Expert Systems with Applications 36 (2009) 8143–8151

Contents lists available at ScienceDirect

Expert Systems with Applications

journal homepage: www.elsevier.com/locate/eswa

Weapon selection using the AHP and TOPSIS methods under fuzzy environment q
Metin Dağdeviren a,*, Serkan Yavuz b, Nevzat Kılınç c
a
Department of Industrial Engineering, Gazi University, 06570 Maltepe, Ankara, Turkey
b
Department of Technology Management, Defense Sciences Institute, 06100, Ankara, Turkey
c
Department of Technical and Project Management, Land Forces Command, 06100 Bakanliklar, Ankara, Turkey

a r t i c l e i n f o a b s t r a c t

Keywords: The weapon selection problem is a strategic issue and has a significant impact on the efficiency of defense
Weapon selection systems. On the other hand, selecting the optimal weapon among many alternatives is a multi-criteria
Multi criteria decision-making decision-making (MCDM) problem. This paper develops an evaluation model based on the analytic hier-
AHP archy process (AHP) and the technique for order performance by similarity to ideal solution (TOPSIS), to
TOPSIS
help the actors in defence industries for the selection of optimal weapon in a fuzzy environment where
Fuzzy set theory
Fuzzy TOPSIS
the vagueness and subjectivity are handled with linguistic values parameterized by triangular fuzzy
numbers. The AHP is used to analyze the structure of the weapon selection problem and to determine
weights of the criteria, and fuzzy TOPSIS method is used to obtain final ranking. A real world application
is conducted to illustrate the utilization of the model for the weapon selection problem. The application
could be interpreted as demonstrating the effectiveness and feasibility of the proposed model.
Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction on manufacturers or vendors, to carry out. For a proper and effec-

tive evaluation, the decision-maker may need a large amount of
One of the most important components of a country’s national data to be analyzed and many factors to be considered (Ayağ &
power is military power and The Armed Forces is the application Özdemir, 2006). Although weapon selection plays an important
instrument of deterrent military power. Regardless from the Forces role in the design of an effective defense system, the publications
(Army, Navy, Air Force, Gendarmerie), the most basic weapon used on this subject are limited. Cheng & Mon (1994) assessed weapon
in the operations of the Armed Forces is infantry rifle. Acquisition systems with AHP based on fuzzy scales. Chen (1996) applied fuzzy
of the infantry rifle satisfying the performance, cost and user- set theory to evaluate weapon systems using fuzzy arithmetic
friendliness (ease of usage) requirements of the user and equipping operations. Cheng & Lin (2002) applied fuzzy decision theory to
the individual fighter with this weapon is a critical factor in evaluate main battle tanks. Mon, Cheng, & Lin (1994) utilized fuzzy
enhancing the operational efficiency of World’s Armed Forces AHP based process on entropy weight to assess weapon systems.
which carry out operations ranging from conventional war to Cheng (1996, 1999) measured naval tactical missile systems and
peace-keeping operations. Another issue increasing the impor- weapon systems with fuzzy AHP and ranking fuzzy numbers,
tance of the infantry rifle used and to be used in the 21st century, respectively. Cheng, Yang, & Hwang (1999) applied AHP to evaluate
is the above-mentioned diversity of operations and the rapid attack helicopters using linguistic variables.
improving technologies which lead to better infantry rifles com- As stated by Cheng (1996, 1999), the performance evaluation
pared to old ones. Owing to the ‘‘21st Century’s Soldier Moderniza- and optimal selection of weapon systems have multi-level and
tion Programmes”, there is plenty of infantry rifle projects either multi-factor features, so such difficulties can be regarded as multi-
on development or production phase by various countries. ple criteria decision-making (MCDM) (Wang & Chang, 2007). Sev-
A proper weapon selection is a very important issue for defense eral methods exist for MCDM (Vincke, 1992; Zeleny, 1982). There
industries due to the fact that improper weapon selection can neg- are no better or worse techniques, but some techniques better suit
atively affect the overall performance and productivity of a defense to particular decision problems than others do (Mergias, Mousta-
system. Selecting the new weapon is a time-consuming and diffi- kas, Papadopoulos, & Loizidou, 2007). The advantage of these
cult process, requiring advanced knowledge and deep experience. methods is that they can account for both financial and non-finan-
So, the process can be a hard task for managers, and also for weap- cial impacts. Among these methods, the most popular ones are
scoring models (Nelson, 1986), analytic hierarchy process (AHP)
q
(Göleç & Tasßkın, 2007), analytic network process (ANP) (Yuksel &
Opinions and results in this paper are of the writers’ and do not reflect any
Dağdeviren, 2007), axiomatic design (AD) (Kulak & Kahraman,
official opinion of Turkish Armed Forces.
* Corresponding author. Tel.: +90 3122317400/2821; fax: +90 3122308434. 2005), utility models (Munoz & Sheng, 1995), TOPSIS (Kahraman,
E-mail address: [email protected] (M. Dağdeviren). Büyüközkan, & Atesß, 2007), ELECTRE (Wang & Triantaphyllou,

0957-4174/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.eswa.2008.10.016
8144 M. Dağdeviren et al. / Expert Systems with Applications 36 (2009) 8143–8151

2005) and PROMETHEE (Dağdeviren, 2008). It is essential to devel- objectives, criteria and alternatives are arranged in a hierarchical
op all elements related to the situation in detail before selecting an structure similar to a family tree. A hierarchy has at least three lev-
appropriate MCDM method to solve the problem under consider- els: overall goal of the problem at the top, multiple criteria that de-
ation (Bufardi, Gheorghe, Kiritsis, & Hirouchakis, 2004; Mergias fine alternatives in the middle, and decision alternatives at the
et al., 2007). The MCDM method choice decision should wait until bottom (Albayrak & Erensal, 2004).
the analyst and the decision-makers understand the problem, the The second step is the comparison of the alternatives and the
feasible alternatives, different outcomes, conflicts between the cri- criteria. Once the problem has been decomposed and the hierarchy
teria and level of the data uncertainty (Mergias et al., 2007). is constructed, prioritization procedure starts in order to determine
The main objective of this study is to propose a systematic eval- the relative importance of the criteria within each level. The pair-
uation model to help the actors in Defence Industries for the selec- wise judgment starts from the second level and finishes in the low-
tion of an optimal weapon among a set of available alternatives. est level, alternatives. In each level, the criteria are compared
Weapon selection problem is both a MCDM problem where many pairwise according to their levels of influence and based on the
criteria should be considered in decision-making, and a problem specified criteria in the higher level (Albayrak & Erensal, 2004).
containing subjectivity, uncertainty and ambiguity in assessment In AHP, multiple pairwise comparisons are based on a standardized
process. Therefore, this study utilizes a MCDM method (AHP) to comparison scale of nine levels (Table 1).
determine the importance weights of evaluation criteria, and fuzzy Let C = {Cj | j = 1, 2, . . . , n} be the set of criteria. The result of the
TOPSIS to obtain the performance ratings of the feasible alterna- pairwise comparison on n criteria can be summarized in an (n  n)
tives in linguistic values parameterized with triangular fuzzy num- evaluation matrix A in which every element aij (i,j = 1,2, . . . ,n) is the
bers. This approach is employed for four reasons: (a) TOPSIS logic is quotient of weights of the criteria, as shown:
rational and understandable; (b) the computation processes are 2 3
straightforward; (c) the concept permits the pursuit of best alter- a11 a12 . . . a1n
natives for each criterion depicted in a simple mathematical form, 6a a22 . . . a2n 7
6 21 7
and (d) the importance weights are incorporated into the compar- A¼6
6 .. .. .. .. 7
7; aii ¼ 1; aji ¼ 1=aij ; aij –0: ð1Þ
ison procedures (Wang & Chang, 2007). Besides, fuzzy TOPSIS has 4 . . . . 5
eliminated many procedures to be performed only in AHP-fuzzy an1 an2 . . . ann
AHP solution and enabled coming to a conclusion in a shorter time.
The full AHP-fuzzy AHP solution is only practically usable if the At the last step, the mathematical process commences to nor-
number of criteria and alternatives is sufficiently low so that the malize and find the relative weights for each matrix. The relative
number of pairwise comparisons performed by evaluator must re- weights are given by the right eigenvector (w) corresponding to
main below a reasonable threshold. For example, if there are n cri- the largest eigenvalue (kmax), as
teria which have been assigned the importance weights and m
alternatives, then to run a full AHP-Fuzzy AHP solution there are Aw ¼ kmax w: ð2Þ
n  m  (m  1)/2 pairwise comparisons remaining to be performed If the pairwise comparisons are completely consistent, the ma-
for running a full AHP-fuzzy AHP solution (Shipley, Korvin, & Obit, trix A has rank 1 and kmax = n. In this case, weights can be obtained
1991; Shyur, 2006). Due to a large number of potential available by normalizing any of the rows or columns of A Wang and Yang
alternative in the literature, a full AHP-fuzzy AHP decision process (2007).
becomes impractical in some cases. Therefore, to avoid an unrea- It should be noted that the quality of the output of the AHP is
sonably large number of pairwise comparisons, the fuzzy TOPSIS strictly related to the consistency of the pairwise comparison judg-
is employed to achieve the final ranking results. ments. The consistency is defined by the relation between the en-
The remainder of this study is structured as follows: Section 2 tries of A: aij  ajk = aik. The consistency index (CI) is
briefly describes the proposed methods. In Section 3, proposed
model for weapon selection is presented and the stages of the pro-
CI ¼ ðkmax  nÞ=ðn  1Þ: ð3Þ
posed approach are explained in detail. How the proposed model is The final consistency ratio (CR), usage of which let someone to
used on a real world example is explained in Section 4. In Section 5, conclude whether the evaluations are sufficiently consistent, is cal-
conclusions and suggestions are discussed. culated as the ratio of the CI and the random index (RI), as
indicated.

2. Methods CR ¼ CI=RI: ð4Þ

The number 0.1 is the accepted upper limit for CR. If the final
2.1. The AHP method consistency ratio exceeds this value, the evaluation procedure
has to be repeated to improve consistency. The measurement of
AHP, developed by Saaty (1980), addresses how to determine consistency can be used to evaluate the consistency of decision-
the relative importance of a set of activities in a multi-criteria deci- makers as well as the consistency of overall hierarchy (Wang &
sion problem. The process makes it possible to incorporate judg- Yang, 2007).
ments on intangible qualitative criteria alongside tangible
quantitative criteria (Badri, 2001). The AHP method is based on
three principles: first, structure of the model; second, comparative
judgment of the alternatives and the criteria; third, synthesis of the Table 1
priorities. In the literature, AHP, has been widely used in solving Nine-point intensity of importance scale and its description.
many complicated decision-making problems (Chan, Kumar, Ti- Definition Intensity of importance
wari, Lau, & Choy, 2007; Dağdeviren & Yüksel, 2008; Kahraman,
Equally important 1
Ruan, & Doğan, 2003; Kulak & Kahraman, 2005). Moderately more important 3
In the first step, a complex decision problem is structured as a Strongly more important 5
hierarchy. AHP initially breaks down a complex multi-criteria deci- Very strongly more important 7
Extremely more important 9
sion-making problem into a hierarchy of interrelated decision ele-
Intermediate values 2, 4, 6, 8
ments (criteria, decision alternatives). With the AHP, the
 M. Dağdeviren et al. / Expert Systems with Applications 36 (2009) 8143–8151 8145

2.2. The TOPSIS method where I0 is associated with the benefit criteria, and I00 is
associated with the cost criteria.
The TOPSIS (technique for order performance by similarity to Step 5: Calculate the separation measures, using the n-dimen-
idea solution) was first developed by Hwang & Yoon (1981). sional Euclidean distance. The separation of each alterna-
According to this technique, the best alternative would be the tive from the positive-ideal solution ðDj Þ is given as
one that is nearest to the positive-ideal solution and farthest from vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u n
the negative ideal solution (Ertugrul & Karakasoglu, 2007). The po- uX

sitive-ideal solution is a solution that maximizes the benefit crite- Dj ¼ t

v ij  v i 2 j ¼ 1; 2; . . . ; J: ð10Þ
i¼1
ria and minimizes the cost criteria, whereas the negative ideal
solution maximizes the cost criteria and minimizes the benefit cri- Similarly, the separation of each alternative from the neg-
teria (Wang & Elhag, 2006). In short, the positive-ideal solution is ative-ideal solution ðDj Þ is as follows:
composed of all best values attainable from the criteria, whereas vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u n
the negative ideal solution consists of all worst values attainable uX

from the criteria (Wang, 2007). There have been lots of studies in Dj ¼t v ij  v i 2 j ¼ 1; 2; . . . ; J: ð11Þ
i¼1
the literature using TOPSIS for the solution of MCDM problems.
(Chen, 2000; Chu, 2002; Chu & Lin, 2002; Lai, Liu, & Hwang,
1994; Wang, Liu, & Zhang, 2005). The TOPSIS method consists of Step 6: Calculate the relative closeness to the idea solution and
the following steps (Shyur & Shih, 2006): rank the performance order. The relative closeness of
the alternative Aj can be expressed as
Step 1: Establish a decision matrix for the ranking. The structure
of the matrix can be expressed as follows: Dj
CC j ¼ ; j ¼ 1; 2; . . . ; J; ð12Þ
Dj þ Dj
2 F1 F2 . . . Fj . . . Fn 3
A1 f11 f12    f1j    f1n where the CC j index value lies between 0 and 1. The larger
6 7 the index value means the better the performance of the
A2 6 f21 f22
6    f2j    f2n 7
7
6 7 alternatives.
... 6 .. ... .
. . 7
.
6 .
6  .  . 7 7 ð5Þ
D¼ 6 7 2.3. The fuzzy TOPSIS method
Ai 6 fi1 fi2
6    fij    fin 7
7
6 7
.. 6 .. .. .. .. 7 The TOPSIS is widely used for tackling ranking problems in real
. 6 .
4 .  .  . 7 5 situations. Despite its popularity and simplicity in concept, this
AJ fJ1 fJ2    fJj    fJn method is often criticized for its inability to adequately handle
the inherent uncertainty and imprecision associated with the
mapping of the decision-maker’s perception to crisp values. In
where Aj denotes the alternatives j, j = 1,2, . . . ,J; Fi repre- the traditional formulation of the TOPSIS, personal judgments
sents ith attribute or criterion, i = 1, 2, . . . , n, related to are represented with crisp values. However, in many practical
ith alternative; and fij is a crisp value indicating the perfor- cases the human preference model is uncertain and decision-mak-
mance rating of each alternative Ai with respect to each ers might be reluctant or unable to assign crisp values to the com-
criterion Fj. parison judgments (Chan & Kumar, 2007). Having to use crisp
Step 2: Calculate the normalized decision matrix R(=[rij]). The values is one of the problematic points in the crisp evaluation pro-
normalized value rij is calculated as: cess. One reason is that decision-makers usually feel more confi-
dent to give interval judgments rather than expressing their
fij judgments in the form of single numeric values. As some criteria
rij ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Pn 2 ¼ 1; 2; . . . ; J; i ¼ 1; 2; . . . ; n: ð6Þ
are difficult to measure by crisp values, they are usually neglected
j¼1 fij
during the evaluation. Another reason is mathematical models
that are based on crisp value. These methods cannot deal with
Step 3: Calculate the weighted normalized decision matrix by decision-makers’ ambiguities, uncertainties and vagueness
multiplying the normalized decision matrix by its associ- which cannot be handled by crisp values. The use of fuzzy set the-
ated weights. The weighted normalized value vij is calcu- ory (Zadeh, 1965) allows the decision-makers to incorporate
lated as: unquantifiable information, incomplete information, non-obtain-
able information and partially ignorant facts into decision model
V ij ¼ wi  r ij ; j ¼ 1; 2; . . . ; J; i ¼ 1; 2; . . . ; n; ð7Þ (Kulak, Durmusoglu, & Kahraman, 2005). As a result, fuzzy TOPSIS
and its extensions are developed to solve ranking and justification
where wi represents the weight of the ith attribute or
problems (Büyüközkan, Feyzioğlu, & Nebol, 2007; Chen & Tsao,
criterion.
2007; Kahraman et al., 2007; Önüt & Soner, 2007; Wang & Elhag,
Step 4: Determine the positive-ideal and negative-ideal 2006; Yang & Hung, 2007; Yong, 2006).
solutions. This study uses triangular fuzzy number for fuzzy TOPSIS. The
   reason for using a triangular fuzzy number is that it is intuitively
A ¼ v 1 ; v 2 ; . . . ; v i easy for the decision-makers to use and calculate. In addition,
      modeling using triangular fuzzy numbers has proven to be an
  ð8Þ
¼ max v ij i 2 I0 ; min v ij i 2 I00 ; effective way for formulating decision problems where the infor-
j j
mation available is subjective and imprecise (Chang & Yeh, 2002;
  Chang, Chung, & Wang, 2007; Kahraman, Besßkese, & Ruan, 2004;
A ¼ v 1 ; v 2 ; . . . ; v i
      Zimmerman, 1996). In practical applications, the triangular form
  ð9Þ
¼ min v ij i 2 I0 ; max v ij j 2 I00 ; of the membership function is used most often for representing
j j
fuzzy numbers (Ding & Liang, 2005; Kahraman et al., 2004; Karsak
8146 M. Dağdeviren et al. / Expert Systems with Applications 36 (2009) 8143–8151

& Tolga, 2001; Xu & Chen, 2007). In the following, some basic where
important definitions of fuzzy sets are given (Chen, 1996; Chen,
Lin, & Huang, 2006; Cheng & Lin, 2002; Önüt & Soner, 2007; Raj  v
~ ij ¼ ~xij  wi .
& Kumar, 1999; Wang & Chang, 2007; Zimmerman, 1996).  A set of performance ratings of Aj (j = 1, 2, . . . , J) with respect to
~ ¼ f~
criteria Ci (i = 1, 2, . . . , n) called X xij ; i ¼ 1; 2; . . . ; n; J ¼
Definition 1. A fuzzy set A ~ in a universe of discourse X is
1; 2; . . . ; Jg.
characterized by a membership function lA~ ðxÞ which associates
 A set of importance weights of each criterion wi (i = 1, 2, . . . ,n).
with each element x in X a real number in the interval [0, 1]. The
~
function value lA~ ðxÞ is termed the grade of membership of x in A.
According to briefly summarized fuzzy theory above, fuzzy
~ can be defined by a trip- TOPSIS steps can be outlined as follows (Önüt & Soner, 2007):
Definition 2. A triangular fuzzy number a
let (a1, a2, a3) shown in Fig. 1. The membership function la~ ðxÞ is
Step 1: Choose the linguistic values ð~ xij ; i ¼ 1; 2; . . . ; n; J ¼
defined.
8 1; 2; . . . ; JÞ for alternatives with respect to criteria. The fuzzy lin-
>
> 0; x < a1 guistic rating ð~ xij Þ preserves the property that the ranges of nor-
>
>
< xa1 ; a1 6 x 6 a2 malized triangular fuzzy numbers belong to [0, 1]; thus, there is
a2 a1
la~ ðxÞ ¼ xa3
ð13Þ no need for normalization.
>
> ; a2 6 x 6 a3
>
> a2 a3
Step 2: Calculate the weighted normalized fuzzy decision
:
0; x > a3 matrix. The weighted normalized value v ~ ij calculated by Eq. (20).
~ and b
Let a ~ be two triangular fuzzy numbers parameterized by Step 3: Identify positive-ideal (A*) and negative ideal (A) solu-
the triplet (a1, a2, a3) and (b1, b2, b3), respectively, then the oper- tions. The fuzzy positive-ideal solution (FPIS,A*) and the fuzzy
ational laws of these two triangular fuzzy numbers are as follows: negative-ideal solution (FNIS, A) are shown in the following
equations:
~ ¼ ða1 ; a2 ; a3 ÞðþÞðb1 ; b2 ; b3 Þ ¼ ða1 þ b1 ; a2 þ b2 ; a3 þ b3 Þ; ð14Þ
~ðþÞb
a  
 
~ ¼ ða ; a ; a ÞðÞðb ; b ; b Þ ¼ ða  b ; a  b ; a  b Þ; ð15Þ
~ðÞb
a A ¼ v~ 1 ; v~ 2 ; . . . ; v~ i ¼ max v ij ji 2 I0 ;
1 2 3 1 2 3 1 3 2 2 3 1 j
 
a ~ ¼ ða þ 1; a2 ; a3 ÞðÞðb1 ; b2 ; b3 Þ ¼ ða1  b1 ; a2  b2 ; a3  b3 Þ; ð16Þ
~ðÞb 00
 min v ij ji 2 I ; i ¼ 1; 2; . . . ; n j ¼ 1; 2; . . . ; J; ð21Þ
~ ¼ ða1 ; a2 ; a3 Þð=Þðb1 ; b2 ; b3 Þ ¼ ða1 =b3 ; a2 =b2 ; a3 =b1 Þ; j
~ð=Þb
a ð17Þ
~ ¼ ðka1 ; ka2 ; ka3 Þ:
a ð18Þ  
A ¼ f v~ 1 ; v~ 2 ; . . . ; v~ i g ¼ min v ij ji 2 I0 ;
j
 
Definition 3. A linguistic variable is a variable values of which are  max v ij ji 2 I00 ; i ¼ 1; 2; . . . ; n j ¼ 1; 2; . . . ; J ð22Þ
linguistic terms (Chen, 2000; Zadeh, 1975). The concept of j

linguistic variable is very useful in dealing with situations which where I0 is associated with benefit criteria and I00 is associated
are too complex or too ill-defined to be reasonably described in with cost criteria.
conventional quantitative expressions (Chen, 2000; Zadeh, 1975). Step 4: Calculate the distance of each alternative from A* and A
For example, ‘‘weight” is a linguistic variable; its values are very using the following equations:
low, low, medium, high, very high, etc. These linguistic values can
also be represented by fuzzy numbers. X
n
Dj ¼ dðv~ ij ; v~ i Þ j ¼ 1; 2; . . . ; J ð23Þ
j¼1
~ ¼ (b , b , b ) be two triangu-
~ ¼ (a1, a2, a3) and b
Definition 4. Let a 1 2 3
lar fuzzy numbers, then the vertex method is defined to calculate X
n
the distance between them. Dj ¼ dðv~ ij ; v~ i Þ j ¼ 1; 2; . . . ; J: ð24Þ
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi j¼1
~ 1
~; bÞ ¼
dða ½ða1  b1 Þ2 þ ða2  b2 Þ2 þ ða3  b3 Þ2 : ð19Þ: Step 5: Calculate similarities to ideal solution.
3
Dj
Definition 5. Considering the different importance values of each CC j ¼ j ¼ 1; 2; . . . J: ð25Þ
Dj þ Dj
criterion, the weighted normalized fuzzy-decision matrix is con-
structed as. Step 6: Rank preference order. Choose an alternative with max-
imum CCj or rank alternatives according to CCj in descending
~ ¼ ½v~ ij  ;
V i ¼ 1; 2; . . . ; n; j ¼ 1; 2; . . . ; J; ð20Þ
nJ order.

µ a~ ( x ) 3. The proposed model

The proposed model for the weapon selection problem, com-

1
posed of AHP and fuzzy TOPSIS methods, consists of three basic
stages: (1) identify the criteria to be used in the model, (2) AHP
computations, (3) evaluation of alternatives with fuzzy TOPSIS
and determination of the final rank.
In the first stage, alternative weapons and the criteria which
will be used in their evaluation are determined and the decision
x hierarchy is formed. AHP model is structured such that the objec-
a1 a2 a3 tive is in the first level, criteria are in the second level and alterna-
tive weapons are on the third level. In the last step of the first stage,
~.
Fig. 1. Triangular fuzzy number a the decision hierarchy is approved by decision-making team.
 M. Dağdeviren et al. / Expert Systems with Applications 36 (2009) 8143–8151 8147

After the approval of decision hierarchy, criteria used in weapon

Forming decision making team
selection are assigned weights using AHP in the second stage. In
this phase, pairwise comparison matrices are formed to determine
the criteria weights. The experts from decision-making team make
Determining alternative weapons
individual evaluations using the scale provided in Table 1, to deter-
mine the values of the elements of pairwise comparison matrices.
Computing the geometric mean of the values obtained from indi-
Determining the criteria to be used in
vidual evaluations, a final pairwise comparison matrix on which evaluation
there is a consensus is found. The weights of the criteria are calcu-
lated based on this final comparison matrix. In the last step of this
phase, calculated weights of the criteria are approved by decision- STAGE 1:
Structuring decision hierarchy Group working
making team.
Weapon ranks are determined by using fuzzy TOPSIS method in
the third stage. Linguistic values are used for evaluation of
alternative weapons in this step. The membership functions of
these linguistic values are shown at (Fig. 2), and the triangular Approve
N decision
fuzzy numbers related with these variables are shown at
hierarchy?
(Table 2).
The weapon having the maximum CCj value is determined as
the optimal weapon according to the calculations by Fuzzy TOPSIS. Y
Ranking of the other weapons is determined according to CCj in
descending order. Schematic diagram of the proposed model for
Assigning criteria weigths via AHP
weapon selection is provided in Fig. 3.

4. A Numeric application of proposed model STAGE 2:

AHP
Proposed model is applied to a real problem in Defence Indus- Approve
try. The aim of our specific study is to assess possible alternative N criteria
weights?
weapon solutions and to help the decision-makers accordingly in
terms of user requirements.
The high technology weapons make significant improvements
Y
in the defense capabilities of the nations. Therefore, selecting the
most proper weapon is of great importance for the nations. But it
is hard to choose the most suitable one among the alternatives Evaluation of alternatives
which dominate each other in different characteristics. For the
application, an expert team was formed from three junior manag-
ers of the Turkish Defence Industry Firms and the authors of this STAGE 3:
Determine the final rank Fuzzy TOPSIS
paper. The criteria to be used in the model were determined by
the expert team. Pairwise comparison matrices used to calculate
criteria weights were also formed by the same team. The applica-
Selecting the optimal weapon

Fig. 3. Schematic diagram of the proposed model for weapon selection.

VL L M H VH E
1.0
tion performed is based on the steps provided in previous section
and explained step by step together with the results.

4.1. Identification of necessary criteria for weapon selection

Criteria to be considered in the selection of infantry rifle are

0 0.2 0.4 0.6 0.8 1 determined by the expert team. Past experience and the back-
ground of the expert team are utilized in the determination of
Fig. 2. Membership functions of linguistic values for criteria rating.
the criteria and 13 important criteria to be used for infantry rifle
selection are established. Criteria and their definitions of impor-
tance are given in Table 3.
Table 2 Following the determination of the criteria, alternative rifles un-
Linguistic values and fuzzy numbers. der development or in usage are investigated and decision-making
Linguistic values Fuzzy numbers team determined five possible weapons suitable for the needs. At
Very low (VL) (0, 0, 0.2)
the end of this pre-study, it turned out to be that these five alter-
Low (L) (0, 0.2, 0.4) native rifles do not have different characteristics with regard to
Medium (M) (0.2, 0.4, 0.6) all 13 criteria. All 5 rifles seemed to satisfy the requirements of
High (H) (0.4, 0.6, 0.8) the 7 criteria presented in Table 3; remaining 6 criteria seemed
Very high (VH) (0.6, 0.8, 1)
to differentiate for the alternative rifles. These 6 criteria are as
Excellent (E) (0.8, 1, 1)
follows: Price (C1), Weight (C2), Barrel (C3), Modularity (C4),
8148 M. Dağdeviren et al. / Expert Systems with Applications 36 (2009) 8143–8151

Table 3
Weapon evaluation criteria and its definition.

Criterion Definition of Importance

Barrel Manufacturing process, Application of thermal processes, chrome coating
Recoil system Normal buffer, tampon, cushioning, wolfram cushioning, usage of gas piston and evacuation channel
Hit precision Dispersion in a certain distance for hit accuracy
Weight The overall weight for soldier’s ease of use
Usability in different climatic conditions and Climatic conditions (gas piston, gas tube, lock mechanism), maintenance (gas piston, field strip and assembly,
maintenance chrome barrel)
Modularity Modular Frame (without Piccatiny rail, welded rail, built of rail) and Modular Handguard (without rail, partial rail,
complete rail)
Trigger system Adjustable trigger, both hands usability
Safety mechanism Needle safety, safety lock
Sight system Portability, adjustable slewing mechanism
Magazine Being composite, transparency, multiple usage
Butt Being composite, graded adjustability, folding, cheek cushioning,
Dust bolt or dust cover Presence or not
Price First time purchase cost

Usability in different climatic conditions and maintenance C(5), re- Table 5

coil (C6). As a result, only these six criteria were used in evaluation Results obtained with AHP.
and decision hierarchy is established accordingly. Decision hierar- Criteria Weights (w) kmax, CI, RI CR
chy structured with the determined alternative weapons and crite-
C1 0.076 kmax = 6.342
ria is provided in Fig. 4. C2 0.296
There are three levels in the decision hierarchy structured C3 0.092 CI = 0.068 0.055
for weapon selection problem. The overall goal of the decision C4 0.251
C5 0.155 RI = 1.24
process determined as ‘‘the selection of the optimal weapon” is
C6 0.130
in the first level of the hierarchy. The criteria are on the second
level and alternative weapons are on the third level of the
hierarchy.
The results obtained from the computations based on the
pairwise comparison matrix provided in Table 4, are presented in
4.2. Calculate the weights of criteria
Table 5.
The C2, C4 and C5 are determined as the three most important
After forming the decision hierarchy for the problem, the
criteria in the weapon selection process by AHP. Consistency ratio
weights of the criteria to be used in evaluation process are calcu-
of the pairwise comparison matrix is calculated as 0.055 < 0.1. So
lated by using AHP method. In this phase, the experts in the expert
the weights are shown to be consistent and they are used in the
team are given the task of forming individual pairwise comparison
selection process.
matrix by using the scale given in Table 1. Geometric means of
these values are found to obtain the pairwise comparison matrix
4.3. Evaluation of alternatives and determine the final rank
on which there is a consensus (Table 4).
At this stage of the decision procedure, the team members were
asked to establish the decision matrix by comparing alternatives
under each of the criteria separately. Fuzzy Evaluation Matrix
Selection of the best established by the evaluation of alternative rifles by linguistic vari-
weapon ables in Table 2, is presented in Table 6. Linguistic variables are in
the upper section of Table 6, the lower section is composed of the
triangular fuzzy numbers which are equivalent of Linguistic
(C1) (C2) (C3) (C4) (C5) (C6)
variables.
After the fuzzy evaluation matrix was determined, the second
step is to obtain a fuzzy weighted decision table. Using the criteria
weights calculated by AHP (Table 5) in this step, the Weighted
A1 A2 A3 A4 A5 Evaluation Matrix is established with Eq. (20). The resulting fuzzy
weighted decision matrix is shown in Table 7.
Fig. 4. The decision hierarchy of weapon selection. According to (Table 7), it is seen that the elements v ~ ij , 8i; j are
normalized positive triangular fuzzy numbers and their ranges be-
long to the closed interval [0, 1]. Thus, we can define the fuzzy po-
sitive-ideal solution (FPIS, A*) and the fuzzy negative-ideal solution
Table 4 (FNIS, A) as v ~ i ¼ ð1; 1; 1Þ and v
~i ¼ ð0; 0; 0Þ for benefit criterion,
The pairwise comparison matrix for criteria. and v i ¼ ð0; 0; 0Þ and v i ¼ ð1; 1; 1Þ for cost criterion. In this prob-
~  ~ 

C1 C2 C3 C4 C5 C6 lem, C1 and C2 are cost criteria whereas the other criteria are ben-
C1 1.0 0.4 0.4 0.3 0.4 0.5
efit criteria. For the third step, the distance of each alternative from
C2 2.1 1.0 3.4 1.9 2.3 2.4 D* and D can be currently calculated using Eq. (23) and Eq. (24).
C3 2.3 0.2 1.0 0.3 0.4 0.4 The fourth step solves the similarities to an ideal solution by Eq.
C4 3.2 0.5 3.2 1.0 2.8 1.9 (25) (Yang & Hung, 2007).
C5 2.4 0.4 2.1 0.3 1.0 1.8
In order to illustrate steps 3 and 4 calculation, CC1 calculation is
C6 1.9 0.4 2.3 0.5 0.5 1.0
used as an example as follows:
 M. Dağdeviren et al. / Expert Systems with Applications 36 (2009) 8143–8151 8149

Table 6
Fuzzy evaluation matrix for the alternative weapons.

C1 C2 C3 C4 C5 C6
A1 Low Excellent Very high Medium High Medium
A2 Medium High High Very high Very high High
A3 Medium Very high Excellent High Medium High
A4 Very high Low Medium Excellent High Medium
A5 High Very high Very high Low Excellent Very high

A1 (0, 0.2, 0.4) (0.8, 1, 1) (0.6, 0.8, 1) (0.2, 0.4, 0.6) (0.4, 0.6, 0.8) (0.2, 0.4, 0.6)
A2 (0.2, 0.4, 0.6) (0.4, 0.6, 0.8) (0.4, 0.6, 0.8) (0.6, 0.8, 1) (0.6, 0.8, 1) (0.4, 0.6, 0.8)
A3 (0.2, 0.4, 0.6) (0.6, 0.8, 1) (0.8, 1, 1) (0.4, 0.6, 0.8) (0.2, 0.4, 0.6) (0.4, 0.6, 0.8)
A4 (0.6, 0.8, 1) (0, 0.2, 0.4) (0.2, 0.4, 0.6) (0.8, 1, 1) (0.4, 0.6, 0.8) (0.2, 0.4, 0.6)
A5 (0.4, 0.6, 0.8) (0.6, 0.8, 1) (0.6, 0.8, 1) (0, 0.2, 0.4) (0.8, 1, 1) (0.6, 0.8, 1)

Weight 0.076 0.296 0.092 0.251 0.155 0.130

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1h i Table 8
D1 ¼ ð0  0Þ2 þ ð0  0:015Þ2 þ ð0  0:030Þ2 Fuzzy TOPSIS results.
3
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1h i Alternatives Dj D
j CCj
þ ð0  0:238Þ2 þ ð0  0:297Þ2 þ ð0  0:297Þ2 A1 3.981 2.044 0.339
3
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi A2 3.761 2.262 0.376
1h i
A3 3.900 2.124 0.353
þ ð1  0:055Þ2 þ ð1  0:074Þ2 þ ð1  0:092Þ2
3 A4 3.724 2.309 0.383
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi A5 3.919 2.110 0.350
1h i
þ ð1  0:050Þ2 þ ð1  0:100Þ2 þ ð1  0:151Þ2
3
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1h i
þ ð1  0:062Þ2 þ ð1  0:093Þ2 þ ð1  0:124Þ2 Table 9
3 Weighted and unweighted rankings.
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1h i
þ ð1  0:026Þ2 þ ð1  0:052Þ2 þ ð1  0:078Þ2 Rank Weighted Weighted Unweighted Unweighted
3 CCj ranking CCj ranking
¼ 3:981; 1 0.383 A4 0.618 A2
2 0.376 A2 0.551 A3
3 0.353 A3 0.550 A4
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 4 0.350 A5 0.549 A5
1 5 0.339 A1 0.508 A1
D1 ¼þ ½ð1  0Þ2 þ ð1  0:015Þ2 þ ð1  0:030Þ2 
3
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
þ ½ð1  0:238Þ2 þ ð1  0:297Þ2 þ ð1  0:297Þ2 
3 Similar calculations are done for the other alternatives and the
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1 results of fuzzy TOPSIS analyses are summarized in Table 8. Based
þ ½ð0  0:055Þ2 þ ð0  0:074Þ2 þ ð0  0:092Þ2  on CCj values, the ranking of the alternatives in descending order
3
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi are A4, A2, A3, A5 and A1. Proposed model results indicate that
1
þ ½ð0  0:050Þ2 þ ð0  0:100Þ2 þ ð0  0:151Þ2  A4 is the best alternative with CC value of 0.383.
3 The case in which criteria weights are not considered, i.e. the
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1 criteria have equal priorities, is analyzed and the CCj values ob-
þ ½ð0  0:062Þ2 þ ð0  0:093Þ2 þ ð0  0:124Þ2 
3 tained in this condition are presented in Table 9 with their compar-
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1 isons with previous values.
þ ½ð0  0:026Þ2 þ ð0  0:052Þ2 þ ð0  0:078Þ2  Based on unweighted CCj values, the ranking of the alternatives
3
in descending order are A2, A3, A4, A5 and A1. The best alternative
¼ 2:044;
has changed according to the unweighted ranking results. The
change in the best alternative when criteria weights are taken into
Dj 2:044
CCj ¼ ¼ ¼ 0:339: account has shown that criteria weights found consistently consti-
Dj þ Dj 3:981 þ 2:044 tute an important phase in decision-making process.

Table 7
Weighted evaluation for the alternative weapons.

C1 C2 C3 C4 C5 C6
A1 (0.000, 0.015, 0.030) (0.238, 0.297, 0.297) (0.055, 0.074, 0.092) (0.050, 0.100, 0.151) (0.062, 0.093, 0.124) (0.026, 0.052, 0.078)
A2 (0.015, 0.030, 0.046) (0.119, 0.178, 0.238) (0.037, 0.055, 0.074) (0.151, 0.201, 0.251) (0.093, 0.124, 0.155) (0.052, 0.078, 0.104)
A3 (0.015, 0.030, 0.046) (0.178, 0.238, 0.297) (0.074, 0.092, 0.092) (0.100, 0.151, 0.201) (0.031, 0.062, 0.093) (0.052, 0.078, 0.104)
A4 (0.046, 0.061, 0.076) (0.000, 0.059, 0.119) (0.018, 0.037, 0.055) (0.201, 0.251, 0.251) (0.062, 0.093, 0.124) (0.026, 0.052, 0.078)
A5 (0.030, 0.046, 0.061) (0.178, 0.238, 0.297) (0.055, 0.074, 0.092) (0.000, 0.050, 0.100) (0.124, 0.155, 0.155) (0.078, 0.104, 0.130)

A* v~ 1 ¼ ð0; 0; 0Þ v~ 2 ¼ ð0; 0; 0Þ v~ 3 ¼ ð1; 1; 1Þ v~ 4 ¼ ð1; 1; 1Þ v~ 5 ¼ ð1; 1; 1Þ v~ 6 ¼ ð1; 1; 1Þ

A v~ 1 ¼ ð1; 1; 1Þ v~ 2 ¼ ð1; 1; 1Þ v~ 3 ¼ ð0; 0; 0Þ v~ 4 ¼ ð0; 0; 0Þ v~ 5 ¼ ð0; 0; 0Þ v~ 6 ¼ ð0; 0; 0Þ
8150 M. Dağdeviren et al. / Expert Systems with Applications 36 (2009) 8143–8151

5. Conclusion and suggestions Büyüközkan, G., Feyzioğlu, O., & Nebol, E. (2007). Selection of the strategic alliance
partner in logistics value chain. International Journal of Production Economics.
doi:10.1016/j.ijpe.2007.01.016.
The weapon selection problem is a strategic issue and has sig- Chan, F. T. S., & Kumar, N. (2007). Global supplier development considering
nificant impacts to the efficiency of a defense system. Several alter- risk factors using fuzzy extended AHP-based approach. OMEGA, 35,
417–431.
natives must be considered and evaluated in terms of many
Chan, F. T. S., Kumar, N., Tiwari, M. K., Lau, H. C. W., & Choy, K. L. (2007). Global
different conflicting criteria in a weapon selection problem, leading supplier selection: a fuzzy-AHP approach. International Journal of Production
to a large set of subjective or ambiguous data. Therefore, an effec- Research. doi:10.1080/00207540600787200.
Chang, Y. H., & Yeh, C. H. (2002). A survey analysis of service quality for domestic
tive evaluation approach is essential to improve decision quality.
airlines. European Journal of Operational Research, 139, 166–177.
This study, presenting a scientific framework to assess weapon sys- Chang, Y. H., Chung, H. Y., & Wang, S. Y. (2007). A survey and optimization-based
tems, uses triangular fuzzy numbers to express linguistic values evaluation of development strategies for the air cargo industry. International
that consider the subjective judgments of evaluators and then Journal of Production Economics, 106, 550–562.
Chen, C. T. (2000). Extensions of the TOPSIS for group decision-making under fuzzy
adopts fuzzy multiple criteria decision-making approach to syn- environment. Fuzzy Sets and Systems, 114, 1–9.
thesize the group decision. Chen, S. M. (1996). Evaluating weapon systems using fuzzy arithmetic operations.
A decision model is provided for weapon selection problem in Fuzzy Sets and Systems, 77, 265–276.
Chen, C. T., Lin, C. T., & Huang, S. F. (2006). A fuzzy approach for supplier evaluation
this study. The proposed model is based on the comparisons of and selection in supply chain management. International Journal of Production
weapon alternatives according to identified criteria. AHP and fuzzy Economics, 102, 289–301.
TOPSIS compound decision-making method has been used in pro- Chen, T. Y., & Tsao, C. Y. (2007). The interval-valued fuzzy TOPSIS methods
and experimental analysis. Fuzzy Sets and Systems. doi:10.1016/j.fss.2007.
posed model. With its above-mentioned structure, the proposed 11.004.
model differs from the present weapon selection literature. AHP Cheng, C. H., & Mon, M. L. (1994). Evaluating weapon system by analytic hierarchy
is used to assign weights to the criteria to be used in weapon selec- process based on fuzzy scales. Fuzzy Sets and Systems, 63, 1–10.
Cheng, C. H. (1996). Evaluating naval tactical missile systems by fuzzy AHP based on
tion, while fuzzy TOPSIS is employed to determine the priorities of
the grade value of membership function. European Journal of Operational
the alternatives. The weights obtained from AHP are included in Research, 96, 343–350.
decision-making process by using them in fuzzy TOPSIS computa- Cheng, C. H. (1999). Evaluating weapon systems using ranking fuzzy numbers. Fuzzy
Sets and Systems, 107, 25–35.
tions and the alternative priorities are determined based on these
Cheng, C. H., Yang, K. L., & Hwang, C. L. (1999). Evaluating attack helicopters by AHP
weights. In this way, weighting of the criteria considered during based on linguistic variable weights. European Journal of Operational Research,
decision-making and evaluation of these criteria are performed 116, 423–435.
simultaneously. Proposed model has significantly increased the Cheng, C. H., & Lin, Y. (2002). Evaluating the best main battle tank using fuzzy
decision theory with linguistic criteria evaluation. European Journal of
efficiency of decision-making process in weapon selection. Fuzzy Operational Research, 142, 174–186.
TOPSIS has eliminated many procedures to be performed only in Chu, T. C. (2002). Selecting plant location via a fuzzy TOPSIS approach. International
AHP-fuzzy AHP solution and enabled coming to a conclusion in a Journal of Advanced Manufacturing Technology, 20, 859–864.
Chu, T. C., & Lin, Y. C. (2002). Improved extensions of the TOPSIS for group decision-
shorter time. Additionally, in the application, it is shown that cal- making under fuzzy environment. Journal of Information and Optimization
culation of the criteria weights is important in TOPSIS method Sciences, 23, 273–286.
Dağdeviren, M., & Yüksel, I. _ (2008). Developing a fuzzy analytic hierarchy process
and they could change the ranking.
(AHP) model for behavior-based safety management. Information Science, 178,
Multiple decision-makers are often preferred rather than a sin- 1717–1733.
gle decision-maker to avoid the bias and minimize the partiality Dağdeviren, M. (2008). Decision making in equipment selection: An integrated
in the decision process (Bilsel, Büyüközkan, & Ruan, 2006). Since approach with AHP and PROMETHEE. Journal of Intelligent Manufacturing, 19,
397–406.
the decisions taken in the Defense Industries of the nations affect
Ding, J. F., & Liang, G. S. (2005). Using fuzzy MCDM to select partners of strategic
the overall community, it is better to come to a final decision alliances for liner shipping. Information Sciences, 173, 197–225.
with the participation of experts. The efficiency and the accuracy _ & Karakasoglu, N. (2007). Performance evaluation of Turkish cement
Ertugrul, I.,
firms with fuzzy analytic hierarchy process and TOPSIS methods. Expert Systems
of the decision could be enhanced by different participants from
with Applications, 36(1), 702–715.
different areas. In addition, mostly qualitative criteria have to Göleç, A., & Tasßkın, H. (2007). Novel methodologies and a comparative study for
be considered in the decision process and the difficulty in deter- manufacturing systems performance evaluations. Information Sciences, 177,
mining the parameters of these criteria force decision-makers to 5253–5274.
Hwang, C. L., & Yoon, K. (1981). Multiple attribute decision making: Methods and
utilize scientific methods. Therefore development of a group deci- applications, A State of the Art Survey. New York: Springer-Verlag.
sion-making system for national defense problems is very useful. Kahraman, C., Ruan, D., & Doğan, I. (2003). Fuzzy group decision-making for facility
This combines the ideas of different individuals by a scientific location selection. Information Sciences, 157, 135–153.
Kahraman, C., Besßkese, A., & Ruan, D. (2004). Measuring flexibility of computer
method. integrated manufacturing systems using fuzzy cash flow analysis. Information
Although the model was developed and tested for use in weap- Sciences, 168, 77–94.
on selection problem, it can also be used with slight modifications Kahraman, C., Büyüközkan, G., & Atesß, N. Y. (2007). A two phase multi-attribute
decision making approach for new product introduction. Information Sciences,
in other decision-making problems in Defense Industries of differ- 177, 1567–1582.
ent countries. Also, mathematical models can be combined with Karsak, E. E., & Tolga, E. (2001). Fuzzy multi-criteria decision-making procedure for
the proposed model. This will improve the proposed method and evaluating advanced manufacturing system investments. International Journal
of Production Economics, 69, 49–64.
is one of the directions in our future research. Kulak, O., & Kahraman, C. (2005). Fuzzy multi-attribute selection among
transportation companies using axiomatic design and analytic hierarchy
References process. Information Sciences, 170, 191–210.
Kulak, O., Durmusoglu, B., & Kahraman, C. (2005). Fuzzy multi-attribute equipment
Albayrak, E., & Erensal, Y. C. (2004). Using analytic hierarchy process (AHP) to selection based on information axiom. Journal of Materials Processing Technology,
improve human performance. An application of multiple criteria decision 169, 337–345.
making problem. Journal of Intelligent Manufacturing, 15, 491–503. Lai, Y. J., Liu, T. Y., & Hwang, C. L. (1994). TOPSIS for MODM. European Journal of
Ayağ, Z., & Özdemir, R. G. (2006). A fuzzy AHP approach to evaluating machine tool Operational Research, 76, 486–500.
alternatives. Journal of Intelligent Manufacturing, 17, 179–190. Mergias, I., Moustakas, K., Papadopoulos, A., & Loizidou, M. (2007). Multi-criteria
Badri, M. A. (2001). A combined AHP-GP model for quality control systems. decision aid approach for the selection of the best compromise management
International Journal of Production Economics, 72, 27–40. scheme for ELVs: The case of Cyprus. Journal of Hazardous Materials, 147,
Bilsel, R. U., Büyüközkan, G., & Ruan, D. (2006). A fuzzy preference-ranking model 706–717.
for a quality evaluation of hospital web sites. International Journal of Intelligent Munoz, A. A., & Sheng, P. (1995). An analytical approach for determining the
Systems, 21, 1181–1197. environmental impact of machining processes. Journal of Materials Processing
Bufardi, A., Gheorghe, A., Kiritsis, R., & Hirouchakis, D. (2004). Multicriteria decision- Technology, 53, 736–758.
aid approach for product end-of-life alternative selection. International Journal Nelson, C. A. (1986). A scoring model for flexible manufacturing system project
of Production Research, 42, 3139–3157. selection. European Journal of Operational Research, 24, 346–359.
 M. Dağdeviren et al. / Expert Systems with Applications 36 (2009) 8143–8151 8151

Mon, D. L., Cheng, C. H., & Lin, J. C. (1994). Evaluating weapon system using fuzzy Wang, J. J., & Yang, D. L. (2007). Using a hybrid multi-criteria decision aid method
analytic hierarchy process based on entropy weight. Fuzzy Sets and Systems, 62, for information systems outsourcing. Computers & Operation Research, 34,
127–134. 3691–3700.
Önüt, S., & Soner, S. (2007). Transshipment site selection using the AHP and TOPSIS Wang, X., & Triantaphyllou, E. (2005). Ranking irregularities when evaluating
approaches under fuzzy environment. Waste Management. doi:10.1016/ alternatives by using some ELECTRE methods. OMEGA. doi:10.1016/j.omega.
j.wasman.2007.05.019. 2005.12.003.
Raj, P. A., & Kumar, D. N. (1999). Ranking alternatives with fuzzy weights using Wang, Y. J. (2007). Applying FMCDM to evaluate financial performance of domestic
maximizing set and minimizing set. Fuzzy Sets and Systems, 105, 365–375. airlines in Taiwan. Expert Systems with Applications. doi:10.1016/j.eswa.2007.
Saaty, T. L. (1980). The analytic hierarchy process. New York: McGraw-Hill. 02.029.
Shipley, M. F., Korvin, D. K., & Obit, R. (1991). A decision making model for multi- Xu, Z. S., & Chen, J. (2007). An interactive method for fuzzy multiple attribute group
attribute problems incorporating uncertainty and bias measures. Computers and decision making. Information Sciences, 177, 248–263.
Operations Research, 18, 335–342. Yang, T., & Hung, C. C. (2007). Multiple-attribute decision making methods for plant
Shyur, H. J. (2006). COTS evaluation using modified TOPSIS and ANP. Applied layout design problem. Robotics and Computer-Integrated Manufacturing, 23,
Mathematics and Computation, 177, 251–259. 126–137.
Shyur, H. J., & Shih, H. S. (2006). A hybrid MCDM model for strategic vendor Yong, D. (2006). Plant location selection based on fuzzy TOPSIS. International Journal
selection. Mathematical and Computer Modelling, 44, 749–761. of Advanced Manufacturing Technology, 28, 839–844.
Vincke, P. (1992). Multicriteria decision aid. New York: Wiley. _ & Dağdeviren, M. (2007). Using the analytic network process (ANP) in a
Yuksel, I.,
Wang, J., Liu, S. Y., & Zhang, J. (2005). An extension of TOPSIS for fuzzy MCDM based SWOT analysis – A case study for a textile firm. Information Sciences, 177,
on vague set theory. Journal of Systems Science and Systems Engineering, 14, 3364–3382.
73–84. Zadeh, L. A. (1965). Fuzzy sets. Information and Control, 8, 338–353.
Wang, Y. M., & Elhag, T. M. S. (2006). Fuzzy TOPSIS method based on alpha level sets Zadeh, L. A. (1975). The concept of a linguistic variable and its application to
with an application to bridge risk assessment. Expert Systems with Applications, approximate reasoning. Information Sciences, 8. pp. 199–249(I), 301–357(II).
31, 309–319. Zeleny, M. (1982). Multiple criteria decision making. New York: McGraw-Hill.
Wang, T. C., & Chang, T. H. (2007). Application of TOPSIS in evaluating initial Zimmerman, H. J. (1996). Fuzzy sets theory and its applications. Boston: Kluwer
training aircraft under a fuzzy environment. Expert Systems with Applications, 33, Academic Publishers.
870–880.