Expert Systems with Applications 38 (2011) 96569664

Contents lists available at ScienceDirect

Expert Systems with Applications

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

Fuzzy AHP approach for supplier selection in a washing machine company

Ozcan Kilincci , Suzan Asl Onal
Dokuz Eylul University, Dept. of Industrial Engineering, Izmir, Turkey

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

Keywords: Supplier selection process has gained importance recently, since the cost of raw materials and component
Multi-criteria decision-making parts constitutes the main cost of a product and most of the rms have to spend considerable amount of
Supplier selection their revenues on purchasing. Supplier selection is one of the most important decision making problems
Analytic hierarchy process including both qualitative and quantitative factors to identify suppliers with the highest potential for
Fuzzy analytic hierarchy process
meeting a rms needs consistently and at an acceptable cost. In this study, supplier selection problem
of a well-known washing machine company in Turkey is investigated and a fuzzy analytic hierarchy pro-
cess based methodology is used to select the best supplier rm providing the most customer satisfaction
for the criteria determined. After the main attributes and sub-attributes for supplier selection are dened
to design the hierarchy structure, the weights of them and alternatives are calculated using fuzzy analytic
hierarchy process approach. The supplier with the highest priority weight is selected as the best supplier.
Macros in MS Excel are written to calculate the priority weights of the alternatives based on the question-
naire forms used to facilitate comparisons of main attributes, sub-attributes and alternatives.
2011 Elsevier Ltd. All rights reserved.

1. Introduction
Increases and varieties of customer demands, advances of re-
cent technologies in communication and information systems,
competition in global environment, decreases in governmental
regulations and increases in environmental consciousness have
forced companies for focusing on supply chain management
(Tracey & Tan, 2001). The supply chain management term de-
nes the integration of activities to procure materials, transforms
them into intermediate goods and nal products, and delivers to
customers. The supply chain consists of all links from suppliers
to customers of a product. Gofn, Szwejczewski, and New (1997)
have stated that supplier management is one of the key issues of
supply chain management because the cost of raw materials and
component parts constitutes the main cost of a product and most
of the rms have to spend considerable amount of their sales rev-
enues on purchasing. In most industries the cost of raw materials
and component parts constitutes the main cost of a product, such
that in some cases it can account for up to 70% (Ghobadian, Stainer,
& Kiss, 1993). In high technology rms, purchased materials and
services represent up to 80% of total product cost (Weber, Current,
& Benton, 1991). Thus the purchasing department can play a key
role in an organizations efciency and effectiveness since the
department has a direct effect on cost reduction, protability and
exibility of a company by selecting the right suppliers signi-
Corresponding author.
E-mail address: ozcan.kilincci@deu.edu.tr (O. Kilincci).
cantly reduces the purchasing costs and improves corporate com-
petitiveness (Ghodsypour & OBrien, 2001).
Supplier selection is a multi-criteria problem which includes
both qualitative and quantitative factors. In order to select the best
suppliers it is necessary to make a trade off between these tangible
and intangible factors some of which may conict (Ghodsypour &
OBrien, 1998). The objective of supplier selection is to identify
suppliers with the highest potential for meeting a rms needs con-
sistently and at an acceptable cost. Selection is a broad comparison
of suppliers using a common set of criteria and measures. How-
ever, the level of detail used for examining potential suppliers
may vary depending on a rms needs.
Although analytical hierarchy process (AHP) is widely preferred
for solving multi-criteria decision-making problems in real situa-
tions, it is insufcient to explain uncertain conditions in especially
pair-wise comparison stage. Most of humans judgments are not
represented as exact numbers. Since some of the evaluation crite-
ria are subjective and qualitative in nature, it is very difcult for
the decision-maker to express the preferences using exact numer-
ical values and to provide exact pair-wise comparison judgments
(Chan & Kumar, 2007). Fuzzy evaluations in decision making pro-
cess are very useful to tackle this disadvantage of AHP.
To select the best supplier for a washing machine company, this
study proposes a fuzzy extended AHP (FEAHP) approach using tri-
angular fuzzy numbers to represent decision makers comparison
judgments and extent analysis method to decide the nal priority
of different decision criteria. The presented fuzzy AHP uses the lin-
guistic variables and triangular fuzzy numbers as a pair-wise com-
parison scale for deriving the priorities of different selection

0957-4174/$ - see front matter 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.eswa.2011.01.159
 O. Kilincci, S.A. Onal / Expert Systems with Applications 38 (2011) 96569664
attributes and sub-attributes. After pair-wise comparisons, the pri-
ority weights for main attributes, sub-attributes and alternatives
are combined to determine the priority weights of the alternative
suppliers. The supplier with the highest priority weight is selected
as the best supplier. Macros in MS Excel are used to calculate the
priority weights of the alternatives based on the questionnaire
forms used to facilitate comparisons of main attributes, sub-
attributes and alternatives. In particular, the approach presented
can handle the inherent uncertainty and imprecision of the human
decision making process and provide the exibility needed for the
decision maker to understand the decision problem.
The remainder of the paper is organized as follows: In Section 2,
researches related with supplier selection problem in literature are
given. In Section 3, we explain AHP, fuzzy AHP, fuzzy numbers, fuz-
zy sets and extent analysis method on fuzzy AHP. In Section 4, a
FEAHP based approach to select the best supplier rm providing
the most satisfaction for the criteria determined is developed and
the steps of each stage of the procedure is explained in detail.
The paper ends with concluding remarks in Section 5.
2. Literature survey
During recent years supply chain management and supplier
selection process have received considerable attention in the liter-
ature. Supplier selection is a multi-criteria problem and there are
not a lot of efcient techniques or algorithms that address this
problem. However three major groups of methods in the literature
are mathematical programming models, cost based models, and
categorical models.
Since supplier selection problems usually have several objec-
tives such as maximization of quality or maximization of prot
or minimization of cost, the problem can be modeled using math-
ematical programming. Weber and Current (1993) proposed a
multi-objective approach to supplier selection to aim at minimiz-
ing the price, maximizing the quality and on time delivery using
systems constraints and policy constraints in a mixed integer
model. Ghodsypour and OBrien (1998) proposed an integration
of AHP and linear programming to consider both tangible and
intangible factors in choosing the best suppliers and placing the
optimum order quantities among them such that the total value
of purchasing becomes maximum. ebi and Bayraktar (2003)
structure the supplier selection problem as an integrated lexico-
graphic goal programming and AHP model including both quanti-
tative and qualitative conicting factors. Wang, Huang, and
Dismukes (2004) use AHP and preemptive goal programming
based multi-criteria decision-making methodology is then devel-
oped to take into account both qualitative and quantitative factors
in supplier selection. Wang and Yang (2009) search supplier selec-
tion in a quantity discount environment using multi objective lin-
ear programming, AHP, and fuzzy compromise programming.
Since price has traditionally been a leading factor, selecting sup-
pliers based on cost has been a common approach. A popular appli-
cation of the cost approach has been calculating the total cost for
each purchase. The total cost of working with each supplier is cal-
culated and the cheapest one is selected. Timmerman (1986) pro-
poses costratio method which collects all costs related to quality,
delivery, and services and shows them as a benet or penalty per-
centage on unit price. Ellram (1990) explains that a formal total
cost approach explicitly recognizes cost factors in addition to price
and argues that any total cost approach should include transporta-
tion costs, receiving costs, quality costs, purchasing administrative
expenses and the price of the item.
Categorical methods are similar to the cost based methods be-
cause where the categorical methods determines the best supplier
using rating values on relevant supplier performance characteristics,
9657
the cost method does same process using dollar gures assigned to
the characteristics. Verma and Pullman (1998) examined the differ-
ence between managers rating of the perceived importance of dif-
ferent supplier attributes and their actual choice of suppliers in an
experimental setting. Humphreys, Wong, and Chan (2003) pre-
sented a framework of environmental criteria which a company
can consider during their supplier selection process.
AHP is widely used in studies related with categorical methods
since it is one of the extensively used multi-criteria decision-
making methods. It effectively keeps both qualitative and quantita-
tive data in decision making problems and is easier to understand
its process. Therefore it has been extensively applied to supplier
selection problems. Muralidharan, Anantharaman, and Deshmukh
(2001) proposed a methodology which makes estimation of the
rating by a group on an individual basis the principle of anonymity.
Handeld, Walton, Sroufe, and Melnyk (2002) illustrated the use of
AHP as a decision support model to help managers understand the
trade-offs between environmental dimensions. Pi and Low (2006)
proposed a supplier evaluation and selection system via Taguchi
loss function and AHP. Sevkli, Koh, Zaim, Demirbag, and Tatoglu
(2007) used data envelopment AHP to select the best supplier for
a TV company. Although Ounnar, Pujo, Mekaouche, and Giambiasi
(2007) did not make a supplier selection in their research, best
relationship between customer and supplier was examined using
most of supplier selection criteria in the literature.
In order to deal with uncertainties of the decision problem and
eliminate the disadvantages of AHP, fuzzy AHP is preferred in sup-
plier selection studies. Altinoz (2001) examined supplier selection
in general and specically in the Textile sector. Feng, Chen, and
Jiang (2005) proposed a comprehensive evaluation method based
on fuzzy decision theory and characteristics of supply chain man-
agement for optimal combination and selection among candidate
suppliers and outsourced parts. Haq and Kannan (2006) aims to
demonstrate how fuzzy AHP model can help in solving supplier
selection decisions in practice. Lu, Wu, and Kuo (2007) adds envi-
ronmental principles into supplier selection process by applying
fuzzy AHP. Chan, Kumar, Tiwari, Lau, and Choy (2008) propose a
fuzzy AHP approach for global supplier selection. Chamodrakas,
Batis, and Martakos (2009) uses fuzzy AHP to select supplier in
electronic sector.
The most common method used in the solution of fuzzy AHP
applications is the extent analysis method proposed by Chang
(1992). Kahraman, Cebeci, and Ulukan (2003) used fuzzy AHP to
select the best supplier rm providing the most satisfaction for
the criteria determined in the white good sector. Same procedure
is used to select the best catering rm providing the most customer
satisfaction by Kahraman, Cebeci, and Ruan (2004). Chan and Ku-
mar (2007) discussed a fuzzy extended AHP (FEAHP) approach
using triangular fuzzy numbers to represent decision makers com-
parison judgments and fuzzy synthetic extent analysis method to
decide the nal priority of different decision criteria.
3. AHP and fuzzy AHP
AHP is a systematic procedure to solve multi-criterion decision
making problems. It includes both subjective and objective evalu-
ation measures, offering a useful hierarchical procedure to control
the consistency of the evaluation measures and alternatives pre-
ferred by decision makers to reduce bias in decision making pro-
cess. The application of AHP to a decision problem involves four
steps (Zahedi, 1986). In structuring of the decision problem into
a hierarchical model, supplier selection problem is dened, objec-
tive is identied, criteria and attributes that must be satised to
objective are recognized. Objective is at rst level, criteria is at sec-
ond level, attributes are at third level, and decision alternatives are
9658 O. Kilincci, S.A. Onal / Expert Systems with Applications 38 (2011) 96569664

at fourth level in hierarchical structure of the problem. In making

pair-wise comparisons and obtaining the judgment matrix, the ele-
ments at a particular level are compared using nine-point numer-
ical scale to dene how much more an element is important than
other. If A and B are the elements to compare, 1 denes that A
and B are equal importance, and 9 denes that A is extreme
importance. All pair-wise comparisons are given in a judgment
matrix. Next step is to determine the local weights and consistency
of comparisons. Local weights of the elements are calculated from
the judgment matrix using eigenvector method. As the compari-
sons in the matrix are made subjectively, consistency ratio can
be computed. If the ratio is less than 0.1 human judgments is
acceptable. In last step, local weights at various levels are aggre-
gated to obtain nal weight of alternatives. The nal weights rep-
resent the rating of the alternatives in achieving the aim of the
multi-criterion decision making problem. Further information
about AHP can be found in Golden, Wasil, and Harker (1989), Saaty
(1980,1990).
Although AHP has been widely used to solve the multi-criterion
decision making problems, its most important disadvantage is that
it use a scale of one to nine which cannot handle uncertainty deci-
sions in comparison of the attributes. All comparisons during AHP
implementation may not include a certainty therefore the decision
maker needs more than nine-point scale to describe the uncer-
tainty. In this condition, linguistic variables and triangular fuzzy
numbers can be used to decide the priority of one decision variable
over other. Synthetic extent analysis method is used to decide the
nal priority weights based on triangular fuzzy numbers and so-
called as fuzzy extended AHP (FEAHP) (Chan & Kumar, 2007).
The FEAHP is the fuzzy extension of AHP to efciently handle
the fuzziness in the decision process to select the best supplier
by using both qualitative and quantitative data in the multi-attri-
bute decision making problems. In this approach triangular fuzzy
numbers are used in replace of nine-point of scale in traditional
AHP and then the synthetic extent value of the pair-wise compar-
ison is calculated by using the extent analysis method. After the
weight vectors are decided and normalized and normalized weight
vectors are determined, the nal priority weights of the alternative
suppliers are computed using the different weights of criteria and
attributes. As a result, the supplier with highest weight is selected
as the best supplier.
3.1. Fuzzy sets and fuzzy numbers
The concept of fuzzy logic was conceived by Zadeh to formulate
conclusions from vague, ambiguous or imprecise information. To
explain this information in mathematics, a fuzzy set which is a
class of objectives with continuum grades of membership was
developed. A membership function in fuzzy sets assigns to each ob-
ject a grade of membership in [0, 1].
A tilde  will be placed above a symbol if the symbol repre-
sents a fuzzy set. A triangular fuzzy number (TFN), M e is shown in
Fig. 1. A TFN is denoted simply as (l, m, u). The parameters l, m,
and u denote the smallest possible value, the most promising value
and the largest possible value that describe a fuzzy event (Kahr-
aman et al., 2004). When l = m = u, it is a non-fuzzy number by con-
vention (Chan & Kumar, 2007).
Each TFN has linear representations on its left and right side
such that its membership function can be dened as (Kahraman
et al., 2004);
8 9
> 0; x < l; > X
>
> >
>
< x  l=m  l; l 6 x 6 m; =
l Me
>
> u  x=u  m; m 6 x 6 u; >
>
>
: >
; P
0; x > u:
e (Kahraman et al., 2004)
Fig. 1. A triangular fuzzy number, M
Fuzzy numbers are intuitively easy to use in expressing the
decision-makers qualitative assessments. A fuzzy number can al-
ways be given by its corresponding left and right representation
of each degree of membership (Kahraman et al., 2004):
e Mly ; Mry l m  ly; u m  uy;
M y 2 0; 1; 2
where l(y)and r(y) denote the left side representation and the right
side representation of a fuzzy number, respectively. The algebraic
operations with fuzzy numbers can be found in Kahraman et al.
(2003).
3.2. Extent analysis method on fuzzy AHP
In crisp AHP, a discrete scale of one to nine is used to decide the
priority of one decision variable over another whereas in fuzzy AHP
fuzzy numbers or linguistic variables are used. In practice, decision
makers usually prefer triangular or trapezoidal fuzzy numbers.
Since fuzzy numbers are used in fuzzy AHP, the solution methods
differentiate from crisp AHP. The most common method used in
the solution of fuzzy AHP applications is the extent analysis meth-
od proposed by Chang (1992).
The extent analysis method is used to consider the extent of an
object to be satised for the goal, that is, satised extent. In the
method, the extent is quantied by using a fuzzy number. On
the basis of the fuzzy values for the extent analysis of each object,
a fuzzy synthetic degree value can be obtained, which is dened as
follows.
In a supplier selection problem, let X = {x1, x2, ..., xn} represent
the elements of the alternatives as an object set and let
U = {u1, u2, ..., um} represent the elements of the supplier selection
criteria as a goal set. According to the method of Changs (1992) ex-
tent analysis, each object is taken and extent analysis for each goal,
gi, is performed respectively. Therefore, m extent analysis values
for each object can be obtained, with the following signs:
M1gi ; M2gi ; . . . ; Mm
gi ; i 1; 2; . . . ; n 3
where all the M jgi j ; 2; :::; m are TFNs.
The steps of Changs extent analysis can be given as in the fol-
lowing (Kahraman et al., 2004):
(1) The value of fuzzy synthetic extent with respect to ith object
is dened as
" #1
m Xn X
m
Si M jgi  M jgi 4
1 j1 i1 j1
To obtain m j
j1 M g i , perform the fuzzy addition operation of m extent
analysis values for a particular matrix such that
 O. Kilincci, S.A. Onal / Expert Systems with Applications 38 (2011) 96569664 9659

! 0 0 0
X
m X
m X
m X
m
W 0 d A1 ; d A2 ; . . . ; d An T ; 13
Mjgi lj ; mj ; uj 5
j1 j1 j1 j1 where Ai (i = 1, 2, . . ., n) are n elements.
P P (4) Via normalization, the normalized weight vectors are
and to obtain ni1 m j 1
j1 M g i  , perform the fuzzy addition operation
W dA1 ; dA2 ; . . . ; dAn T ; 14
of Mjgi j 1; 2; :::; m values such that
! where W is a non-fuzzy number. This gives the priority weights of
X
n X
m X
n X
n X
n
one alternative over another.
M jgi li ; mi ; ui 6
i1 j1 i1 i1 i1
4. Application of supplier selection process in the white good
and then compute the inverse of the vector in Eq. (6) such that sector
0 1
" #1
Xn X
m B 1 1 1 C One of the biggest white good manufacturers in Europe, estab-
M jgi B
@P ; n ; n C 7
n P P A lished in Turkey, wants to select the best supplier rm for one its
i1 j1 ui mi li
i1 i1 i1
critical parts used in the production of washing machines. The
white good company has refrigerator, washing machine, and A/C
(2) The degree of possibility of M2 l2 ; m2 ; u2 P factories which are sought for its products in over 90 countries
M1 l1 ; m1 ; u1 is dened as for example England, France, Germany and Spain. The annual
VM 2 P M 1 supbminlM1 x; lM2 yc 8 capacity of the washing machine factory, which won the title of
yPx the most productive facility in the world in the sector by the
and can be equivalently expressed as follows: number of products per capita, is 2 million units/year. With the
washing machines produced in the factory, doing an export in
VM 2 P M 1 hgtM1 \ M2 lM2 d 9 the amount of 225 million USD annually is foreseen.
8 9 The manufacturing company discussed in the study is new in
< 1;
> if m2 P m1 ; >
= this sector and it is trying to increase its customer base. Therefore
0; if l1 P u2 ; 10 the production facilities and the ability of the supplier to increase
>
: >
l1 u2
m2 u2 m1 l1
; otherwise; ; its capacity should be taken into account to judge the best supplier
rm. Also due to the white good sector, there are seasonal uctu-
where d is the ordinate of the highest intersection point D between ations. The suppliers capacity should be exible enough to meet
lM1 and lM2 . In Fig. 2, the intersection between M1 and M2 can be the changes in the market demand. Since the production is based
seen. on assembly, all of the raw materials are outsourced. In consider-
To compare M1 and M2, we need both the values of VM 1 P M 2 ation that the capacity of the factory is 2 million units/year, there
and VM 2 P M 1 . are a huge number of raw materials supplied. Therefore we can
(3) The degree of possibility for a convex fuzzy number to be conclude that the supplier selection is one of the most important
greater than k convex fuzzy numbers Mi (i = 1, 2, ..., k) can be de- decision-making problems in this washing machine company.
ned by
4.1. Fuzzy AHP procedure for the supplier selection problem
VM P M1 ; M 2 ; . . . ; Mk VM P M 1 andM
P M2 and . . . andM P M k min VM P M i ; i In this supplier selection problem, the relative importance of
1; 2; 3; . . . ; k: 11 different decision criteria involves a high degree of subjective judg-
ment and individual preferences. The linguistic assessment of hu-
Assume that man feelings and judgments are vague and it is not reasonable to
0 represent them in terms of precise numbers. It feels more con-
d Ai min VSi P Sk : 12
dent to give interval judgments. Therefore triangular fuzzy num-
For k = 1, 2, ..., n; k i. Then the weight vector is given by bers were used in this problem to decide the priority of one
decision variable over another. The triangular fuzzy numbers were
determined from reviewing literature (Kahraman et al., 2003).
Then synthetic extent analysis method was used to decide the nal
priority weights based on triangular fuzzy numbers and so-called
as fuzzy extended AHP. In the following sections, the main steps
of the method will be explained in detail.
Step 1: Dene the main attributes and sub-attributes for supplier
selection to design the fuzzy analytical hierarchy process tree structure.
First the overall objective of the supplier selection problem has
been identied which was selection of the best supplier rm. In
the white good sector a lot of criteria should be taken into account
because the competition is really high. All of the possible important
criteria which could affect the supply of the critical part have been
discussed with an expert in the Production Planning Department.
Also other supplier selection studies in the literature were reviewed
with the expert. By combining the attributes determined by the ex-
pert and the attributes used in the literature, the main attributes
and the sub-attributes in the study were determined.
First three main attributes have been identied. The main attri-
butes determined by the expert were supplier, product perfor-
Fig. 2. The intersection between M1 and M2 (Zhu, Jing, & Chang, 1999). mance and service performance. The discussion has been further
9660 O. Kilincci, S.A. Onal / Expert Systems with Applications 38 (2011) 96569664

prolonged to decide the 14 sub-attributes with three potential sup- (g) Working with Kanban approach: According to Kanban
pliers. The sub-attributes determined by the expert were nancial approach, the supplier rm keeps the stock of the product
status, management approach, technical ability, quality systems being purchased. By this way, the rm gets rid of the stock
and process, geographical location, production facility and capac- cost. Also according to this approach, the rm only pays
ity, working with Kanban approach, product price, handling, prod- for the amount of product it uses.
uct quality, follow-up, technical support, lead time and
professionalism. The main and sub-attributes have been discussed 4.1.2. Product performance criteria
in detail below. Product performance criteria can be used to examine important
characteristics of the product being purchased. It depends on the
4.1.1. Supplier criteria type of the product being considered. The sub-criteria of the prod-
Supplier criteria evaluates whether the supplier ts its supply uct performance criteria are explained below:
and technology strategy. They are developed to measure important
aspects of the suppliers business: nancial status, management (a) Product price: Product price is a high percentage of in total
approach, technical ability, quality systems and processes, geo- cost of the product. Therefore purchasing department wants
graphical location, production facilities and capacity, and working to purchase the product with minimum price to decrease the
with Kanban approach. The sub-criteria of the supplier criteria are total cost.
explained below: (b) Handling: Handling includes of the transportation, storage,
receiving, and packaging operations. First three operations
(a) Financial status: One of the key elements in relations of them are important in view of stockout condition and
between rm and supplier is long-term commitment. Finan- rapid delivery time. Also packaging provides not to damage
cial status of the supplier can be a good indicator to provide the product during transportation.
this key element. (c) Product Quality: The quality of the product being purchased
(b) Management approach: If supplier has a strategic manage- from supplier provides the quality of the product being pro-
ment and applies new methods in management, the rm duced by the rm (or customer). Performance criteria of the
will easily construct good relations with the supplier. suppliers quality can be measured using rejection rate of the
(c) Technical ability: Technical abilities represent that supplier product which is dened in terms of the number of parts
ensures future improvements according to changing in cus- rejected by the customers because of quality problems.
tomers needs.
(d) Quality systems and process: To provide a high-quality prod- 4.1.3. Service performance criteria
uct, supplier should have a quality system including quality Service performance criteria can be used to evaluate the bene-
assurance, quality control procedures, quality control charts, ts provided by the suppliers services. Service performance crite-
documentation, continuously quality improvement, etc. ria should always be included in the supplier selection criteria
(e) Geographical location: The location of the suppliers rm is because any purchase involves some degree of service, such as or-
important because of shorter delivery time, lower transpor- der processing, delivery and support. When assessing the tness of
tation cost, and faster technical support. services of the supplier rms, follow-up, technical support, lead
(f) Production facility and capacity: Suppliers production facili- time and professionalism will be used. The sub-criterias of the ser-
ties should meet customers specic requirements. When vice performance criteria are explained below:
the customer develops new product or improves current
product according to market demand the suppliers facilities (a) Follow-up: The supplier should give information to customer
should be available to produce it. Also the supplier can about the production steps, its warehousing operations, and
increase its capacity to supply more orders from the transportation processes. Customer is informed delay in the
customer. shipment of the product.

Selection of the best supplier firm

Supplier Criteria Product Performance Criteria Service Performance Criteria

Kanban Approach

Technical Support
Technical Ability

Quality Systems

Professionalizm
Product Quality
FinancialStatus

Product Price
Working with
Geographical
Management

Lead Time
Follow-up
Handling
Location

Capacity

Supplier A Supplier B Supplier C

Fig. 3. The hierarchy of the supplier selection problem.

 O. Kilincci, S.A. Onal / Expert Systems with Applications 38 (2011) 96569664 9661

(b) Technical support: During the production of the product, cus- Table 2
tomer can need technical support from the supplier to make The fuzzy evaluation matrix with respect to the goal with linguistic variables.

changes in composition of the material to solve problems. Supplier Product Service

(c) Lead time: Lead time is the time from the moment the sup- performance performance
plier receives an order to the moment it ships it in the Supplier WP
absence of nished goods or intermediate inventory. It takes Product WP FSP
to actually manufacture the order without any inventory performance
Service performance
other than raw materials or supply parts. Therefore it
directly effects delivery time.
(d) Professionalism: Professionalism includes the knowledge,
accuracy, attitude and reliability of the supplier rm. Table 3
The fuzzy evaluation matrix with respect to the goal with triangular fuzzy numbers.

After the main attributes, sub-attributes and alternatives were Supplier Product Service
determined, the hierarchy of the supplier selection problem was performance performance
structured. Fig. 3 shows the structuring of the supplier selection Supplier (1, 1, 1) (2/3, 1, 3/2) (2/3, 1, 3/2)
problem hierarchy of four levels. The top level of the hierarchy rep- Product (2/3, 1, 3/ (1, 1, 1) (3/2, 2, 5/2)
performance 2)
resents the ultimate goal of the problem which is selection of the
Service (2/3, 1, 3/ (2/5, 1/2, 2/3) (1, 1, 1)
best supplier rm. The second level of the hierarchy is grouped under performance 2)
three categories, which are supplier attribute, product performance
attribute and service performance attribute. At the third level, these
ple, the calculation of the priority weights of the main attributes
main attributes are decomposed into various sub-attributes that
will be explained in detail below.
may affect the suppliers choice. Finally, the bottom level of the hier-
In Table 2, the fuzzy evaluation matrix with respect to the goal
archy represents the three alternative suppliers.
with linguistic variables can be found. By using the values in Table
Step 2: Calculate the weights of the main attributes, sub-attributes
1, the linguistic variables in the comparison matrix were converted
and alternatives. After the construction of the hierarchy, the differ-
to triangular fuzzy numbers. The fuzzy evaluation matrix with re-
ent priority weights of each main attribute, sub-attribute and alter-
spect to the goal with triangular fuzzy numbers can be seen in Ta-
native were calculated using the fuzzy AHP approach. The
ble 3. In order to nd the priority weights of the main attributes,
comparison of the importance of one main attribute, sub-attribute
rst the fuzzy synthetic extent values of the attributes were calcu-
and alternative over another were achieved by the help of the
lated by using Eq. (4). The different values of fuzzy synthetic extent
questionnaire. The questionnaires facilitate the answering of
of the three different main attributes were denoted by SS, SPP, SSP.
pair-wise comparison questions. The preference of one measure
over another was decided by the available research and the expe- SS 2:34; 3; 4  1=12:17; 1=9:5; 1=7:58
rience of the expert.
0:19; 0:32; 0:53 15
First the expert compared the main attributes with respect to
the main goal; then the expert compared the sub-attributes with
SPP 3:17; 4; 5  1=12:17; 1=9:5; 1=7:58
respect to the main attributes. At the end, the expert compared
the supplier rms with respect to each sub-attribute. The expert 0:26; 0:42; 0:66 16
used the linguistic variables to make the pair-wise comparisons.
Then the linguistic variables were converted to triangular fuzzy SSP 2:07; 2:5; 3:17  1=12:17; 1=9:5; 1=7:58
numbers. Table 1 shows the linguistic variables and their corre-
0:17; 0:26; 0:42 17
sponding triangular fuzzy numbers. Owing to the limited space,
the pair-wise comparison matrix of main attributes with respect The degree of possibility of Si over Si (i j) was determined by using
to the goal will be presented here. Eqs. (9) and (10).
After the pair-wise comparison matrices were formed, the con-
0:26  0:53
sistency of the pair-wise judgment of each comparison matrix was VSS P SPP 0:73 18
0:32  0:53  0:42  0:26
checked using the calculation method of consistency index and
consistency ratios in crisp AHP.
VSS P SSP 1 19
Each triangular fuzzy number, M = (l, m, u) in the pair-wise
comparison matrix was converted to a crisp number using
VSPP P SS 1 20
M crisp 4  m l u=6. After the fuzzy comparison matrices
were converted into crisp matrices, the consistency of each matrix
VSPP P SSP 1 21
was checked by the method in crisp AHP (Kwong & Bai, 2003).
After calculation, the consistency ratio of each comparison matrix 0:19  0:42
was found to be under 0.10. So we can conclude that the consis- VSSP P SS 0:79 22
0:26  0:42  0:32  0:19
tency of the pair-wise judgments in all matrices is acceptable.
Then the priority weights of each main attribute, sub-attribute
0:26  :42
and alternative were calculated using FAHP method. As an exam- VSSP P SPP 0:50 23
0:26  :42  0:42  0:26
With the help of Eq. (12), the minimum degree of possibility was
Table 1 stated as below:
The linguistic variables and their corresponding fuzzy numbers.
0
d S min0:73; 1 0:73 24
Equally preferred (EP) (1, 1, 1)
Weakly preferred (WP) (2/3, 1, 3/2) 0
Fairly strongly preferred (FSP) (3/2, 2, 5/2) d PP min1; 1 1 25
Very strongly preferred (VSP) (5/2, 3, 7/2)
Absolutely preferred (AP) (7/2, 4, 9/2) 0
d SP min0:79; 0:50 0:50 26
9662 O. Kilincci, S.A. Onal / Expert Systems with Applications 38 (2011) 96569664

Table 4
Priority vectors for the decision hierarchy.

Variables in level 1 Level 1 priorities Variables in level 2 Level 2 priorities Variables in level 3 Level 3 priorities
Supplier criteria (S) 0.33 Financial status (FS) 0.18 Supplier A 0.16
Supplier B 0.84
Supplier C 0
Management (M) 0 Supplier A 1
Supplier B 0
Supplier C 0
Technical ability (TA) 0.22 Supplier A 0.58
Supplier B 0.42
Supplier C 0
Quality systems (QS) 0.31 Supplier A 0.58
Supplier B 0.42
Supplier C 0
Geographical location (GL) 0.04 Supplier A 0
Supplier B 0
Supplier C 1
Capacity (C) 0.25 Supplier A 0.58
Supplier B 0.42
Supplier C 0
Working with Kanban 0 Supplier A 0.33
approach (WWKA)
Supplier B 0.22
Supplier C 0.45
Product performance 0.45 Product price (PP) 0.16 Supplier A 0.45
criteria (PP)
Supplier B 0.22
Supplier C 0.33
Handling (H) 0 Supplier A 0.58
Supplier B 0.42
Supplier C 0
Product quality (PQ) 0.84 Supplier A 0.45
Supplier B 0.33
Supplier C 0.22
0.22 Follow-up (F) 0 Supplier A 0.45
Supplier B 0.22
Supplier C 0.33
Technical support (TS) 0.23 Supplier A 0.45
Supplier B 0.33
Supplier C 0.22
Lead time (LT) 0.69 Supplier A 0.58
Supplier B 0.42
Supplier C 0
Professionalism (P) 0.08 Supplier A 0.58
Supplier B 0.42
Supplier C 0

Therefore the weight vector was given as W0 = (0.73, 1, 0.50). After
normalization process, the weight vector of the main attributes
which are supplier attribute, product performance attribute and
service performance attribute was found to be
WG = (0.33, 0.45, 0.22)T. We can conclude that the most important
attribute in the supplier selection process is product performance
attribute because it has the highest priority weight. Supplier attri-
bute is the next preferred attribute.
The same calculations were applied to the other pair-wise com-
parison matrices and the priority weights of each main attribute,
sub-attribute and alternative were calculated. The priority weights
of each main attribute, sub-attribute, and alternative can be found
in Table 4.
Step 3: Compute the overall score of each supplier and choose the
best supplier. In the last step, the priority weights of the main attri-
butes and sub-attributes were combined to determine priority
weights of the alternative suppliers. In Table 5, each column of
the matrix was multiplied by the priority weight at the top of
the column and then those values were added up for each row.
At the end, the priority weights of the alternatives with respect
to supplier attribute were calculated.
The same calculations have been applied to the sub-attributes
of product performance attribute and service performance attri-
bute and the priority weights of the alternatives with respect to
product performance and service performance attributes have
been calculated. The priority weights can be seen in Tables 6 and 7.

Table 5
Sub-attributes of supplier criteria.

FS M TA QS GL C WWKA Alternative priority weight

Weight 0.18 0 0.22 0.31 0.04 0.25 0
Alternative
Supplier A 0.16 1 0.58 0.58 0 0.58 0.33 0.48
Supplier B 0.84 0 0.42 0.42 0 0.42 0.22 0.48
Supplier C 0 0 0 0 1 0 0.45 0.04
 O. Kilincci, S.A. Onal / Expert Systems with Applications 38 (2011) 96569664 9663

Table 6 5. Conclusion
Sub-attributes of product performance criteria.

PP H PQ Alternative priority weight Supplier selection is a broad comparison of suppliers using a

Weight 0.16 0 0.84 common set of criteria and measures to identify suppliers with
Alternative the highest potential for meeting a rms needs consistently and
Supplier A 0.45 0.58 0.45 0.45 at an acceptable cost. Selecting the right suppliers signicantly re-
Supplier B 0.22 0.42 0.33 0.31
duces the purchasing costs and improves corporate competitive-
Supplier C 0.33 0 0.22 0.24
ness therefore supplier selection one of the most important
decision making problems.
In this study, a fuzzy AHP based approach is presented to select
Table 7 the best supplier rm providing the most customer satisfaction for
Sub-attributes of service performance criteria. the criteria determined for a washing machine company. Three
main attributes and 14 sub-attributes are determined based on lit-
F TS LT P Alternative priority weight
erature survey and the experience of the expert in the Production
Weight 0 0.23 0.69 0.08
Planning Department. Comparison of the main attributes and sub-
Alternative
Supplier A 0.45 0.45 0.58 0.58 0.55 attributes and suppliers are made using a questionnaire. The
Supplier B 0.22 0.33 0.42 0.42 0.40 weights of the main attributes, sub-attributes and alternatives
Supplier C 0.33 0.22 0 0 0.05 are calculated using fuzzy AHP approach. Best supplier is selected
after computing the overall score of each supplier.
To select the best supplier, all of the calculations in fuzzy AHP
are written using macros in MS Excel. It has several advantages.
Table 8
The supplier selection problem can be solved in a shorter and more
Main-attributes of the goal.
accurate way. It permits making hand-made errors while making
S PP SP Alternative priority weight the mathematical calculations. Also future changes in the prefer-
Weight 0.33 0.45 0.22 ences of the main attributes, sub-attributes or alternatives can be
Alternative easily reected on the priority weights of the alternatives without
Supplier A 0.48 0.45 0.55 0.48 calculating all of the steps from the beginning once more. In this
Supplier B 0.48 0.31 0.40 0.39
Supplier C 0.04 0.24 0.05 0.13
study, supplier selection process is presented for only one impor-
tant part of the washing machine. Fuzzy AHP approach can be eas-
ily adopted to select best supplier for each of other important parts
At the end, the priority weights of the alternatives with respect of the washing machine using the macros in MS Excel.
to the main attributes were combined and the priority weights of This study can be extended to incorporate the suppliers capac-
the alternatives were determined. The priority weights of the alter- ity constraints in the supplier selection process. Thus an integrated
native suppliers can be seen in Table 8. AHP and mathematical programming model can be proposed to
In order to shorten the solution process for the supplier selec- determine the best suppliers and the optimal order quantities from
tion problem, macros in Excel were used to calculate the priority the chosen suppliers. Also, the macros in MS Excel can be inte-
weights of the alternatives based on the questionnaire forms used grated into ERP software. Suppliers information, such as delivery
to facilitate comparisons of main attributes, sub-attributes and time, product quality, professionalism and etc., can be collected
alternatives. The expert lled out the questionnaire forms. Some from database of ERP. This integration may reduce time consuming
example parts of questionnaire form are given in appendices. Then efforts in supplier selection process.
macros in Excel made all of the calculations mentioned above to
calculate the priority weights of the alternative rms. The priority
Appendix A. Appendices
weights for the alternatives were found to be (0.48, 0.39, 0.13).
According to the nal score, Supplier A is the most preferred sup-
Some parts of questionnaire forms, which are written in MS Ex-
plier because it has the highest priority weight and Supplier B is
cel using macros, used to facilitate comparisons of main and sub-
the next recommended alternative supplier.
attributes.

Table A1.1
Questionnaire form used to facilitate comparisons of main attributes.

With respect to: the Best supplier Firm Importance (or preference) of one main-attribute over another
Questi ons Attributes Absolute Very strong Fairly Strong Weak Equal Weak Fairly Strong Very strong Absolute Attributes
Ql Supplier X Product
Q2 Supplier X Service
Q3 Product X Service

Table A1.2
Questionnaire form used to facilitate comparisons of sub-attributes of service.

With respect to: Service Importance (or preference) of one sub-attribute over another
Questions Sub-attributes Absolute Very strong Fairly Strong Weak Equal Weak Fairly Strong Very strong Absolute Sub-attributes
Q28 Follow-up X Technical Support
Q29 Follow-up X Lead Time
Q30 Follow-up X Professionalism
Q31 Technical Support X Lead Time
Q32 Technical Support X Professionalism
Q33 Lead Time X Professionalism
9664 O. Kilincci, S.A. Onal / Expert Systems with Applications 38 (2011) 96569664
With respect to the overall goal selection of the best supplier
rm
Q1. How important is supplier when it is compared with
product?
Q2. How important is supplier when it is compared with service?
Q3. How important is product when it is compared with service?
Table A1.1
With respect to the main attribute service
Q28. How important is follow-up when it is compared with tech-
nical support?
Q29. How important is follow-up when it is compared with lead
time?
Q30. How important is follow-up when it is compared with
professionalism?
Q31. How important is technical support when it is compared
with lead time?
Q32. How important is technical support when it is compared
with professionalism?
Q33. How important is lead time when it is compared with
professionalism?
Table A1.2
References
Altinoz, C. (2001). Supplier selection in textiles: A fuzzy approach. Ph.D. Thesis.
Raleigh, USA: North Carolina State University.
ebi, F., & Bayraktar, D. (2003). An integrated approach for supplier selection.
Logistics Information Management, 16(6), 395400.
Chamodrakas, I., Batis, D., & Martakos, D. (2009). Supplier selection in electronic
marketplaces using satiscing and fuzzy AHP. Experts Systems with Applications,
37(1), 490498.
Chan, F. T. S., Kumar, N., Tiwari, M. K., Lau, H. C. W., & Choy, K. L. (2008). Global
supplier selection: a fuzzy-AHP approach. International Journal of Production
Research, 46(14), 38253857.
Chan, F. T. S., & Kumar, N. (2007). Global supplier development considering risk
factors using fuzzy extended AHP-based approach. Omega, 35, 417431.
Chang, D. Y. (1992). Extent analysis and synthetic decision. Optimization Techniques
and Applications, 1, 352355.
Ellram, L. M. (1990). The supplier selection decision in strategic partnerships.
Journal of Purchasing and Materials Management, 26(4), 814.
Feng, D., Chen, L., & Jiang, M. (2005). Vendor selection in supply chain system: An
approach using fuzzy decision and AHP. In Proceedings of ICSSM05 international
conference on services systems and services management, Changqing, China (pp.
721725).
Ghobadian, A., Stainer, A., & Kiss, T. (1993). A computerized vendor rating system. In
Proceedings of the rst international symposium on logistics (pp. 321328). The
University of Nottingham, Nottingham, UK.
Ghodsypour, S. H., & OBrien, C. (1998). A decision support system for supplier
selection using an integrated analytic hierarchy process and linear
programming. International Journal of Production Economics, 56(57), 199212.
Ghodsypour, S. H., & OBrien, C. (2001). The total cost of logistics in supplier
selection, under conditions of multiple sourcing, multiple criteria and capacity
constraint. International Journal of Production Economics, 73, 1527.
Gofn, K., Szwejczewski, M., & New, C. (1997). Managing suppliers: When fewer can
mean more. International Journal of Physical Distribution & Logistics Management,
27(7), 422436.
Golden, B. L., Wasil, E. A., & Harker, P. T. (1989). The analytic hierarchy process. Berlin,
Heidelberg: Springer.
Handeld, R., Walton, S. V., Sroufe, R., & Melnyk, S. A. (2002). Applying
environmental criteria to supplier assessment: A study in the application of
the Analytical Hierarchy Process. European Journal of Operational Research, 141,
7087.
Haq, A. N., & Kannan, G. (2006). Fuzzy analytical hierarchy process for evaluating
and selecting a vendor in a supply chain model. International Journal of Advanced
Manufacturing Technology, 29, 826835.
Humphreys, P. K., Wong, Y. K., & Chan, F. T. S. (2003). Integrating environmental
criteria into the supplier selection process. Journal of Materials Processing
Technology, 138, 349356.
Kahraman, C., Cebeci, U., & Ruan, D. (2004). Multi-attribute comparison of catering
service companies using fuzzy AHP: The case of Turkey. International Journal of
Production Economics, 87, 171184.
Kahraman, C., Cebeci, U., & Ulukan, Z. (2003). Multi-criteria supplier selection using
fuzzy AHP. Logistics Information Management, 16(6), 382394.
Kwong, C. K., & Bai, H. (2003). Determining the importance weights for the
customer requirements in QFD using a fuzzy AHP with an extent analysis
approch. IEE Transactions, 35(7), 619626.
Lu, L. Y. Y., Wu, C. H., & Kuo, T. C. (2007). Environmental principles applicable
to green supplier evaluation by using multi-objective decision
analysis. International Journal of Production Research, 45(18-19), 4317
4331.
Muralidharan, C., Anantharaman, N., & Deshmukh, S. G. (2001). Vendor rating in
purchasing scenario: A condence interval approach. International Journal of
Operations & Production Management, 21(10), 13051325.
Ounnar, F., Pujo, P., Mekaouche, L., & Giambiasi, N. (2007). Customersupplier
relationship management in an intelligent supply chain network. Production
Planning and Control, 18(5), 377387.
Pi, W.-N., & Low, C. (2006). Supplier evaluation and selection via Taguchi loss
functions and an AHP. International Journal of Advanced Manufacturing
Technology, 27(56), 625630.
Saaty, T. L. (1980). The analytic hierarchy process. New York: McGraw-Hill.
Saaty, T. L. (1990). How to make a decision: The analytic hierarchy process.
European Journal of Operational Research, 48, 926.
Sevkli, M., Koh, S. C. L., Zaim, S., Demirbag, M., & Tatoglu, E. (2007). An application of
data envelopment analytic hierarchy process for supplier selection: A case
study of BEKO in Turkey. International Journal of Production Research, 45(9),
19732003.
Timmerman, E. (1986). An approach to vendor performance evaluation. Journal of
Purchasing and Materials Management, 22(4), 29.
Tracey, M., & Tan, C. L. (2001). Empirical analysis of supplier selection and
involvement, customer satisfaction, and rm performance. Supply Chain
Management: An International Journal, 6(4), 174188.
Verma, R., & Pullman, M. E. (1998). An analysis of the supplier selection process.
Omega, 26(6), 739750.
Wang, T.-Y., & Yang, Y.-H. (2009). A fuzzy model for supplier selection in quantity
discount environments. Expert Systems with Applications, 36(10), 12179
12187.
Wang, G., Huang, S. H., & Dismukes, J. P. (2004). Product-driven supply chain
selection using integrated multi-criteria decision-making methodology.
International Journal of Production Economics, 91, 115.
Weber, C. A., & Current, J. R. (1993). A multiobjective approach to vendor selection.
European Journal of Operational Research, 68, 173184.
Weber, C. A., Current, J. R., & Benton, W. C. (1991). Vendor selection criteria and
methods. European Journal of Operational Research, 50, 218.
Zahedi, F. (1986). The analytic hierarchy processa survey of the method and its
applications. Interfaces, 16(4), 96108.
Zhu, K.-J, Jing, Y., & Chang, D.-Y. (1999). A discussion on extent analysis method and
applications of fuzzy AHP. European Journal of Operational Research, 116,
450456.