used in supplier selec-
tion problems [16,24]. This method uses linguistic variables to
express the comparative judgments given by decision makers. Let
X = {x
1
, x
i
, . . ., x
n
} represent an object set and G = {g
1
, g
j
, . . ., g
m
}
a goal set. In the method proposed by Chang [10], each object, x
i
,
is taken and extent analysis is performed for each goal, g
j
. Thus,
m extent analysis values for each object can be obtained, with the
following signs:
M
1
gi
, M
j
gi
, . . ., M
m
gi
, i = 1, 2, . . ., n (27)
where all the M
j
gi
(j =1, 2, . . ., m) are triangular fuzzy numbers.
The method follows the steps described next.
(i) Compute the value of the fuzzy synthetic extent with respect
to the ith object according to Eq. (28).
S
i
=

m
j=1
M
i
gi

n
i=1

m
j=1
M
i
gi
_
1
(28)
where

m
j=1
M
i
gi
is obtained by performing the fuzzy addition
operation of m extent analysis values for a particular matrix
such that

m
j=1
M
i
gi
=
_

m
j=1
l
j
,

m
j=1
m
j
,

m
j=1
u
j
_
(29)
and
_

n
i=1

m
j=1
M
i
gi
_
1
is given by
_

n
i=1

m
j=1
M
i
gi
_
1
=
_
1

n
i=1

m
j=1
u
i
,
1

n
i=1

m
j=1
m
i
,
1

n
i=1

m
j=1
l
i
_
(30)
(ii) Compute the degree of possibility of S
2
(l
2
, m
2
, u
2
)S
1
=
(l
1
, m
1
, u
1
), where S
2
and S
1
are given by Eq. (28). The degree
of possibility between two fuzzy synthetic extents is dened
as in Eq. (31)
V(S
2
S
1
) = sup
yx
[min(
S
2
(y),
S
1
(x))] (31)
which can be equivalently expressed as in Eqs. (32) and (33).
V(S
2
S
1
) = hgt(S
1
S
2
) =
S
2
(d) (32)
Fig. 2. The intersection between S
1
and S
2
.

S
2
(d) =
_

_
1, if m
2
m
1
0, if l
1
u
2
l
1
u
2
(m
2
u
2
) (m
1
l
1
)
, otherwise
_

_
(33)
In Eqs. (32) and (33), d represents the ordinate of the highest
intersection point D between
S
1
and
S
2
, as it can be seen in
Fig. 2. The comparison between M
1
and M
2
requires the values
of V(S
2
S
1
) and V(S
1
S
2
).
(iii) Compute the degree of possibility for a convex fuzzy number
to be greater than k convex fuzzy numbers S
i
(i =1, . . ., k). This
is calculated according to Eq. (34).
V(SS
1
, S
2
, . . ., S
k
)
= V[(SS
1
) and(SS
2
) and and(SS
k
)]
= minV(SS
i
), i = 1, 2, . . ., k.
(34)
(iv) Compute the vector W

, which is given by Eq. (35).

W

= (d

(A
1
), d

(A
2
), . . ., d

(A
k
))
T
(35)
assuming that
d

(A
i
) = minV(S
i
S
j
), for i = 1, 2, .., k, j = 1, 2, . . ., k, k / = j (36)
The normalized vector is indicated by
W = (d(A
1
), d(A
2
), . . ., d(A
k
))
T
(37)
where W is a non-fuzzy number calculated for each comparison
matrix.
4. Application case in the automotive industry
A manufacturer of transmission cables for motorcycles needs
to select a supplier of metallic components used in a variety of
transmission cables. To select the best alternative, ve potential
suppliers were evaluated against ve decision criteria. The evalu-
ation of the potential suppliers in each criterion was made based
on linguistic judgments given by the decision makers, a group of
employees from the quality and purchase areas of the company.
The criteria were dened by the decision makers, as follows:

Quality (C
1
): related to quality of conformance, quality manage-
ment and after sale service quality.

Price (C
2
): related to the acquisition cost.

Delivery (C
3
): related to delivery time and reliability.

Supplier prole (C
4
): related to supplier reputation and nancial
health.

Supplier relationship (C
5
): related to the degree of cooperation
and trust in the buyersupplier relationship.
F.R. Lima Junior et al. / Applied Soft Computing 21 (2014) 194209 199
Fig. 3. Linguistic scale of the weights of the criteria.
Fig. 4. Linguistic scale of the ratings of the alternatives.
The Fuzzy TOPSIS and Fuzzy AHP methods were applied to this
case, as described next.
4.1. Fuzzy TOPSIS application
Evaluations of the weight of the criteria and the ratings of the
alternatives were made by the decision makers according to the
linguistic terms depicted in Figs. 3 and 4 respectively. Based on
Chen [17], triangular fuzzy numbers (TFN) were used to specify the
linguistic values of these variables, as presented in Tables 3 and 4.
Table 5 presents the linguistic judgments of the weights of the
criteria and the ratings of the alternatives for the three decision
makers involved in the selection process. The linguistic variables
shown in Table 5 are converted into TFN. Table 6 presents the
parameters of the TFNresulting fromaggregation of the judgments
Table 3
Linguistic scale to evaluate the weight of the criteria.
Linguistic terms Fuzzy triangular number
Of little importance (VL) (0.0, 0.0, 0.25)
Moderately important (MI) (0.0, 0.25, 0.50)
Important (I) (0.25, 0.50, 0.75)
Very important (VI) (0.50, 0.75, 1.0)
Absolutely important (AI) (0.75, 1.0, 1.0)
Table 4
Linguistic scale to evaluate the ratings of the alternative suppliers.
Linguistic terms Fuzzy triangular number
Very low (VL) (0.0, 0.0, 2.5)
Low (L) (0.0, 2.5, 5.0)
Good (G) (2.5, 5.0, 7.5)
High (H) (5.0, 7.5, 10.0)
Excellent (EX) (7.5, 10.0, 10.0)
Table 5
Linguistic ratings of the alternative suppliers by different decision makers.
C
1
C
2
C
3
C
4
C5
DM1
A
1
G H G G L
A
2
VH VH VH VH H
A
3
H H VH G H
A
4
G H G H H
A5 G H H G G
Weights of criteria VI AI VI I VI
DM2
A
1
G G G H G
A
2
VH VH VH VH VH
A
3
VH H VH H VH
A
4
H VH H H H
A5 H H H H H
Weights of criteria AI AI VI I I
DM3
A
1
H G G G H
A
2
VH VH VH VH VH
A
3
VH H VH G VH
A
4
G VH H H G
A5 L G G G G
Weights of criteria AI AI AI VI I
presented in Table 5, which represents the fuzzy decision matrix.
Thenormalizedfuzzydecisionmatrixandtheweightednormalized
fuzzydecisionmatrixarerepresentedrespectivelyinTables 7and8.
According to Chen [17], the Fuzzy Positive Ideal Solution (FPIS,
A
+
) and the Fuzzy Negative Ideal Solution (FNIS, A

) were dened
as
A
+
= [(1, 1, 1), (1, 1, 1), (1, 1, 1), (1, 1, 1), (1, 1, 1)]
A

= [(0, 0, 0), (0, 0, 0), (0, 0, 0), (0, 0, 0), (0, 0, 0)]
The distances d
+
i
and d

i
of the ratings of each alternative from
A
+
and A

, calculated according to Eqs. (23)(25), are presented in

Tables 9 and 10 respectively.
The global performance of each supplier alternative is given by
the closeness coefcient, CC
i
, calculatedas inEq. (26) andpresented
Table 6
Fuzzy numbers of the aggregated ratings of the alternative suppliers.
C
1
C
2
C
3
C
4
C5
A
1
(3.33, 5.83, 8.33) (3.33, 5.83, 8.33) (2.50, 5.00, 7.50) (3.33, 5.83, 8.33) (0.25, 5.00, 7.50)
A
2
(7.50, 10.0, 10.0) (7.50, 10.0, 10.0) (7.50, 10.0, 10.0) (7.50, 10.0, 10.0) (6.67, 9.17, 10.0)
A
3
(6.67, 9.17, 10.0) (5.00, 7.50, 10.0) (7.50, 10.0, 10.0) (3.33, 5.83, 8.33) (6.67, 9.17, 10.0)
A
4
(3.33, 5.83, 8.33) (6.67, 9.17, 10.0) (4.17, 6.67, 9.17) (5.00, 7.50, 10.0) (4.17, 6.67, 9.17)
A5 (2.50, 5.00, 7.50) (4.17, 6.67, 9.17) (4.17, 6.67, 9.17) (3.33, 5.83, 8.33) (3.33, 5.83, 8.33)
Weights of criteria (0.67, 0.92, 1.00) (0.75, 1.00, 1.00) (0.67, 92, 1.00) (0.42, 0.67, 0.92) (0.33, 0.58, 0.83)
200 F.R. Lima Junior et al. / Applied Soft Computing 21 (2014) 194209
Table 7
Normalized fuzzy decision matrix.
C
1
C
2
C
3
C
4
C5
A
1
(0.33, 0.58, 0.83) (0.33, 0.58, 0.83) (0.25, 0.50, 0.75) (0.33, 0.58, 0.83) (0.25, 0.50, 0.75)
A
2
(0.75, 1.00, 1.00) (0.75, 1.00, 1.00) (0.75, 1.00, 1.00) (0.75, 1.00, 1.00) (0.67, 0.92, 1.00)
A
3
(0.67, 0.92, 1.00) (0.50, 0.75, 1.00) (0.75, 1.00, 1.00) (0.33, 0.58, 0.83) (0.67, 0.92, 1.00)
A
4
(0.33, 0.58, 0.83) (0.67, 0.92, 1.00) (0.42, 0.67, 0.92) (0.50, 0.75, 1.00) (0.42, 0.67, 0.92)
A5 (0.25, 0.50, 0.75) (0.42, 0.67, 0.92) (0.42, 0.67, 0.92) (0.33, 0.58, 0.83) (0.33, 0.58, 0.83)
Table 8
Weighted normalized fuzzy decision matrix.
C
1
C
2
C
3
C
4
C5
A
1
(0.22, 0.53, 0.83) (0.25, 0.58, 0.83) (0.17, 0.45, 0.75) (0.13, 0.39, 0.76) (0.08, 0.29, 0.63)
A
2
(0.50, 0.92, 1.00) (0.56, 1.00, 1.00) (0.50, 0.92, 1.00) (0.31, 0.67, 0.92) (0.22, 0.53, 0.83)
A
3
(0.44, 0.84, 1.00) (0.38, 0.75, 1.00) (0.50, 0.92, 1.00) (0.14, 0.39, 0.76) (0.22, 0.53, 0.83)
A
4
(0.22, 0.53, 0.83) (0.50, 0.92, 1.00) (0.28, 0.61, 0.92) (0.21, 0.50, 0.92) (0.14, 0.39, 0.76)
A5 (0.16, 0.46, 0.75) (0.31, 0.67, 0.92) (0.28, 0.61, 0.92) (0.14, 0.39, 0.76) (0.11, 0.34, 0.69)
Table 9
Distances of the ratings of each alternative from A
+
with respect to each criterion.
C
1
C
2
C
3
C
4
C5 d
+
i
d(A
1
, A
+
) 0.53 0.50 0.59 0.62 0.70 2.95
d(A
2
, A
+
) 0.29 0.25 0.29 0.44 0.53 1.81
d(A
3
, A
+
) 0.33 0.39 0.29 0.62 0.53 2.17
d(A
4
, A
+
) 0.53 0.29 0.48 0.54 0.62 2.47
d(A5, A
+
) 0.59 0.44 0.48 0.62 0.66 2.80
Table 10
Distances of the ratings of each alternative from A

with respect to each criterion.

C
1
C
2
C
3
C
4
C5 d

i
d(A
1
, A

) 0.59 0.60 0.52 0.50 0.40 2.61

d(A
2
, A

) 0.83 0.88 0.83 0.68 0.59 3.81

d(A
3
, A

) 0.80 0.75 0.83 0.50 0.59 3.47

d(A
4
, A

) 0.59 0.83 0.66 0.61 0.50 3.19

d(A5, A

) 0.52 0.69 0.66 0.50 0.45 2.80

in Table 11. Finally, this calculation led to the outranking presented
in Table 11, meaning that supplier A
5
is the best alternative, fol-
lowed by A
1
, A
2
, A
3
and A
4
, in this order.
4.2. Fuzzy AHP application
Thelinguistic terms presentedinFig. 5wereusedbythedecision
makers to comparatively evaluate the weight of the criteria and the
ratings of the alternatives. Following Chang [10], TFN were used
to specify the linguistic values of these variables, as presented in
Table 12.
Table 11
Outranking of alternative suppliers according to Fuzzy TOPSIS.
Suppliers CC
i
Rank
A
1
0.47 5th
A
2
0.68 1st
A
3
0.62 2nd
A
4
0.56 3rd
A5 0.50 4th
Table 12
Comparative linguistic scale for ratings of alternatives and weights of criteria.
Linguistic terms Fuzzy triangular number
Equally preferable (EQ) (1.0, 1.0, 3.0)
Slightly preferable (SP) (1.0, 3.0, 5.0)
Fairly preferable (FP) (3.0, 5.0, 7.0)
Extremely preferable (XP) (5.0, 7.0, 9.0)
Absolutely preferable (AP) (7.0, 9.0, 9.0)
Table 13 presents the comparative judgments of the weights
of the criteria made by the three decision makers involved already
converted into TFN. The results of aggregation of these fuzzy values
arepresentedinTable14andwereobtainedbythearithmetic mean
of the judgments.
Likewise, the fuzzy values of the aggregated comparative judg-
ments of the alternative suppliers for each criterion made by the
three decision makers are presented in Tables 1519.
Fig. 5. Comparative linguistic scale of the weights of the criteria and ratings of the
alternatives.
F.R. Lima Junior et al. / Applied Soft Computing 21 (2014) 194209 201
Table 13
Comparative judgments of the weights of the criteria made by decision makers.
C
1
C
2
C
3
C
4
C5
DM1
C
1
(1.0, 1.0, 1.0) (0.14, 0.20, 0.33) (1.0, 1.0, 3.0) (7.0, 9.0, 9.0) (1.0, 3.0, 5.0)
C
2
(3.0, 5.0, 7.0) (1.0, 1.0, 1.0) (5.0, 7.0, 9.0) (7.0, 9.0, 9.0) (3.0, 5.0, 7.0)
C
3
(0.33, 1.0, 1.0) (0.11, 0.14, 0.20) (1.0, 1.0, 1.0) (3.0, 5.0, 7.0) (1.0, 3.0, 5.0)
C
4
(0.11, 0.11, 0.14) (0.11, 0.11, 0.14) (0.14, 0.20, 0.33) (1.0, 1.0, 1.0) (0.20, 0.33, 1.0)
C5 (0.20, 0.33, 1.0) (0.14, 0.20, 0.33) (0.20, 0.33, 1.0) (1.0, 3.0, 5.0) (1.0, 1.0, 1.0)
DM2
C
1
(1.0, 1.0, 1.0) (1.0, 3.0, 5.0) (3.0, 5.0, 7.0) (3.0, 5.0, 7.0) (3.0, 5.0, 7.0)
C
2
(0.20, 0.33, 1.0) (1.0, 1.0, 1.0) (3.0, 5.0, 7.0) (1.0, 3.0, 5.0) (1.0, 3.0, 5.0)
C
3
(0.14, 0.20, 0.33) (0.14, 0.20, 0.33) (1.0, 1.0, 1.0) (1.0, 1.0, 3.0) (1.0, 1.0, 3.0)
C
4
(0.14, 0.20, 0.33) (0.20, 0.33, 1.0) (0.33, 1.0, 1.0) (1.0, 1.0, 1.0) (1.0, 1.0, 3.0)
C5 (0.14, 0.20, 0.33) (0.20, 0.33, 1.0) (0.33, 1.0, 1.0) (0.33, 1.0, 1.0) (1.0, 1.0, 1.0)
DM3
C
1
(1.0, 1.0, 1.0) (1.0, 1.0, 3.0) (1.0, 1.0, 3.0) (1.0, 3.0, 5.0) (1.0, 3.0, 5.0)
C
2
(0.33, 1.0, 1.0) (1.0, 1.0, 1.0) (1.0, 1.0, 3.0) (1.0, 3.0, 5.0) (1.0, 3.0, 5.0)
C
3
(0.33, 1.0, 1.0) (0.33, 1.0, 1.0) (1.0, 1.0, 1.0) (1.0, 1.0, 3.0) (1.0, 3.0, 5.0)
C
4
(0.20, 0.33, 1.0) (0.20, 0.33, 1.0) (0.33, 1.0, 1.0) (1.0, 1.0, 1.0) (1.0, 1.0, 3.0)
C5 (0.20, 0.33, 1.0) (0.20, 0.33, 1.0) (0.20, 0.33, 1.0) (0.33, 1.0, 1.0) (1.0, 1.0, 1.0)
Table 14
Fuzzy numbers of the aggregated weights of the criteria.
C
1
C
2
C
3
C
4
C5
C
1
(1.0, 1.0, 1.0) (0.71, 1.40, 2.78) (1.67, 2.33, 4.33) (3.67, 5.67, 7.0) (1.67, 3.67, 5.67)
C
2
(1.18, 2.11, 3.00) (1.0, 1.0, 1.0) (3.00, 4.33, 6.33) (3.0, 5.0, 6.33) (1.67, 3.67, 5.67)
C
3
(0.27, 0.73, 0.78) (0.19, 0.45, 0.51) (1.0, 1.0, 1.0) (1.67, 3.0, 5.0) (1.00, 2.33, 4.33)
C
4
(0.19, 0.44, 0.49) (0.17, 0.26, 0.71) (0.23, 0.51, 0.78) (1.0, 1.0, 1.0) (0.73, 0.78, 2.33)
C5 (0.18, 0.29, 0.78) (0.18, 0.29, 0.78) (0.24, 0.56, 1.00) (0.56, 1.67, 2.33) (1.0, 1.0, 1.0)
Table 15
Fuzzy numbers of the supplier alternative ratings related to criterion C
1
.
A
1
A
2
A
3
A
4
A5
A
1
(1.00, 1.00, 1.00) (0.73, 1.44, 3.00) (0.18, 0.29, 0.78) (0.73, 1.44, 3.00) (0.73, 0.78, 2.33)
A
2
(0.51, 1.44, 2.33) (1.00, 1.00, 1.00) (1.40, 2.11, 3.67) (1.67, 2.33, 4.33) (1.40, 2.78, 4.33)
A
3
(1.67, 3.67, 5.67) (0.49, 1.40, 2.11) (1.00, 1.00, 1.00) (1.40, 2.78, 4.33) (1.40, 2.78, 4.33)
A
4
(0.51, 1.44, 2.33) (0.27, 0.73, 0.78) (0.45, 1.18, 2.11) (1.00, 1.00, 1.00) (1.13, 1.89, 3.00)
A5 (0.56, 1.67, 2.33) (0.45, 1.18, 2.11) (0.45, 1.18, 2.11) (0.71, 2.07, 3.44) (1.00, 1.00, 1.00)
Table 16
Fuzzy numbers of the supplier alternative ratings related to criterion C
2
.
A
1
A
2
A
3
A
4
A5
A
1
(1.00, 1.00, 1.00) (0.43, 1.10, 1.78) (0.47, 1.22, 2.33) (0.45, 0.51, 1.44) (0.73, 0.78, 2.33)
A
2
(4.07, 5.44, 6.33) (1.00, 1.00, 1.00) (1.00, 1.67, 3.67) (0.73, 0.78, 2.33) (1.40, 2.78, 4.33)
A
3
(0.73, 2.11, 3.67) (0.29, 0.78, 1.00) (1.00, 1.00, 1.00) (0.73, 0.78, 2.33) (1.67, 3.00, 5.00)
A
4
(1.44, 3.00, 4.33) (0.56, 1.67, 2.33) (0.56, 1.67, 2.33) (1.00, 1.00, 1.00) (1.67, 3.00, 5.00)
A5 (0.78, 1.67, 3.00) (0.45, 1.18, 2.11) (0.23, 0.51, 0.78) (0.23, 0.51, 0.78) (1.00, 1.00, 1.00)
Table 17
Fuzzy numbers of the supplier alternative ratings related to criterion C
3
.
A
1
A
2
A
3
A
4
A5
A
1
(1.00, 1.00, 1.00) (1.67, 3.00, 5.00) (1.00, 1.67, 3.67) (1.00, 1.67, 3.67) (0.73, 1.44, 3.00)
A
2
(0.22, 0.51, 0.78) (1.00, 1.00, 1.00) (1.00, 1.00, 3.00) (0.73, 0.78, 2.33) (0.73, 1.44, 3.00)
A
3
(0.28, 0.78, 1.00) (0.33, 1.00, 1.00) (1.00, 1.00, 1.00) (1.00, 1.67, 3.67) (0.72, 0.73, 2.11)
A
4
(0.28, 0.78, 1.00) (0.56, 1.67, 2.33) (0.29, 0.78, 1.00) (1.00, 1.00, 1.00) (0.73, 1.44, 3.00)
A5 (0.51, 1.44, 2.33) (0.51, 1.44, 2.33) (1.22, 2.33, 3.00) (0.51, 1.44, 2.33) (1.00, 1.00, 1.00)
Table 18
Fuzzy numbers of the supplier alternative ratings related to criterion C
4
.
A
1
A
2
A
3
A
4
A5
A
1
(1.00, 1.00, 1.00) (0.73, 0.78, 2.33) (1.67, 3.66, 5.67) (0.73, 1.44, 3.00) (1.78, 2.51, 3.44)
A
2
(0.56, 1.66, 2.33) (1.00, 1.00, 1.00) (3.67, 5.00, 7.00) (1.40, 2.77, 4.33) (0.71, 2.06, 3.44)
A
3
(0.18, 0.29, 0.78) (0.20, 0.43, 0.47) (1.00, 1.00, 1.00) (0.73, 0.78, 2.33) (0.47, 0.56, 1.67)
A
4
(0.51, 1.44, 2.33) (0.45, 1.18, 2.11) (0.56, 1.67, 2.33) (1.00, 1.00, 1.00) (0.47, 1.22, 2.33)
A5 (0.44, 1.11, 1.84) (1.13, 1.89, 3.00) (0.78, 2.33, 3.67) (0.73, 2.11, 3.67) (1.00, 1.00, 1.00)
202 F.R. Lima Junior et al. / Applied Soft Computing 21 (2014) 194209
Table 19
Fuzzy numbers of the supplier alternative ratings related to criterion C5.
A
1
A
2
A
3
A
4
A5
A
1
(1.00, 1.00, 1.00) (1.40, 2.11, 3.67) (1.40, 2.11, 3.67) (0.73, 1.44, 3.00) (0.45, 0.51, 1.44)
A
2
(0.49, 1.40, 2.11) (1.00, 1.00, 1.00) (1.00, 1.67, 3.67) (2.07, 2.77, 4.33) (0.46, 0.56, 1.67)
A
3
(0.49, 1.40, 2.11) (0.29, 0.78, 1.00) (1.00, 1.00, 1.00) (0.73, 1.44, 3.00) (0.46, 0.56, 1.67)
A
4
(0.51, 1.44, 2.33) (0.48, 1.38, 2.07) (0.51, 1.44, 2.33) (1.00, 1.00, 1.00) (0.70, 0.71, 2.07)
A5 (1.44, 3.00, 4.33) (0.78, 2.33, 3.67) (0.78, 2.33, 3.67) (1.89, 3.00, 3.67) (1.00, 1.00, 1.00)
Table 20
Consistency Ratios of comparative matrices.
C
1
C
2
C
3
C
4
C5 Weight
DM1 0.16 0.10 0.08 0.13 0.15 0.12
DM2 0.18 0.18 0.19 0.15 0.17 0.09
DM3 0.15 0.16 0.16 0.14 0.17 0.15
The consistency ratios (CR) for each comparative matrix were
calculated according to Saaty [65] and Facchinetti et al. [66] and
are presented in Table 20. As it can be seen, all the values of CR
are below0.20, which conrms the consistency of the comparative
judgments.
The values of the fuzzy synthetic extent for the criteria matrix
are:
S
C1
= (8.71, 14.07, 20.78)
_
1
65.94
,
1
44.48
,
1
27.17
_
= (0.13, 0.32, 0.76)
S
C2
= (9.84, 16.11, 22.33)
_
1
65.94
,
1
44.48
,
1
27.17
_
= (0.15, 0.36, 0.82)
S
C3
= (4.13, 7.51, 11.62)
_
1
65.94
,
1
44.48
,
1
27.17
_
= (0.06, 0.17, 0.42)
S
C4
= (2.32, 2.99, 5.31)
_
1
65.94
,
1
44.48
,
1
27.17
_
= (0.04, 0.07, 0.20)
S
C5
= (2.16, 3.80, 5.89)
_
1
65.94
,
1
44.48
,
1
27.17
_
= (0.03, 0.08, 0.22)
The degrees of possibility of these fuzzy values, computed as in
Eqs. (32) and (33) are:
V(S
C1
S
C2
) = 0.93
V(S
C1
S
C3
) = 1.00
V(S
C1
S
C4
) = 1.00
V(S
C1
S
C5
) = 1.00
V(S
C2
S
C1
) = 1.00
V(S
C2
S
C3
) = 1.00
V(S
C2
S
C4
) = 1.00
V(S
C2
S
C5
) = 1.00
V(S
C3
S
C1
) = 0.67
V(S
C3
S
C2
) = 0.59
V(S
C3
S
C4
) = 1.00
V(S
C3
S
C5
) = 1.00
V(S
C4
S
C1
) = 0.20
V(S
C4
S
C2
) = 0.14
V(S
C4
S
C3
) = 0.57
V(S
C4
S
C5
) = 0.90
V(S
C5
S
C1
) = 0.27
V(S
C5
S
C2
) = 0.20
V(S
C5
S
C3
) = 0.65
V(S
C5
S
C4
) = 1.00
Therefore, the weight vector W

, computed as in Eqs. (35) and

(36), is:
d

(C
1
) = V(S
C1
S
C2
, S
C3
, S
C4
, S
C5
)
= min(0.93, 1.00, 1.00, 1.00) = 0.93
d

(C
2
) = V(S
C2
S
C1
, S
C3
, S
C4
, S
C5
)
= min(1.00, 1.00, 1.00, 1.00) = 1.00
d

(C
3
) = V(S
C3
S
C1
, S
C2
, S
C4
, S
C5
)
= min(0.67, 0.59, 1.00, 1.00) = 0.59
d

(C
4
) = V(S
C4
S
C1
, S
C2
, S
C3
, S
C5
)
= min(0.20, 0.14, 0.57, 0.90) = 0.14
F.R. Lima Junior et al. / Applied Soft Computing 21 (2014) 194209 203
Table 21
Weight vectors of the criteria and alternative suppliers.
C
1
C
2
C
3
C
4
C5
A
1
0.17 0.15 0.24 0.23 0.20
A
2
0.23 0.26 0.18 0.26 0.20
A
3
0.24 0.21 0.18 0.11 0.16
A
4
0.20 0.24 0.18 0.18 0.18
A5 0.16 0.14 0.22 0.22 0.26
Weights of criteria 0.33 0.35 0.21 0.05 0.07
Table 22
Global performance of alternatives and outranking.
Supplier Global performance Rank
A
1
0.80 5th
A
2
1.00 1st
A
3
0.90 2nd
A
4
0.87 3rd
A5 0.81 4th
d

(C
5
) = V(S
C5
S
C1
, S
C2
, S
C3
, S
C4
)
= min(0.27, 0.20, 0.65, 1.00) = 0.20
W

= (0.93, 1.00, 0.59, 0.14, 0.20)

After normalization the weight vector is (0.33, 0.35, 0.21, 0.05,
0.07).
Calculation of the weight vectors for the alternative evaluation
matrices followed the same procedure. The weight vectors from
Tables 1519 are respectively (0.71, 0.97, 1.00, 0.83, 0.69), (0.58,
1.00, 0.80, 0.94, 0.53), (1.00, 0.77, 0.76, 0.78, 0.93), (0.88, 1.00, 0.40,
0.68, 0.82) and (0.80, 0.80, 0.63, 0.69, 1.00).
Table 21summarizes the normalizedweight vectors of the crite-
ria and alternative suppliers.
For supplier alternative A
1
, the global performance was com-
puted as:
D(A
1
) = (d

(A
1
C1
) d

(C
1
) +d

(A
1
C2
) d

(C
2
) +d

(A
1
C3
)
d

(C
3
) +d

(A
1
C4
) d

(C
4
) +d

(A
1
C5
) d

(C
5
)) = 0.60
The global performance for the other alternative suppliers was
computed similarly. Table 22 presents the global performance for
all the alternatives and their ranking position. Therefore, following
this procedure, similarly to the application of Fuzzy TOPSIS, sup-
plier A
5
is the best evaluated alternative, followed by A
1
, A
2
, A
3
and
A
4
, in this order.
5. Comparative analysis of Fuzzy TOPSIS and Fuzzy AHP
methods
Comparison of both methods was based on a set of required
characteristics of the techniques so as to adequately deal with the
problemof supplier selection, as presentedinSection2. The follow-
ing factors were considered: adequacy to changes of alternatives
or criteria; agility in the decision process; computational complex-
ity; adequacy to supporting group decision making; the amount of
alternative suppliers and criteria; and modeling of uncertainty.
5.1. Adequacy to changes of alternatives
In the supplier selection process, the evaluation of a different
set of supply alternatives may require the inclusion or exclusion
of alternatives. In this case, the selection method must produce a
consistent preference order of alternatives.
In the Fuzzy AHP application case, with ve alternatives
and equal weights for all the criteria, the outranking is
A
2
>A
5
>A
1
>A
4
>A
3
, as illustrated in Fig. 6a. To test the Fuzzy AHP
method, an additional supply alternative (A
6
) was evaluated. Five
tests were performed, each one with an additional alternative with
a rating equal to one of the ve existing alternatives. For most of the
tests, the results have shown no signicant changes in the alterna-
tive ranking. However, when the additional supply alternative has
a rating equal to the best alternative (A
2
in Fig. 6a), the resulting
preference order changes considerably. In this case, what was the
worst alternative, A
3
, becomes the best one, as shown in Fig. 6b,
which is not expected in supplier selection problems. This inver-
sion of alternatives, known as ranking reversal, already pointed out
by other studies as a awin the AHP method [6769] also happens
in the Fuzzy AHP. On the other hand, in the Fuzzy TOPSIS applica-
tion case, the same sequence of tests has caused no change at all
to the alternative nal scores. The order of preferences remained
the same in all the tests, with the additional alternative having the
same ranking number as the equal rating alternative.
5.2. Adequacy to changes of criteria
Insomepurchasingsituations, wemayalsoneedtochangesome
of the criteria used to evaluate the suppliers. In this case, the crite-
ria importance order produced by the selection method must be
consistent as well.
In the Fuzzy AHP application case, with the ve criteria and
respective weights, the criteria importance order given by the
method was C
2
>C
1
>C
3
>C
5
>C
4
. To evaluate the effect of adding
a new criterion, ve tests were carried out, each one with the
additional criterion with a weight equal to one of the ve existing
criteria. For most of the tests, the results have shown no signicant
changes in the importance order. Yet, when the additional criterion
Fig. 6. Results of tests of changes of alternatives, Fuzzy AHP.
204 F.R. Lima Junior et al. / Applied Soft Computing 21 (2014) 194209
Fig. 7. Results of tests of changes of criteria, Fuzzy AHP.
has a weight equal to the one of C
5
, then there is an inversion of
the importance order. This indicates that the ranking reversal can
also happen when there is a change of criteria. In the Fuzzy TOPSIS
application case, adding a newcriterion has caused no change at all
to the criteria importance order.
Furthermore, a test was performed to evaluate the effect
of excluding a criterion. The starting point was the criteria
importance order given by the Fuzzy AHP application case
(C
2
>C
1
>C
3
>C
5
>C
4
), as illustrated in Fig. 7a. When the criterion C
5
was excluded, although the importance order was kept the same,
the weight of criterion C
4
was reduced to zero. Because of the com-
parative judgment, the degree of possibility V(S
C4
S
C1
) equals to
zero. As the calculation of the weight vector uses the MIN opera-
tor, the resulting weight of criterion C
4
is null. Consequently this
criterion does not count at all to the evaluation of the alternatives.
The same effect was observed with the exclusion of a second cri-
terion, C
4
. It was observed that this problem of nulling the weight
of a criterion will always happen when the difference between the
synthetic extents of two criteria are large enough such that there
is no intersection between them and consequently the degree of
possibility is zero. Appendix A presents another example in detail
in which 5 criteria are considered for weighting and the problem
of null weight happens for 2 of them. On the other hand, in the
Fuzzy TOPSIS application case, this problem did not happen. This
is due to the fact that the criteria matrix (

W) is computed using
arithmetic mean between fuzzy numbers, which will never lead to
a null weight.
5.3. Agility in the decision process
This factor evaluates the amount of judgments required from
specialists in both methods. Considering n the number of suppli-
ers and m the number of criteria, in the Fuzzy TOPSIS method, m
judgments for each of the n alternatives are required, in addition to
the m judgments related to the weight of the criteria. This can be
expressed as in Eq. (38).
J
TOPSIS
n,m
= m+nm = m(n +1) (38)
In the case of the Fuzzy AHP method, the number of required
judgments for a decision matrix A
ixi
is:
J
A
i,i
= i
i 1
2
(39)
Since there are m matrices of size nn (one for each decision
criterion) in addition to the decision matrix of size mmrelated to
the weight of the criteria, the total number of required judgment is
J
AHP
n,m
= m
m1
2
+m
_
n
n 1
2
_
(40)
Based on Eqs. (38) and (40), Fig. 8 presents the number of judg-
ments for both methods when the number of alternatives and
criteria vary from2 to 9. It can be seen that as the number of criteria
and alternatives increase, the number of required judgment using
Fuzzy AHP is in general greater than that using Fuzzy TOPSIS. In the
applicationcase, the Fuzzy TOPSIS required30judgments while the
Fuzzy AHP required 60 judgments. If there are 9 alternatives and
9 criteria, the Fuzzy AHP requires four times more judgments than
the Fuzzy TOPSIS. On the other hand, when there are few criteria
and alternative (J
22
, J
23
, J
24
, J
32
), the required judgments using
Fuzzy TOPSIS is greater thanwhenusing Fuzzy AHP. Anexceptionis
made to J
33
and J
25
, when both methods require the same num-
ber of judgments. Therefore, it can be said that the Fuzzy TOPSIS
method performs better than the Fuzzy AHP in regard to the level
of interaction with decision makers to data collection. In this sense,
Fuzzy TOPSIS provides greater agility in the decision process than
Fuzzy AHP.
5.4. Computational complexity
The computational complexity of both methods was evaluated
considering only the time complexity. Similarly to Chang [10], the
time complexity, T, was appraised based on the number of times of
multiplications withinthe algorithms. Inthis study, exponentiation
and logical operations were additionally used as a measure of time
complexity.
Considering there are n alternative suppliers and m criteria,
the Fuzzy TOPSIS method requires 3nm operations to compute
Fig. 8. results of tests of agility in the decision process.
F.R. Lima Junior et al. / Applied Soft Computing 21 (2014) 194209 205
Fig. 9. Results of tests of time complexity.
the normalized decision matrix, 3nm operations to compute the
weighted decision matrix and 14nm operations to compute the
distances d
+
i
and d

i
. Therefore, the time complexity, T
n,m
, of the
Fuzzy TOPSIS method is given by Eq. (41).
T
n,m
= 3nm+3nm+7nm+7nm = 20nm (41)
Following the same approach, the Fuzzy AHP method requires
6m(n+1) operations to compute the fuzzy synthetic extent to all
the decision matrices, nm(n 1) +n(n 1) to compute the degrees
of possibility, n(m+1) to normalize the vector W

and nally nm
operations tocompute the global performance. Thus, the time com-
plexity, T
n,m
, of the Fuzzy AHP method is given by Eq. (42).
T

n,m
= 6m(n +1) +nm(n 1) +n(n 1) +n(m+1) +nm
= n
2
(m+1) +m(7n +6) (42)
The graphics in Fig. 9a show the time complexity variation as a
function of number of alternatives for different numbers of criteria
for bothmethods. It canbe seenthat ingeneral Fuzzy AHP performs
better than Fuzzy TOPSIS. In the application case, the Fuzzy TOP-
SIS method required 500 operations while the Fuzzy AHP method
required 355 operations. However, when the consistency tests of
the judgment matrices are performed, the time complexity of the
Fuzzy AHP method, T
n,m
, increases by a factor of 4n(m+1). In this
case, the Fuzzy TOPSIS method performs slightly better than the
Fuzzy AHP as it increases the number of alternatives, although in
most cases Fuzzy AHP still performs better, as shown in Fig. 9b.
In the application case, even with the consistency test, the Fuzzy
AHP method required 475 operations, slightly less than the Fuzzy
TOPSIS.
5.5. Adequacy to supporting group decision making
Both methods allowaggregation of judgments of more than one
decision maker. In the case of the Fuzzy TOPSIS method, aggrega-
tion of different judgments is made according to Eqs. (12) and (13)
for the weights of the criteria and the ratings of the alternative sup-
pliers. In the case of the Fuzzy AHP, although this is not explicitly
considered in the method proposed by Chang [10], he suggests that
aggregation be made using the arithmetic mean of the judgments.
Since the amount of data required by the Fuzzy AHP method
is greater than that required by the Fuzzy TOPSIS, increasing
the number of decision makers will consequently cause a larger
increase in the time complexity of the Fuzzy AHP when compared
with the TOPSIS method. Therefore, although both methods sup-
port group decision making, due to the impact on time complexity,
the Fuzzy TOPSIS method is preferable.
Eventhoughbothmethods compute aggregationbasedonfuzzy
arithmetic mean, an alternative approach would be to weight the
judgments of the different decision makers and aggregate the data
bycomputing a weightedmean. For instance, the procurement staff
is better able to judge the performance of suppliers and therefore
their judgments should be more relevant than the judgments of
others not so involved with procurement.
5.6. The number of alternative suppliers and criteria
The Fuzzy TOPSIS method does not impose any restriction on
the number of alternatives or criteria used in the selection pro-
cess. On the other hand, the comparative analysis of the Fuzzy
AHP method imposes some limitation on the number of criteria
and alternatives. Saaty [65] suggests that the number criteria or
alternatives to be compared using AHP be limited to nine so as not
to compromise human judgment and its consistency. This sugges-
tion applies equally to the Fuzzy AHP method. In the application
case, withvecriteriaandvealternatives, theuseof theFuzzyAHP
method was perfectly viable. Although the limitation of the num-
ber of criteria can be alleviated by deploying the criteria into the
Fuzzy AHP hierarchy structure, the number of alternatives imposes
a real limitation. Therefore, the choice of the method depends on
the particularities of the circumstances at hand. For instance, when
selecting a new supplier for a new product, with many potential
suppliers, the Fuzzy TOPSIS is a better choice.
5.7. Modeling of uncertainty
Both methods utilize fuzzy set theory to deal with the inherent
lack of precision of the data used in the supplier selection deci-
sion process. In both methods the fuzzy number morphology is the
main resource for quantifying imprecision. Due to the vagueness of
judgments of qualitative variables, the parameters of the triangu-
lar membership functions can be chosen so as to better represent
the linguistic terms used by each decision maker to evaluate the
alternatives regarding different decision criteria.
206 F.R. Lima Junior et al. / Applied Soft Computing 21 (2014) 194209
Table 23
Comparative linguistic judgments for replacing of supplier.
Regarding C
1
A
1
A
2
Regarding C
3
A
1
A
2
A
1
(1.0, 1.0, 1.0) (0.2, 0.33, 1.0) A
1
(1.0, 1.0, 1.0) (1.0, 1.0, 3.0)
A
2
(1.0, 3.0, 5.0) (1.0, 1.0, 1.0) A
2
(0.33, 1.0, 1.0) (1.0, 1.0, 1.0)
Regarding C
2
A
1
A
2
Regarding C
4
A
1
A
2
A
1
(1.0, 1.0, 1.0) (0.11, 0.14, 0.14) A
1
(1.0, 1.0, 3.0) (0.2, 0.33, 1.0)
A
2
(7.0, 9.0, 9.0) (1.0, 1.0, 1.0) A
2
(1.0, 3.0, 5.0) (1.0, 1.0, 1.0)
Table 24
Summarized comparative analysis of Fuzzy TOPSIS and Fuzzy AHP.
Parameters of
comparison
Comparison of Fuzzy TOPSIS and Fuzzy AHP
Adequacy to changes
of alternatives
Fuzzy AHP is prone to ranking reversal when
including a new alternative, while Fuzzy TOPSIS
produces consistent preference order
Adequacy to changes
of criteria
In Fuzzy AHP ranking reversal also happens when
a new criteria is included. Fuzzy TOPSIS produces
consistent important order. Fuzzy AHP can yield
null weight for selected criteria, while Fuzzy
TOPSIS never leads to a null weight
Agility in the
decision process
Fuzzy TOPSIS performs better than Fuzzy AHP in
most cases except when there are very few criteria
and suppliers
Time complexity Fuzzy AHP performs better than Fuzzy TOPSIS in
most cases. If Fuzzy AHP consistency test is
included, as it increases the number of
alternatives, then Fuzzy TOPSIS surpass Fuzzy AHP
Support to group
decision making
Adequate for both methods. Aggregation based on
fuzzy arithmetic mean. Alternative approach could
be based on weighted mean
Number of criteria
and alternative
suppliers
No Limitation for Fuzzy TOPSIS. Fuzzy AHP limit
the number of criteria and alternative. Fuzzy AHP
allows inclusion of subcriteria into a hierarchy
structure
Modeling of
uncertainty
Both methods are adequate to deal with
imprecision and subjectivity in supplier selection
problems. But Fuzzy AHP is more appropriate
when the purpose is to replace a supplier
In the case of Fuzzy AHP method, the use of pairwise com-
parisons by means of comparative linguistic variables is itself a
way to deal with imprecision. This feature makes this method
more appropriate than the Fuzzy TOPSIS when the purpose is to
replace a supplier. As an example, Table 23 presents the judg-
ments of two suppliers when compared in respect to four criteria
of equal weights aiming at evaluating the benet of replacing A
1
for A
2
. In this case, the supplier global performance given as a rel-
ative measure facilitates the interpretation of the decision makers
(equivalent to 0.32 for A
1
and 0.68 for A
2
). Other advantages of the
Fuzzy AHP compared to Fuzzy TOPSIS in this example are fewer
judgments and less computational complexity.
6. Conclusion
This paper presented a new study comparing the Fuzzy AHP
and the Fuzzy TOPSIS methods in regard to seven factors that
are particularly relevant to the problem of supplier selection.
This paper also presented the application of both methods to a
case of supplier selection, in order to illustrate and clarify the
use of these techniques for the problem of supplier selection. The
comparative evaluation of the techniques in respect to changes
of alternatives or criteria, agility and computational complexity
was based on computational tests considering several scenarios
of supplier selection. The performance of the methods concerning
changes of alternative or criteria was evaluated through ve tests
based on inclusion and exclusion of alternative or criteria. As for
agility and computational complexity, assessment was based on
64 tests considering supplier selection scenarios ranging from
2 to 9 alternatives and criteria. Comparison of the adequacy to
supporting group decision making was based on the analysis of
equations of both methods. For the other factors, comparison was
based on qualitative analysis of the algorithms of both methods.
The comparative analysis of Fuzzy AHP and Fuzzy TOPSIS
has shown some interesting outcomes that one should take into
account so as to better align the technique to the particularities of
the problem at hand. The obtained results concerning the analysis
of the seven factors are valid for the context of supplier selection.
For other decision making problems, changes of alternatives or
criteria, agility and computational complexity may also be relevant
and therefore the conclusions are also applicable to them.
Table 24 presents a summary of the ndings. Regarding the fac-
tor adequacy to changes of alternatives and criteria, it can be seen
that in some situations Fuzzy AHP causes the effect known as rank-
ing reversal, changing the preference order of alternatives and the
importance order of criteria. Ertugrul and Karakasoglu [24] point
this effect when a non-optimal alternative is introduced. However,
this study has shown by a numerical example that the ranking
reversal inFuzzy AHPalsohappens whenanoptimumalternative is
introduced. On the other hand, Fuzzy TOPSIS produces very consis-
tent results. Further research could explore alternative approaches
to avoid the rank reversal in Fuzzy AHP. Another problem caused
by Fuzzy AHP related to the criteria importance order is nulling the
weight of the least important criterion, because of the MINoperator
used in the computation of the degree of possibility.
Concerning the agility inthe decisionprocess, Fuzzy TOPSIS per-
forms better than Fuzzy AHP in most cases except when there are
very fewcriteria andsuppliers. Inaddition, the increase inthe num-
ber of supplier alternatives imposes some limitation to Fuzzy AHP.
As for the Fuzzy TOPSIS, this is not a restriction to the use of the
method. Inthe case of the number of criteria, the intrinsic limitation
imposed by the Fuzzy AHP method can be overcome by deploying
thecriteriaintotheFuzzyAHPhierarchystructure. At thesametime
that the Fuzzy TOPSIS does not constrain the number of criteria, it
does not allowthedeployment of thecriteriaintosubcriteria, which
can be understood as a weakness of the method when applied to
the problemof supplier selection. Afurther studycouldfocus onthe
adaptation of the Fuzzy TOPSIS so as to accommodate the criteria
and subcriteria into the decision matrix.
As for the time complexity, it is in general lower for Fuzzy AHP
than for Fuzzy TOPSIS. However, if the Fuzzy AHP decision matrix
consistency test is performed, which is frequently needed, than the
advantage of the Fuzzy AHP method is not so pronounced. This
conclusiondiffers fromthat made by Ertugrul and Karakasoglu [24]
who states with no further detail that Fuzzy AHP requires more
complex computations than Fuzzy TOPSIS.
Both methods adequately support group decision making. It is
worth to mention that weighted mean could be used to aggre-
gate judgments instead of the arithmetic mean commonly used. By
doing that, one could give different importance to different deci-
sion makers. Although both methods are equally adequate to deal
with the lack of precision of scores of alternatives as well as the
F.R. Lima Junior et al. / Applied Soft Computing 21 (2014) 194209 207
relative importance of different criteria, it is worth noting that the
Fuzzy AHP is more appropriate than the Fuzzy TOPSIS when the
purpose is to replace a supplier.
Finally, some genuine contributions of this study can be pointed
out:

It is the rst study to analyze the adequability of MCDM meth-

ods to the problem of supplier selection taking into account the
alignment of the particularities of the problem with the charac-
teristics of the techniques. A study such as this can contribute to
helping researchers and practitioners to choose more effective
approaches to supplier selection;

It complements the study by Ertugrul and Karakasoglu [24] not

only by considering another problem domain but also by includ-
ing numerical examples to comparatively test the techniques. It
alsoincludes other comparativecriteriasuchas agilityinthedeci-
sionprocess, modelingof uncertaintyandadequacytosupporting
group decision making;

Apart fromthe comparative analysis, another contribution of this

study is the proposition of a set of seven factors for evaluation of
MCDM methods. In this sense, this set of factors can be further
used as a framework to assess the adequacy of other techniques
to the problem of supplier selection;

It is the rst study to discuss and bring numeric examples of

the problem of null weight of criteria of the Fuzzy AHP method.
Further research could test other fuzzy operators such as the
arithmetic mean and T-norms in order to avoid this problem.
Acknowledgments
To FAPESP (So Paulo State Science Foundation) and CAPES for
supporting this research project. To the anonymous reviewers for
their helpful contributions.
Appendix A. Example of the problem of null weight in
Fuzzy AHP
Tables A1 and A2 present the comparative judgments of the
importanceof 5criteria. TableA1presents thejudgments inlinguis-
tic terms and Table A2 shows themconverted to the corresponding
triangular fuzzy numbers, as presented in Table 12 of the paper.
The consistency ratio for this matrix is 0.15.
The values of the fuzzy synthetic extent for the criteria matrix
are:
S
C1
= (10.14, 14.20, 18.33)
_
1
75.48
,
1
57.96
,
1
37.69
_
= (0.13, 0.24, 0.47)
Table A1
Matrix for comparative judgments of 5 criteria in linguistic terms.
C
1
C
2
C
3
C
4
C5
C
1
1/FP EQ AP SP
C
2
XP AP FP
C
3
FP SP
C
4
1/SP
C5
S
C2
= (19.0, 27.0, 33.0)
_
1
75.48
,
1
57.96
,
1
37.69
_
= (0.25, 0.46, 0.85)
S
C3
= (5.44, 10.14, 14.20)
_
1
75.48
,
1
57.96
,
1
37.69
_
= (0.07, 0.17, 0.36)
S
C4
= (1.56, 1.75, 2.62)
_
1
75.48
,
1
57.96
,
1
37.69
_
= (0.02, 0.03, 0.06)
S
C5
= (1.54, 4.86, 7.33)
_
1
75.48
,
1
57.96
,
1
37.69
_
= (0.03, 0.08, 0.21)
The degrees of possibility for the criteria for the ve criteria are:
V(S
C1
S
C2
) = 0.50
V(S
C1
S
C3
) = 1.00
V(S
C1
S
C4
) = 1.00
V(S
C1
S
C5
) = 1.00
V(S
C2
S
C1
) = 1.00
V(S
C2
S
C3
) = 1.00
V(S
C2
S
C4
) = 1.00
V(S
C2
S
C5
) = 1.00
V(S
C3
S
C1
) = 0.77
V(S
C3
S
C2
) = 0.29
V(S
C3
S
C4
) = 1.00
V(S
C3
S
C5
) = 1.00
V(S
C4
S
C1
) = 0.00
V(S
C4
S
C2
) = 0.00
V(S
C4
S
C3
) = 0.00
V(S
C4
S
C5
) = 0.39
Table A2
Matrix for comparative judgments of 5 criteria in fuzzy numbers.
C
1
C
2
C
3
C
4
C5
C
1
(1.00, 1.00, 1.00) (0.14, 0.20, 0.33) (1.00, 1.00, 3.00) (7.00, 9.00, 9.00) (1.00, 3.00, 5.00)
C
2
(3.00, 5.00, 7.00) (1.00, 1.00, 1.00) (5.00, 7.00, 9.00) (7.00, 9.00, 9.00) (3.00, 5.00, 7.00)
C
3
(0.33, 1.00, 1.00) (0.11, 0.14, 0.20) (1.00, 1.00, 1.00) (3.00, 5.00, 7.00) (1.00, 3.00, 5.00)
C
4
(0.11, 0.11, 0.14) (0.11, 0.11, 0.14) (0.14, 0.20, 0.33) (1.00, 1.00, 1.00) (0.20, 0.33, 1.00)
C5 (0.20, 0.33, 1.00) (0.14, 0.20, 0.33) (0.20, 0.33, 1.00) (1.00, 3.00, 5.00) (1.00, 1.00, 1.00)
208 F.R. Lima Junior et al. / Applied Soft Computing 21 (2014) 194209
V(S
C5
S
C1
) = 0.34
V(S
C5
S
C2
) = 0.00
V(S
C5
S
C3
) = 0.61
V(S
C5
S
C4
) = 1.00
Therefore, the weight vector W

is:
d

(C
1
) = V(S
C1
S
C2
, S
C3
, S
C4
, S
C5
)
= min(0.50, 1.00, 1.00, 1.00) = 0.50
d

(C
2
) = V(S
C2
S
C1
, S
C3
, S
C4
, S
C5
)
= min(1.00, 1.00, 1.00, 1.00) = 1.00
d

(C
3
) = V(S
C3
S
C1
, S
C2
, S
C4
, S
C5
)
= min(0.77, 0.29, 1.00, 1.00) = 0.29
d

(C
4
) = V(S
C4
S
C1
, S
C2
, S
C4
, S
C5
)
= min(0.00, 0.00, 0.00, 0.39) = 0.00
d

(C
5
) = V(S
C5
S
C1
, S
C2
, S
C3
, S
C4
)
= min(0.34, 0.00, 0.61, 1.00) = 0.00
W

= (0.50.1.00, 0.29, 0.00, 0.00)

After normalization the weight vector is (0.28, 0.56, 0.16, 0.00,
0.0). This weight vector shows criteria C
4
and C
5
as having null
weight which does not correspond to the comparative judgment.
References
[1] C. Chen, C. Lin, S. Huang, Afuzzy approachfor supplier evaluationand selection,
Int. J. Prod. Econ. 102 (2006) 289301.
[2] W.A.S. Castro, O.D.C. Gomez, L.F.O. Franco, Seleccion de proveedores: uma
aproximacion al estado del arte, Cuaderno de Administracin 22 (2009)
145167.
[3] P. Thrulogachantar, S. Zailani, The inuence of purchasing strategies on manu-
facturing performance, J. Manuf. Technol. Manage. 22 (2011) 641663.
[4] B. Ha, Y. Park, S. Cho, Suppliers affective trust andtrust incompetency inbuyer,
Int. J. Oper. Prod. Manage. 31 (2011) 5677.
[5] L. De Boer, E. Labro, P. Morlacchi, A review of methods supporting supplier
selection, Eur. J. Purch. Supply Manage. 7 (2001) 7589.
[6] C. Wu, D. Barnes, A literature review of decision-making models and
approaches for partner selectioninagilesupplychains, J. Purch. SupplyManage.
17 (2011) 256274.
[7] L. De Boer, L.V.D. Wegen, J. Telgen, Outranking methods in support of supplier
selection, Eur. J. Purch. Supply Manage. 4 (1998) 109118.
[8] A. Amindoust, S. Ahmed, A. Saghania, A. Bahreininejad, Sustainable supplier
selection: a ranking model based onfuzzy inference system, Appl. Soft Comput.
J. 12 (2012) 16681677.
[9] J. Chai, J.N.K. Liu, E.W.T. Ngai, Applicationof decision-makingtechniques insup-
plier selection: a systematic review of literature, Expert Syst. Appl. 40 (2013)
38723885.
[10] D.Y. Chang, Applications of the extent analysis method on fuzzy-AHP, Eur. J.
Oper. Res. 95 (1996) 649655.
[11] C. Kahraman, U. Cebeci, Z. Ulukan, Multi-criteria supplier selection using fuzzy,
Logist. Inform. Manage. 16 (2003) 382394.
[12] F.T.S. Chan, N. Kumar, Global supplier development considering risk factors
using fuzzy extended AHP-based approach, Omega 35 (2007) 417431.
[13] A.H.I. Lee, A fuzzy supplier selection model with the consideration of benets,
opportunities, costs and risks, Expert Syst. Appl. 36 (2009) 28792893.
[14] I. Chamodrakas, D. Batis, D. Martakos, Supplier selection in electronic market-
places using satisfying and Fuzzy-AHP, Expert Syst. Appl. 37 (2010) 490498.
[15] M. Zeydan, C. C olpan, C. C obanoglu, A combined methodology for sup-
plier selection and performance evaluation, Expert Syst. Appl. 38 (2011)
27412751.
[16] O. Kilincci, S.A. Onal, Fuzzy-AHP approach for supplier selection in a washing
machine company, Expert Syst. Appl. 38 (2011) 96569664.
[17] C.T. Chen, Extensions of the TOPSIS for group decision-making under fuzzy
environment, Fuzzy Sets Syst. 114 (2000) 19.
[18] M. Dagdeviren, S. Yavuz, N. Kilinci, Weapon selection using the AHP and
TOPSIS methods under fuzzy environment, Expert Syst. Appl. 36 (2009)
81438151.
[19] G. Bykzkan, M.S. Ersoy, Applying fuzzy decision making approach to IT
outsourcing supplier selection, Eng. Technol. 55 (2009) 411415.
[20] A. Awasthi, S.S. Chauhan, S.K. Goyal, A fuzzy multicriteria approach for eval-
uating environmental performance of suppliers, Int. J. Prod. Econ. 126 (2010)
370378.
[21] C. Liao, H. Kao, An integrated fuzzy-TOPSIS and MCGP approach to sup-
plier selection in supply chain management, Expert Syst. Appl. 38 (2011)
1080310811.
[22] F. Jolai, S.A. Yazdian, K. Shahanaghi, M.A. Khojasteh, Integrating fuzzy TOPSIS
and multi-period goal programming for purchasing multiple products from
multiple suppliers, J. Purch. Supply Manage. 17 (2011) 4253.
[23] A. Zouggari, L. Benyoucef, Simulation based fuzzy TOPSIS approach for group
multi-criteria supplier selection, Eng. Appl. Artif. Intell. 25 (3) (2012) 507519.
[24] I. Ertugrul, N. Karakasoglu, Comparisonof fuzzy AHP andfuzzy TOPSIS methods
for facility location selection, Int. J. Adv. Manuf. Technol. 39 (2008) 783795.
[25] J.W.M. Bertrand, J. Fransoo, Operations management research methodologies
using quantitative modeling, Int. J. Oper. Prod. Manage. 22 (2002) 241264.
[26] C.W. Faris, P.J. Robinson, Y. Wind, Industrial Buying and Creative Marketing,
Allyn & Bacon, Boston, 1967.
[27] P. Kraljic, Purchasing must become supply management, Harvard Bus. Rev.
(September/October) (1983) 109117.
[28] M. Frdell, Criteria for achieving efcient contractorsupplier relations, Eng.
Constr. Archit. Manage. 18 (2011) 381393.
[29] V.R. Kannan, K.C. Tan, Supplier selection and assessment: their impact on busi-
ness performance, J. Supply Chain Manage. 38 (2002) 1121.
[30] C.C. Katsikeas, N.G. Paparoidamis, E. Katsikea, Supply source selection criteria:
the impact of supplier performance on distributor performance, Ind. Market.
Manage. 33 (2004) 755764.
[31] K. Kirytopoulos, V. Leopoulos, D. Voulgaridou, Supplier selection in pharma-
ceutical industry: an analytic network process approach, Benchmark. Int. J. 15
(2008) 494516.
[32] A.F. Guneri, A. Yucel, G. Ayyildiz, An integrated fuzzy-LP approach for a supplier
selection problem in supply chain management, Expert Syst. Appl. 36 (2009)
92239228.
[33] C. Shen, K. Yu, Enhancing the efcacy of supplier selection decision-making on
the initial stage of newproduct development: a hybrid fuzzy approach consid-
ering the strategic and operational factors simultaneously, Expert Syst. Appl.
36 (2009) 1127111281.
[34] S.M. Ordoobadi, Development of a supplier selection model using fuzzy logic,
Supply Chain Manage. Int. J. 14 (2009) 314327.
[35] F.E. Boran, S. Genc , M. Kurt, D. Akay, A multi-criteria intuitionistic fuzzy group
decision making for supplier selection with TOPSIS method, Expert Syst. Appl.
36 (2009) 1136311368.
[36] S.H. Amin, J. Razmi, Anintegratedfuzzy model for supplier management: a case
study of ISP selection and evaluation, Expert Syst. Appl. 36 (2009) 86398648.
[37] C. Ku, C. Chang, H. Ho, Global supplier selection using fuzzy analytic hierarchy
process and fuzzy goal programming, Qual. Quant. 44 (2010) 623640.
[38] W. Wang, A fuzzy linguistic computing approach to supplier evaluation, Appl.
Math. Model. 34 (2010) 31303141.
[39] C. Lin, C. Chen, Y. Ting, An ERP model for supplier selection in electronics
industry, Expert Syst. Appl. 38 (2011) 17601765.
[40] G. Bykzkan, G. C ifc i, A novel fuzzy multi-criteria decision framework for
sustainable supplier selection with incomplete information, Comput. Ind. 62
(2011) 164174.
[41] W. Ho, X. Xu, P.K. Dey, Multi-criteria decision making approaches for supplier
evaluationandselection: aliteraturereview, Eur. J. Oper. Res. 202(2010) 1624.
[42] L.A. Zadeh, Fuzzy sets, Inform. Control 8 (1965) 338353.
[43] M.C.Y. Tam, V.N.R. Tummala, An application of the AHP in vendor selection of
a telecommunications system, Omega 29 (2001) 171182.
[44] M. Hudym cov, M. Benkov, J. Pcsov, T.

Skovrnek, Supplier selection based
on multi-criterial AHP method, Acta Montanistica Slovaca 15 (2010) 249255.
[45] C. Gencer, D. Grpinar, Analytic network process in supplier selection: a case
study in an electronic rm, Appl. Math. Model. 31 (2007) 24752486.
[46] R.F. Saen, Restricting weights in supplier selection decisions in the presence of
dual-role factors, Appl. Math. Model. 34 (2010) 28202830.
[47] D.A.R. Carrera, V. Mayorga, Supply chain management: a modular fuzzy infer-
ence system approach in supplier selection for new product development, Int.
J. Intell. Manuf. 19 (2008) 112.
[48] B. Hsu, C. Chiang, M. Shu, Supplier selection using fuzzy quality data and their
applications to touch screen, Expert Syst. Appl. 37 (2010) 61926200.
[49] Z. Liao, J. Rittscher, Integration of supplier selection, procurement lot sizing
and carrier selection under dynamic demand conditions, Int. J. Prod. Econ. 107
(2007) 502510.
[50] G.A. Keskin, S. Ilhan, C. zkan, The fuzzy ART algorithm: a categorization
method for supplier evaluation and selection, Expert Syst. Appl. 37 (2010)
12351240.
F.R. Lima Junior et al. / Applied Soft Computing 21 (2014) 194209 209
[51] S. nut, S.S. Kara, E. Isik, Long term supplier selection using a combined fuzzy
MCDM approach: a case study for a telecommunication company, Expert Syst.
Appl. 36 (2009) 38873895.
[52] S. Vinodh, R.A. Ramiya, S.G. Gautham, Application of fuzzy analytic network
process for supplier selection in a manufacturing organization, Expert Syst.
Appl. 38 (2011) 272280.
[53] S.I. Omurca, An intelligent supplier evaluation, selection and development sys-
tem, Appl. Soft Comput. 13 (2013) 690697.
[54] M. Bevilacqua, F.E. Ciarapica, G. Giacchetta, A fuzzy-QFD approach to supplier
selection, J. Purch. Supply Manage. 12 (2006) 1427.
[55] M. Dursun, E.E. Karsak, A QFD-based fuzzy MCDM approach for supplier selec-
tion, Appl. Math. Model. 37 (2013) 58645875.
[56] A.A. Amid, S.H. Ghodsypour, C.O. Brien, Fuzzy multiobjective linear model
for supplier selection in supply chain, Int. J. Prod. Econ. 104 (2008)
394407.
[57] F. Arikan, A fuzzy solution approach for multi objective supplier selection,
Expert Syst. Appl. 40 (2013) 947952.
[58] K. Shahanaghi, S.A. Yazdian, Vendor selection using a new fuzzy group TOPSIS
approach, J. Uncertain Syst. 3 (2009) 221231.
[59] E. Bottani, A. Rizzi, A fuzzy-TOPSIS methodology to support out-
sourcing of logistics services, Supply Chain Manage. Int. J. 11 (2006)
294308.
[60] A. Shemshadi, H. Shirazi, M. Toreihi, M.J. Tarokh, A fuzzy VIKOR method for
supplier selection based on entropy measure for objective weighting, Expert
Syst. Appl. 38 (2011) 1216012167.
[61] M.R. Mahmoud, L.A. Garcia, Comparison of different multicriteria evaluation
methods for the Red Bluff diversion dam, Environ. Model. Softw. 15 (2000)
471478.
[62] H.J. Zimmermann, Fuzzy Set Theory and Its Applications, second ed., Kluwer
Academic, Boston, 1991.
[63] L.A. Zadeh, Outline of a new approach to the analysis of complex systems and
decision processes, IEEE Trans. Syst. Man Cybernet. 3 (1973) 2844.
[64] W. Pedrycz, F. Gomide, Fuzzy Systems Engineering Toward Human-centric
Computing, Wiley, New Jersey, 2007.
[65] T.L. Saaty, The Analytic Hierarchy Process, rst ed., McGraw Hill, New York,
1980.
[66] G. Facchinetti, R.G. Ricci, S. Muzzioli, Note onranking fuzzy triangular numbers,
Int. J. Intell. Syst. 13 (1998) 613622.
[67] V. Belton, A.E. Gear, Onashort-comingof Saatys methodof analytic hierarchies,
Omega 11 (1983) 228230.
[68] T.L. Saaty, Making and validating complex decisions with the AHP/ANP, J. Syst.
Sci. Syst. Eng. 14 (2005) 136.
[69] S. Zahir, Normalisation and rank reversals in the additive analytic hierarchy
process: a new analysis, Int. J. Oper. Res. 4 (2009) 446467.