Omega 35 (2007) 417 431 www.elsevier.

com/locate/omega

Global supplier development considering risk factors using fuzzy extended AHP-based approach
Felix T.S. Chan , Niraj Kumar
Department of Industrial and Manufacturing Systems Engineering, University of Hong Kong, Pokfulam, Hong Kong Received 22 November 2004; accepted 8 August 2005 Available online 29 September 2005

Abstract Global supplier development is a multi-criterion decision problem which includes both qualitative and quantitative factors. The global supplier selection problem is more complex than domestic one and it needs more critical analysis. The aim of this paper is to identify and discuss some of the important and critical decision criteria including risk factors for the development of an efcient system for global supplier selection. Fuzzy extended analytic hierarchy process (FEAHP) based methodology will be discussed to tackle the different decision criteria like cost, quality, service performance and suppliers prole including the risk factors involved in the selection of global supplier in the current business scenario. FEAHP is an efcient tool to handle the fuzziness of the data involved in deciding the preferences of different decision variables. The linguistic level of comparisons produced by the customers and experts for each comparison are tapped in the form triangular fuzzy numbers to construct fuzzy pair-wise comparison matrices. The implementation of the system is demonstrated by a problem having four stages of hierarchy which contains different criteria and attributes at wider perspective. The proposed model can provide not only a framework for the organization to select the global supplier but also has the capability to deploy the organizations strategy to its supplier. 2005 Elsevier Ltd. All rights reserved.
Keywords: Supply chain; Global supplier selection; Analytic hierarchy process; Fuzzy logic; Multi-attribute decision making problem

1. Introduction The current globalized market trend identies the necessity of the establishment of long-term business relationship with competitive global suppliers spread around the world. The selection of different unfamiliar international suppliers is a very critical multi-attribute decision-making problem. It depends on the broad comparison of suppliers using a common set of traditional criteria and measures. Different conicting selection
This paper was processed by Associate Editor Erwin Pesch.

Corresponding author. Tel.: +852 2859 7059;

fax: +852 2858 6535. E-mail address: ftschan@hkucc.hku.hk (F.T.S. Chan). 0305-0483/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.omega.2005.08.004

decisions make the task more complicated and risk prone. Research results indicated that supplier selection process is one of the most signicant variables, which has a direct impact on the performance of an organization. As the organization becomes more and more dependent on their suppliers, the direct and indirect consequences of poor decision making will become more critical. In a typical supply chain, coordination between manufacturer and suppliers is the important link in the distribution channel. The global competitive environment drives organizations highly dependent on the success of supplier selection process. Any deciency in coordination of the process will lead to excessive delays and poor customer service. In fact, suppliers are manufacturers external organizations or

418

F.T.S. Chan, N. Kumar / Omega 35 (2007) 417 431

business partners, and indeed their performance will decide the future performance of the whole supply chain. An effective methodology for global supplier selection is the demand from the current business scenario. The supplier development or selection problem considered in this paper is free from any quantity constraints and all the alternative suppliers have the ability to satisfy the manufacturers minimum criteria of demand, quality level and delivery schedule. Here one supplier is able to satisfy the entire need for a particular product. The problem of global supplier selection of a manufacturing company is formulated using different available literatures and the experience of the experts. The paper considers the important decision criteria like cost, quality, service performance and suppliers prole including the risk factors involved in deciding the best potential supplier. The risk factors discussed in this paper can properly take care of some of the relevant issues like political stability, geographical location, economic condition, and effect of terrorism on supplier selection process which requires much attention. The different decision criteria may vary depending on the need of the organization, its preferences about different criteria and the technological strategy. This may not be easy to convert its needs into useful criteria because needs are often expressed as a general qualitative concepts while criteria should be quantitatively evaluated. The overall objective of the selection is to identify the potential supplier which can stand on the organizations specic decision criteria. The decision criteria should be appropriate to its planned level of effort and many conicting criteria like low price with high quality, more advanced technology with high purchasing costs and excessive tariffs, etc., should be analyzed properly. During the global supplier selection the knowledge and critical information are taken from the group of experts in the respective areas and according to the priority of the rm the corresponding importance of one criterion over another can be decided. The analytic hierarchy process (AHP) [1] is widely used for tackling multi-attribute decision-making problems in real situations. In spite of its popularity and simplicity in concept, this method is often criticized for its inability to adequately handle the inherent uncertainty and imprecision associated with the mapping of the decision-makers perception to exact numbers. In the traditional formulation of the AHP, humans judgements are represented as exact (or crisp, according to the fuzzy logic terminology) numbers. However, in many practical cases the human preference model is uncertain and decision makers might be reluctant or unable to assign exact numerical values to the comparison judgements.

Since some of the evaluation criteria are subjective and qualitative in nature, it is very difcult for the decisionmaker to express the preferences using exact numerical values and to provide exact pairwise comparison judgements. It is more desirable for him to use interval or fuzzy evaluations. To improve the AHP method and to facilitate global supplier selection process, the paper discusses a fuzzy extended AHP (FEAHP) approach using triangular fuzzy numbers to represent decision makers comparison judgements and fuzzy synthetic extent analysis [2] method to decide the nal priority of different decision criteria. The fuzzy set theory resembles human reasoning in its use of approximate information and uncertainty to generate decisions. It has the advantage of mathematically represent uncertainty and vagueness and provide formalized tools for dealing with the imprecision intrinsic to many problems. The proposed FEAHP uses the triangular fuzzy numbers as a pair-wise comparison scale for deriving the priorities of different selection criteria and attributes. The weight vectors with respect to each element under a certain criterion is developed using the principle of the comparison of fuzzy numbers. As a result, the priority weights of the each supplier is calculated and based on that, the global suppliers are selected. In particular, the approach developed can adequately handle the inherent uncertainty and imprecision of the human decision making process and provide the exibility and robustness needed for the decision maker to understand the decision problem. These merits of the approach developed would facilitate its use in real-life situations for making effective decisions. The remainder of the paper is organized as follows: Section 2 discusses the past researches available in the area of the supplier selection and fuzzy decision making approaches. Section 3 contains the discussion about the different decision criteria and attributes considered in the selection of the global supplier. In Section 4, the proposed FEAHP is discussed. Section 5 discusses the complete implementation of the FEAHP approach. The priority weights computed for different criteria, attributes and alternatives are also discussed in this section. Section 6 contains the conclusion and the future direction of the research particularly in context with global supplier selection.

2. Background and motivation In the past, several methodologies have been proposed for the supplier selection problem but most of

F.T.S. Chan, N. Kumar / Omega 35 (2007) 417 431

419

them have limited their studies to domestic sourcing only [35]. As a result, many important criteria which are critical in deciding the global supplier have not been discussed properly. Some of the well-known examples of systematic analysis for domestic supplier selection include a categorical method, weighted point method [6], matrix approach [7], vendor performance matrix approach [8], vendor prole analysis (VPA) [9], analytic hierarchy process (AHP) [10,11], and multiple objective programming (MOP) [1216]. Although both AHP and MOP have some advantages over other existing approaches, they are still suffering from some criticisms. AHP cannot effectively take into account risk and uncertainty in assessing the suppliers performance because it presumes that the relative importance of attributes affecting the suppliers performance is known with certainty [17]. MOPs inherent computational complexity often prohibits consideration of many attributes essential for supplier selection. In MOP problem formulation, many of the objectives are regarded as constraints but this can be avoided in converting the priorities into the ratings with regard to each criterion using pair-wise comparison. The drawback of MOP is that it requires arbitrary aspiration levels and cannot accommodate subjective criteria [18]. Apart from the above-mentioned techniques, Choy and Lee [19] used the case-based reasoning approach for intelligent supplier selection to enhance the performance of the selection as compared to traditional approaches. Verma and Pullman [20] studied the perceived importance of supplier selection criteria and identify the relative weights of the attributes in actual selection of suppliers. Liu et al. [21] used data envelopment analysis (DEA) to compare the performance evaluation of different supplier for best selection. OBrien and Ghodsypour [22] proposed integration of an AHP and linear programming to consider both tangible and intangible factors in choosing the best suppliers and placing optimum order quantities among them. Weber and Ellram [14] explored the use of a multi-objective programming approach as a method for supplier selection in just in time (JIT) setting. Wang et al. [23] used the advantages of AHP and preemptive goal programming to incorporate both quantitative and qualitative factor in supplier selection problem. The mathematical programming model applied to supplier selection has problems in including qualitative criteria that are very important in the decision making, especially for supplier partnership policies [24]. Fuzzy extended approach is applied in this paper, since it is a more efcient approach in treating the fuzziness of data involved in analyzing the qualitative factors than other meth-

ods such as weighting or mathematical programming models [25]. With reference to the past literature, it can be observed that only limited discussion of global supplier selection problem which in fact is much different from the domestic one and involves much complexity in terms of the selection of the different effective criteria [26]. In this paper, we have used the attributes which are the deciding factors in global supplier selection. This paper goes beyond the previous literature and tried to consider all the critical criteria, which have not got much attention to decide the best global partner. In this paper FEAHP is used to tackle the global supplier selection because of the ability of fuzzy sets to resemble with the human-decision nature. Some preliminary work in fuzzy AHP were presented by Van laarhoven and Pedrycz [27] and Buckley [28]. They extended Saatys AHP [1] to deal with the imprecision and subjectiveness of the pairwise comparison process using fuzzy utilities represented by fuzzy numbers. Their fuzzy utilities need to be ranked to prioritize the concerned alternatives. This ranking methods can be quite complex and may produce unreliable results [29]. Mikhailov [30] proposed fuzzy preference programming method to derive optimal crisp priorities, which are obtained from fuzzy pairwise comparison judgements based on -cuts decomposition of the fuzzy judgements into a series of interval comparisons. However, Fuzzy preference programming method claimed its superiority over some of the existing fuzzy prioritization methods [27,28] but the mathematical complexity involved may restrict its practicability. Chiou et al. [31] used a fuzzy hierarchical analytic process to determine the weights of criteria from subjective judgements and a non-additive integral technique to evaluate the performance of sustainable development strategies for aquatic products processors. Chang [32] introduced a new extent analysis approach for the synthetic extent values of the pairwise comparison for handling fuzzy AHP. The proposed FEAHP with extent analysis is simple and easy for implementation to prioritize decision variables as compared with the conventional AHP [33]. Kahraman et al. [34] used the fuzzy AHP for domestic supplier selection with only 3 criteria and 11 attributes and neglected the many important criteria which create the uncertainty in supplying the products, i.e. the risk factors involved in global supplier selection. Apart from the above-mentioned applications, fuzzy AHP has also been used in different type of problem environment [3537]. The objective of this paper is to discuss the relevant decision criteria like cost, quality, service performance, and suppliers prole including the risk factors involved

420

F.T.S. Chan, N. Kumar / Omega 35 (2007) 417 431

in global supplier selection and to provide an efcient approach to handle it effectively. The risk factors, discussed in this paper, address some of the important and interesting issues like terrorism, government stability, economic status, and geographical location of supplier to match the supplier selection process with the current political and business scenario for greater efciency. This research can provide an appropriate and suitable approach to tackle the fuzziness involved in the human decision making, with the aim to remove complex and unreliable process of comparing fuzzy utilities for efciently solving the global supplier selection problem. This can provide guidelines to the decision makers for taking decisions regarding their partner selection with great precision considering the current relevant issues.

3.1. Overall cost of the product (C1 ) This criterion is one of the important criteria in assessing the global supplier because it can dictate the international procurement cost. Prot maximization cannot be achieved without the cost minimization. The factors (attributes) affecting this criterion can be stated as follows: Product price (A1 ): The rm always requires the minimum price of the product to increase the protability. The rm therefore must nd a low-cost supply base where it can minimize manufacturing cost related to the production of the product. The processing cost, maintenance cost, warranty cost, and other costs related to the manufacturing of the product determine the total price of the product. Freight cost (A2 ): In global sourcing it is customary to check the high freight cost of each supplier with a great precision. This contains the lengthy distribution channel cost, transport expenses, inventory cost, handling and packaging cost, damages during transportation, and insurance costs. These costs are usually high in context with international suppliers, therefore various freight terms should be carefully analyzed. Tariff and custom duties (A3 ): Different countries have different norms of imposing tariff and custom duties on the goods and services purchased. This should be considered with precision during global supplier selection and preferences should be given to the supplier country having less duties and taxes, as a result, it may help in reducing the nal price of the product to the customers. 3.2. Quality of the product (C2 ) The most important factor leading to international sourcing is the high quality of the products, which makes a good impression to the customer. The quality of the product can be measured in terms of following attributes: Rejection rate of the product (A4 ): The rejection rate of the product is dened in the terms of the number of parts rejected by the customers in xed time period because of some quality problems. It also includes the defective parts detected in the incoming quality control and the production line. Increased lead time (A5 ): The defective parts, which are not detected in quality control process but noticed during production, can increase the lead time of production. It should be analyzed properly to know the bad effect of the poor quality products on the production schedule.

3. Selection criteria for global suppliers Miller [38] observed that most decision makers cannot simultaneously handle more than 7 to 9 factors when making a decision. As such it is necessary to break down the complex problem into more manageable subproblems. Owing to the large number of factors affecting the supplier selection decision, an orderly sequence of steps should be required to tackle it. The problem taken here has four level of hierarchy, and the different decision criteria, attributes and the decision alternatives, will be further discussed. The main objective here is the selection of best global supplier for a manufacturing rm. Application of common criteria to all suppliers makes objective comparisons possible. The criteria which are considered here in selection of the global supplier are: Overall cost of the product Quality of the product Service performance of supplier Suppliers prole Risk factor

These criteria can be decomposed into various other attributes. The main criteria and numbers of attributes relevant to global supplier selection are described below. The criteria are denoted by Ci , attributes by Aj , and alternatives by Sk (where, i, j, k = 1, 2, . . .). The hierarchy of the selection criteria, attributes, and decision alternatives can be seen in Fig. 1. In the hierarchy, the overall objective is placed at level 1, criteria at level 2, attributes at level 3, and the decision alternatives at level 4.

F.T.S. Chan, N. Kumar / Omega 35 (2007) 417 431

421

Level 1: Overall Objective (O)

Global supplier selection

Level 2: Criteria (C)

Cost (C1)

Quality (C2)

Service performance (C)

Suppliers profile (C )

Risk factor (C )

Level 3: Attributes (A)

A1 A3

A2

A4 A6

A5 A7

A8 A10

A9 A11

A12 A14

A13 A15

A16 A18

A17 A19

Level 4: Decision Alternatives(S)

Supplier1 (S1)

Supplier2 (S2)

Supplier3 (S3)

Fig. 1. Hierarchy for the global supplier selection.

Quality assessment (A6 ): This encounters the issues like whether or not the frequent quality assessment of the parts has been done by the supplier. Accordingly the buyer should investigate whether or not potential suppliers are certied for strict quality assurance and have a strong commitment for preventing quality failures. Remedy for quality problems (A7 ): This attribute helps in investigating the suppliers ability to solve the quality problems detected by the manufacturer. The ability of the supplier to efciently tackle the quality problems should also be included in the analysis. 3.3. Service performance (C3 ) The performance of the supplier in providing service to the manufacturer is the prime criteria to decide its suitability for a particular product. The good service given by the supplier may help in increase the customer base and therefore, this criterion is important in global supplier selection. It is analyzed based on the following attributes: Delivery schedule (A8 ): The ability of the supplier to follow the predened delivery schedule is always the prime criteria for selection in this fast moving world. In the process of selection of the most appropriate supplier, the manufacturer should access the complete sup-

ply chain network on time and have the ability to follow the exact delivery schedule according to the customers demand. Technological and R&D support (A9 ): Technology is advancing at a very fast pace in this competitive world to satisfy the customer rst and get its appreciation. Furthermore, suppliers are more likely to assume greater responsibility for outsourced design, engineering service, prototype development and research to increase the performance of the products. The suppliers ability to provide advanced technological and R&D support to produce a good product is of prime concern in global supplier selection. Response to changes (A10 ): The ability of the supplier to change according to the customers demand, price structure, order frequency and current business scenario has greater impact on the process of sourcing overseas. It can affect the performance of the rm in case of urgent and uncertain demand. More exible supplier in terms of the demand constraints can be chosen for better performance towards customer. Ease of communication (A11 ): The ease of communication and negotiability with the suppliers decide the long-term relation between the supplier and manufacturer. Since languages, business customs, ethics and communication devices vary from country to country,

422

F.T.S. Chan, N. Kumar / Omega 35 (2007) 417 431

the manufacturing rm should consider attributes such as cultural similarity, ethical standard and electronic data interchange capabilities in order to ensure effective communication and negotiation with the foreign supplier. 3.4. Suppliers prole (C4 ) The performance and past history of the suppliers help in taking decisions for its selection. The value-ofpartnership should be analyzed based on its track record to decide the superiority of it over other alternative suppliers. Some of the important characteristics of suppliers are summarized below: Financial status (A12 ): The nancial status of the supplier can be analyzed by getting the information about the annual turn over of the supplier and their nancial structure based on the past history. Foreign suppliers shaky nancial situation can gradually deteriorate the long-term relation with the manufacturer. Financial stability is the necessary requisite for long-term partnership between supplier and the manufacturer. Customer base (A13 ): The response of the customers towards the supplier is one of the important factors to decide the performance of the supplier. Suppliers with good customer base should be preferred than the others with no satisfactory customers. Performance history (A14 ): The performance history of the supplier should be analyzed carefully keeping in mind the competitive nature of the supplier, its past production schedule, response to market, and its ability to make commercial relations and business references. Production facility and capacity (A15 ): The production facilities and ability of the supplier to increase its capacity should also be taken into account to judge the best one. The potential production capability of each supplier should be analyzed to meet a specied production plan and also to develop a new product according to the market demand. 3.5. Risk factor (C5 ) Owing to a number of exogenous factors inuencing international sourcing, global supplier selection is much riskier than its domestic counterpart [26]. Consequently, the global supplier selection decision is most strongly affected by perceived risks. Some of the risk factors identied and analyzed in this paper could not be seen in any previous study, which have the signicant capability to affect the selection process of the global supplier.

They can be stated as below: Geographical location (A16 ): The location of the supplier and its physical and social status should be analyzed properly before selection of global partner. The mother country of the supplier, the location of plant, the nature of natural calamities, and other factors should be checked before the selection because for long-term relation it may create problems in the supply of the goods. Political stability (A17 ): The political status of the suppliers country and its nature towards the business policies may affect the long-term relationship between the supplier and the manufacturer. The more stable government should be preferred. Because during the change of political leadership, different foreign countries policies will be changed accordingly and this may create big problems in further maintaining the relationship with supplier. In this connection, this must be analyzed in great precision with the help of the relevant experts. Economy (A18 ): The economic status of the suppliers country may affect the currency exchange rate, local price control and so forth. This can result in higher hidden costs for international sourcing and therefore it should also be taken into account during the global supplier selection process. Terrorism (A19 ): Terrorism is identied as a great enemy of the business process after 9/11 incident. At present, manufacturers are more concerned about this issue because it can hamper the delivery schedule and badly affect the performance of the rm towards the customers. The suppliers policies to effectively tackle these issues should be analyzed properly. The perception and assessment of risk associated with terrorism may be biased and subjective so great precision is required to decide its preference over other alternative risk factors. Risk associated with terrorism depends on the type of the product, region of the supplier, business strategy of the manufacturer and also the mind set of the decision maker. For example, the risk is higher in case of the supply of oil and gas products through pipeline from terrorism affected areas, therefore this criterion has much importance for a specic type of product. Different suppliers have different policies for the avoidance of the possible terrorist threat on the supply of the products. The performance measurement method of the effective policies can also inuence the decision makers assessment for the best supplier. A comprehensive study of the inuence of possible terrorist threat on the supply of the products has been required to efciently decide its priority over other risk factors. The above-mentioned criteria and attributes help in deciding the best global supplier for an organization. The preferences of one over other have been

F.T.S. Chan, N. Kumar / Omega 35 (2007) 417 431

423

decided by the experts and researchers of the respective eld. 4. Fuzzy extended AHP 4.1. Background AHP has been widely used to address the multicriterion decision making problems. However, it has been generally criticized because of the use of a discrete scale of one to nine which cannot handle the uncertainty and ambiguity present in deciding the priorities of different attributes. The relative importance of different decision criteria in global supplier selection involves a high degree of subjective judgment and individual preferences. The hierarchy of the decision variables is the subject of a pairwise comparison of the AHP. In conventional AHP, the pairwise comparison is established using a nine-point scale which converts the human preferences between available alternatives as equally, moderately, strongly, very strongly or extremely preferred. Even though the discrete scale of AHP has the advantages of simplicity and ease of use, it is not sufcient to take into account the uncertainty associated with the mapping of ones perception to a number [33]. The linguistic assessment of human feelings and judgements are vague and it is not reasonable to represent it in terms of precise numbers. It feels more condent to give interval judgements than xed value judgements. Hence, triangular fuzzy numbers are 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). Fuzzy set theory has proven advantages within vague, imprecise and uncertain contexts and it resembles human reasoning in its use of approximate information and uncertainty to generate decisions. It was specially designed to mathematically represent uncertainty and vagueness and provide formalized tools for dealing with the imprecision intrinsic to many decision problems. Fuzzy set theory implements classes and grouping of data with boundaries that are not sharply dened (i.e. fuzzy). Fuzzy set theory includes the fuzzy logic, fuzzy arithmetic, fuzzy mathematical programming, fuzzy graph theory and fuzzy data analysis, usually the term fuzzy logic is used to describe all of these. The FEAHP is the fuzzy extension of AHP to efciently handle the fuzziness of the data involved in the decision of best global supplier. It is easier to understand and it can effectively handle both qualitative and quantitative data in the multi-attribute decision making problems. In this

~ N

1.0

Nl(y)

Nr(y)

0.0

N n1 n2 n3

Fig. 2. A triangular fuzzy number, N.

approach triangular fuzzy numbers are used for the preferences of one criterion over another and then by using the extent analysis method, the synthetic extent value of the pairwise comparison is calculated. Based on this approach, the weight vectors are decided and normalized, thus the normalized weight vectors will be determined. As a result, based on the different weights of criteria and attributes the nal priority weights of the alternative global suppliers are decided. The highest priority would be given to the supplier with highest weight. 4.2. Triangular fuzzy number and representation of preferences A fuzzy set [39,40] is characterized by a membership function, which assigns to each object a grade of membership ranging between 0 and 1. In this set the general terms such as large, medium, and small each will be used to capture a range of numerical values. A fuzzy number is a special fuzzy set, such that M = {(x, M (x), x R},where the value of x lies on the real line R1 i.e. < x < and M (x) is a continuous mapping from R1 to the close interval [0,1]. If n1 , n2 and n3 , respectively, denote the smallest possible value, the most promising value and the largest possible value that describe a fuzzy event then the triangular fuzzy number (TFN) can be denoted as a triplet (n1 , n2 , n3 ) where, n1 n2 n3 . When n1 = n2 = n3 , it is a non-fuzzy number by convention. The membership function can be dened as
N (x) =

(x n1 )/(n2 n1 ), x [n1 , n2 ] (n3 x)/(n3 n2 ), x [n2 , n3 ] 0 otherwise

(1)

A TFN N is shown in Fig. 2.

424

F.T.S. Chan, N. Kumar / Omega 35 (2007) 417 431

Equal

Moderate

Fairly Strong

Very Strong

Absolute

1.0 N(x)

N1

N2

N3

N4

N5

N6

N7

N8

N9

The value of fuzzy synthetic extent with respect to the ith object is dened as 1
m

Fi =
j =1
0.5

Noi
j

i=1 j =1 j

Noi
j

(3)

0.0

The value of m=1 Noi can be found by performing j the fuzzy addition operation of m extent analysis values from a particular matrix such that
m j =1

Fig. 3. The membership functions of the triangular fuzzy numbers.

Noi =
j

n1j ,
j =1 j =1 n i=1

n2j ,
j =1 j m j =1 Noi ]

n3j

(4)

Fuzzy numbers are intuitively easy to use in expressing the decision-makers qualitative assessments. A fuzzy number can be given by its corresponding left and right representation of each degree of membership: N = (N l(y) , N r(y) ) = (n1 +(n2 n1 )y, n3 +(n3 n2 )y),

and the value of [

can be obtained by
j

performing the fuzzy addition operation of Noi (j = 1, 2, . . . , m) such that

n m

i=1 j =1

Noi =
i=1 j

n1j ,
i=1

n2j ,
i=1

n3j

(5)

y[0, 1],

(2)

where l(y) and r(y) denote the left- and right-hand side representation of a fuzzy number, respectively. The algebraic operations with fuzzy numbers used in this paper can be found in [32]. TFNs N1 , N3 , N5 , N7 and N9 are used to represent the pairwise comparison of decision variables from Equal to Absolutely preferred, and TFNs N2 , N4 , N6 and N8 represent the middle preference values between them. Fig. 3 shows the membership functions of the TFNs, Ni =(ni1 , ni2 , ni3 ), where i =1, 2, . . . , 9 and ni1 , ni2 , ni3 are the lower, middle and upper values of the fuzzy number Ni respectively. Higher the value of (ni3 ni1 ) or (ni1 ni3 ) signies the greater fuzziness of the judgement. 4.3. Calculation of priority weights at different level of hierarchy If the object set is denoted by P = {p1 , p2 , . . . , pn } and the objective set is denoted by Q={q1 , q2 , . . . , qm }, then according to the concept of extent analysis [2] each object is taken and extent analysis for each objective Oi is performed, respectively. Therefore the m extent analysis values for each object are obtained with the following signs:
1 2 m Noi , Noi , . . . , Noi , j

m 1 can be calculated by the inand [ n i=1 j =1 Noi ] verse of Eq. (5) as follows: 1

i=1 j =1

Noi
j

, n i=1 n3i

, n i=1 n2i

1
n i=1 n1i

(6)

the degree of possibility of N1 = (n11 , n12 , n13 ) N2 = (n21 , n22 , n23 ) is dened as V (N1 N2 ) = sup [min(
x y N1 (x), N2 (y))]

(7)

when a pair (x, y) exists such that x y and N1 (x) = N2 (y) = 1, then we have V (N1 N2 ) = 1. Since N1 and N2 are convex fuzzy numbers so, V (N1 and V (N2 N1 ) = hgt(N1 N2 ) =
N1 (d),

N2 ) = 1

if n11

n21

(8)

where d is the ordinate of the highest intersection point D between N1 and N2 . when N1 = (n11 , n12 , n13 ) and N2 = (n21 , n22 , n23 ) then ordinate of D is computed by V (N2 N1 ) = hgt(N1 N2 ) n11 n23 = . (n22 n23 ) (n12 n11 ) (9)

where i = 1, 2, . . . , n,

where all the Noi (j = 1, 2, . . . , m) are triangular fuzzy numbers.

For the comparison of N1 and N2 , both the values of V (N1 N2 ) and V (N2 N1 ) are required.

F.T.S. Chan, N. Kumar / Omega 35 (2007) 417 431

Table 1 The Fuzzy evaluation of criteria with respect to the overall objective
O C1 C2 C3 C4 C5 C1 (1, 1, 1) (2/5, 1/2, 2/3) (2/5, 1/2, 2/3) (2/7, 1/3, 2/5) (2/7, 1/3, 2/5) C2 (3/2, 2, 5/2) (1, 1, 1) (2/5, 1/2, 2/3) (2/7, 1/3, 2/5) (2/7, 1/3, 2/5) C3 (3/2, 2, 5/2) (3/2, 2, 5/2) (1, 1, 1) (2/5, 1/2, 2/3) (2/5, 1/2, 2/3) C4 (5/2, 3, 7/2) (5/2, 3, 7/2) (3/2, 2, 5/2) (1, 1, 1) (2/5, 1/2, 2/3) C5 (5/2, 3, 7/2) (5/2, 3, 7/2) (3/2, 2, 5/2) (3/2, 2, 5/2) (1, 1, 1)

425

WO 0.45 0.35 0.09 0.06 0.05

The weight vector from Table 1 is calculated as WO = (0.45, 0.35, 0.09, 0.06, 0.05)T .

The degree possibility for a convex fuzzy number to be greater than k convex fuzzy numbers Ni (i = 1, 2, . . . , k) can be dened by V (N N1 , N2 , . . . , Nk ) = V [(N N1 ) and (N N2 ) and . . . and (N = min V (N Ni ), i = 1, 2, . . . , k. if m(Pi ) = min V (Fi Fk ),

prolonged to decide the nineteen attributes (or subcriteria) with three potential suppliers. The following steps have been considered to form the hierarchy: (1) Dene the global supplier selection problem. (2) Identify the overall objective. What the rm trying to accomplish? (3) Identify the criteria and attributes that must be satised to fulll the overall objectives. (4) Identify decision alternatives or outcomes. (5) Structure the hierarchy placing the objective at rst level, criteria at second level, attributes at third level, and decision alternatives at fourth level. The whole hierarchy of the selection of best global supplier can be easily visualized from Fig. 1. After the construction of the hierarchy the different priority weights of each criteria, attributes and alternatives are calculated using the FEAHP approach. The comparison of the importance or preference of one criterion, attribute or alternative over another can be done with the help of the questionnaire. The preference of one measure over another is decided by the available research, the current business scenario and by the experience of the experts. The method of calculating priority weights of the different decision alternatives using FEAHP is discussed below. First the fuzzy evaluation matrix of the criteria is constructed by the pairwise comparison of the different criterion relevant to the overall objective using triangular fuzzy numbers, which is shown in Table 1. The consistency of the pairwise judgement of each comparison matrix is also checked using the calculation method of consistency index and consistency ratio discussed in [33]. The value of fuzzy synthetic extent with respect to each criterion is calculated by using Eq. (3) and the formula for algebraic operations of the fuzzy set.

Nk )] (10) (11)

for k = 1, 2, . . . , n; k = i. then the weight vector is given by Wp = (m(P1 ), m(P2 ), . . . , m(Pn ))T , (12)

where Pi (i = 1, 2, . . . , n) are n elements. After normalizing Wp , we get the normalized weight vectors W = (w(P1 ), w(P2 ), . . . , w(Pn ))T , (13)

where W is a non-fuzzy number and this gives the priority weights of one alternative over other.

5. Application and discussion of FEAHP in global supplier selection The problem discussed here is concerned with a manufacturing company, searching the best global supplier for one of its most critical parts used in assembling process. The company wants to take into account all the possible important criteria which can affect the supply of the critical part. A decision-making group is formed which consists of the experts from each strategic decision area. Detailed discussion on every criterion, attribute and alternative supplier have been conducted and based on primary discussion, ve critical criteria have been identied. The discussion has been further

426

F.T.S. Chan, N. Kumar / Omega 35 (2007) 417 431

Table 2 The fuzzy evaluation of the attributes with respect to criterion C1

C1 A1 A2 A3 A1 (1, 1, 1) (2/5, 1/2, 2/3) (2/5, 1/2, 2/3) A2 (3/2, 2, 5/2) (1, 1, 1) (2/5, 1/2, 2/3) A3 (3/2, 2, 5/2) (3/2, 2, 5/2) (1, 1, 1) WC1 0.58 0.30 0.12

Table 3 The fuzzy evaluation of the attributes with respect to criterion C2

C2 A4 A5 A6 A7 A4 (1, 1, 1) (2/5, 1/2, 2/3) (2/3, 1, 3/2) (2/7, 1/3, 2/5) A5 (3/2, 2, 5/2) (1, 1, 1) (2/3, 1, 3/2) (2/5, 1/2, 2/3) A6 (2/3, 1, 3/2) (2/3, 1, 3/2) (1, 1, 1) (2/5, 1/2, 2/3) A7 (5/2, 3, 7/2) (3/2, 2, 5/2) (3/2, 2, 5/2) (1, 1, 1) WC2 0.41 0.21 0.26 0.12

The different values of fuzzy synthetic extent with respect to the ve different criteria are denoted by F1 , F2 , F3 , F4 , and F5 , respectively. F1 = (9, 11, 13) (1/39.62, 1/33.32, 1/27.52) = (0.23, 0.33, 0.47); F2 = (7.9, 9.5, 11.17) (1/39.62, 1/33.32, 1/27.52) = (0.20, 0.28, 0.41); F3 = (4.8, 6, 7.34) (1/39.62, 1/33.32, 1/27.52) = (0.12, 0.18, 0.27); F4 = (3.46, 4.16, 4.97) (1/39.62, 1/33.32, 1/27.52) = (0.09, 0.12, 0.18); F5 = (2.36, 2.66, 3.14) (1/39.62, 1/33.32, 1/27.52) = (0.06, 0.08, 0.11). The degree of possibility of Fi over Fj (i = j ) can be determined by Eqs. (8) and (9). V (F1 V (F1 V (F2 V (F2 F2 ) = 1, F4 ) = 1, F1 ) = V (F1 V (F1 F3 ) = 1, F5 ) = 1,

V (F5 V (F5

F1 ) = 0.92, F3 ) = 0.11,

V (F5 V (F5

F2 ) = 0.82, F4 ) = 0.33.

With the help of Eq. (11), the minimum degree of possibility can be stated as below: m(C1 ) = min V (F1 F2 , F3 , F4 , F5 ) = min (1, 1, 1, 1) = 1. Similarly, m(C2 ) = 0.78, m(C3 ) = 0.21, m(C4 ) = 0.14, m(C5 ) = 0.11. Therefore the weight vector is given as Wc = (1, 0.78, 0.21, 0.14, 0.11)T and after the normalization process, the weight vector with respect to decision criteria C1 , C2 , C3 , C4 , and C5 can be presented as follows: Wo = (0.45, 0.35, 0.09, 0.06, 0.05)T . The complete result is shown in Table 1. Now the different attributes are compared under each of the criterion separately by following the same procedure as discussed above. Whenever the value of (n11 n23 ) > 0, the elements of the matrix must be normalized and then do the same process to nd the weight vector of each attribute. The fuzzy evaluation matrices of attributes and the weight vectors of each attribute are shown in Tables 26. Similarly the fuzzy evaluation matrices of decision alternatives and corresponding weight vector of each alternative with respect to corresponding attributes are determined. The priority weights of suppliers with respect to the each criterion are given by adding the weights per supplier multiplied by weights of the corresponding attributes. The results are shown in Tables 711.

0.23 0.41 = 0.78, (0.28 0.41) (0.33 0.23) V (F2 F4 ) = 1, V (F2 F5 ) = 1.

F3 ) = 1,

Similarly, V (F3 F1 ) = 0.21, V (F3 V (F3 F4 ) = 1, V (F3 F5 ) = 1, V (F4 V (F4 F1 ) = 0.31, F3 ) = 0.5, V (F4 V (F4

F2 ) = 0.41,

F2 ) = 0.14, F5 ) = 1,

F.T.S. Chan, N. Kumar / Omega 35 (2007) 417 431

Table 4 The fuzzy evaluation of the attributes with respect to criterion C3
C3 A8 A9 A10 A11 A8 (1, 1, 1) (2/5, 1/2, 2/3) (2/7, 1/3, 2/5) (2/9, 1/4, 2/7) A9 (3/2, 2, 5/2) (1, 1, 1) (2/7, 1/3, 2/5) (2/7, 1/3, 2/5) A10 (5/2, 3, 7/2) (5/2, 3, 7/2) (1, 1, 1) (2/5, 1/2, 2/3) A11 (7/2, 4, 9/2) (5/2, 3, 7/2) (3/2, 2, 5/2) (1, 1, 1)

427

WC3 0.43 0.23 0.29 0.05

Table 5 The fuzzy evaluation of the attributes with respect to criterion C4

C4 A12 A13 A14 A15 A12 (1, 1, 1) (2/5, 1/2, 2/3) (2/5, 1/2, 2/3) (2/9, 1/4, 2/7) A13 (3/2, 2, 5/2) (1, 1, 1) (2/7, 1/3, 2/5) (2/5, 1/2, 2/3) A14 (3/2, 2, 5/2) (2/5, 1/2, 2/3) (1, 1, 1) (2/5, 1/2, 2/3) A15 (7/2, 4, 9/2) (3/2, 2, 5/2) (3/2, 2, 5/2) (1, 1, 1) WC4 0.61 0.06 0.21 0.12

Table 6 The fuzzy evaluation of the attributes with respect to criterion C5

C5 A16 A17 A18 A19 A16 (1, 1, 1) (2/3, 1, 3/2) (2/3, 1, 3/2) (2/5, 1/2, 2/3) A17 (2/3, 1, 3/2) (1, 1, 1) (2/5, 1/2, 2/3) (2/5, 1/2, 2/3) A18 (2/3, 1, 3/2) (3/2, 2, 5/2) (1, 1, 1) (2/5, 1/2, 2/3) A19 (2/3, 1, 3/2) (3/2, 2, 5/2) (3/2, 2, 5/2) (1, 1, 1) WC5 0.27 0.43 0.30 0.00

Table 7 Summary combination of priority weights: attributes of criterion C1

Weight Alternatives S1 S2 S3 A1 0.58 0.71 0.13 0.16 A2 0.30 0.44 0.36 0.20 A3 0.12 0.69 0.08 0.23 Alternative priority weight 0.63 0.19 0.18

Table 8 Summary combination of priority weights: attributes of criterion C2

Weight Alternatives S1 S2 S3 A4 0.41 0.51 0.23 0.26 A5 0.21 0.51 0.23 0.26 A6 0.26 0.69 0.08 0.23 A7 0.12 0.87 0.00 0.13 Alternative priority weight 0.60 0.16 0.24

Finally the priority weights of each supplier can be calculated by weights per supplier multiplied by weights of the corresponding criterion. The highest score of the

supplier gives the idea about the best global supplier of the manufacturing company for supply of the parts. The results are shown in Table 12.

428

F.T.S. Chan, N. Kumar / Omega 35 (2007) 417 431

Table 9 Summary combination of priority weights: attributes of criterion C3

Weight Alternatives S1 S2 S3 A8 0.43 0.27 0.18 0.55 A9 0.23 0.69 0.08 0.23 A10 0.29 0.05 0.64 0.31 A11 0.05 0.49 0.32 0.19 Alternative priority weight 0.31 0.30 0.39

Table 10 Summary combination of priority weights: attributes of criterion C4

Weight Alternatives S1 S2 S3 A12 0.61 0.83 0.17 0.00 A13 0.06 0.45 0.45 0.10 A14 0.21 0.69 0.08 0.23 A15 0.12 0.33 0.33 0.34 Alternative priority weight 0.72 0.19 0.09

Table 11 Summary combination of priority weights: attributes of criterion C5

Weight Alternatives S1 S2 S3 A16 0.27 0.72 0.00 0.28 A17 0.43 0.49 0.32 0.19 A18 0.30 0.83 0.17 0.00 A19 0.00 0.27 0.18 0.55 Alternative priority weight 0.65 0.19 0.16

Table 12 Summary combination of priority weights: main criteria of the overall objective
Weight Alternatives S1 S2 S3 C1 0.45 0.63 0.19 0.18 C2 0.35 0.60 0.16 0.24 C3 0.09 0.31 0.30 0.39 C4 0.06 0.72 0.19 0.09 C5 0.05 0.65 0.19 0.16 Alternative priority weight 0.60 0.19 0.21

Thus the supplier 1 is the best global supplier and the alternative supplier after this is supplier 3.

According to the nal score, the supplier S1 is the most preferred supplier because it has the highest priority weight and supplier S3 is the next recommended alternative supplier. Through the illustration of the FEAHP model, it is found that the global supplier selection problem can be solved in a structural and simple manner without involving much complexity. The sensitivity of each supplier with respect to the attributes and main criteria can be shown in Figs. 4 and 5. The nal priority weights of each supplier can be seen from

Fig. 6. Because of the space limitation, the step by step computations and fuzzy comparison matrices of all the attributes are not shown. The important results are also shown in Figs. 46. The nal priority weights of different criteria shows that the overall cost of the product carries the highest priority and it is followed by quality of the product, service performance, supplier prole, and risk factor. Under the overall cost of the product the attribute price of the product is generally decided by considering the

F.T.S. Chan, N. Kumar / Omega 35 (2007) 417 431

S1 S2 S3

429

Priority weights

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 A12 A13 A14 A15 A16 A17 A18 A19 Attributes

Fig. 4. Sensitivity of each supplier with respect to the attributes.

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 C1

Priority weights

S1 S2 S3

C2

C3 Criteria

C4

C5

Fig. 5. Sensitivity of each supplier with respect to the criteria.

0.6 0.5 Priority weights 0.4 0.3 0.2 0.1 0 S1 S2 Suppliers S3

transportation of defective products can unnecessarily increase the expenses of the manufacturer. The rejection of the products from customer or from the production line badly affects the performance of the manufacturer in this competitive business environment. Therefore, rejection rate of the product identied as one of the most important attribute under the criterion of quality of the product and carries higher priority weights in this particular study. Apart from this delivery schedule under the service performance, nancial status of the rm under suppliers prole and political stability under the risk factor criterion carries high priority weight and it seems to be practically very relevant in the current political and business environment. The perception and assessment of risk associated with terrorism may be biased and subjective, thus great precision is required to decide its preference over other alternative risk factors.

6. Conclusion and future work In this paper a fuzzy extended AHP (FEAHP) approach has been presented to select the best global supplier for a manufacturing rm to supply one of its most critical parts used in assembling process. The main criteria and attributes have been decided which are based on the current business scenario and experience of the experts in the respective elds. The large number of criteria and attributes demonstrated the complexities being involved in the selection of global supplier. Each factor affecting the supply of the product have been analyzed and discussed. After 9/11 incident the business community became more concerned about the risk factors involved in the business processing. Different risk factors have been considered in making the decision of

Fig. 6. Final priority weights of each supplier.

processing cost, maintenance cost, warranty cost and other cost related to the manufacturing of the product. These factors contribute most of the part of the overall cost of the product as compare to freight cost and tariff and custom duties in this particular study, therefore it has got a higher priority weight. Processing and

430

F.T.S. Chan, N. Kumar / Omega 35 (2007) 417 431 [4] Partovi FY, Burton J, Banerjee A. Application of analytic hierarchy process in operations management. International Journal of Operations and Production Management 1990;10(3):23041. [5] Boer L, Wegen L, Telgen J. Outranking method in support of supplier selection. European Journal of Operational Research 1998;4:10918. [6] Timmerman E. An approach to vendor performance evaluation. Journal of Purchasing and Materials Management 1986;26(4): 28. [7] Gregory RE. Source selection: a matrix approach. Journal of Purchasing and Materials Management 1986;22(2):249. [8] Soukup WR. Supplier selection strategies. Journal of Purchasing and Materials Management 1986;23(2):712. [9] Thompson KN. Supplier prole analysis. Journal of Purchasing and Materials Management 1990;26(1):118. [10] Narasimhan R. An analytical approach to supplier selection. Journal of Purchasing and Materials Management 1983;19(4):2732. [11] Nydick RL, Hill RP. Using the analytical hierarchy process to structure the supplier selection procedure. Journal of Purchasing and Materials Management 1992;28(2):316. [12] Buffa FP, Jackson WM. A goal programming model for purchase planning. Journal of Purchasing and Materials Management 1983;19(3):2734. [13] Sharma D, Benton WC, Srivastava R. competitive strategy and purchasing decisions. Proceedings of the annual national conference of the decision sciences institute, 1989. p. 108890. [14] Weber CA, Ellram LM. Supplier selection using multi-objective programming: a decision support system approach. International Journal of Physical Distribution & Logistics Management 1993;23(2):414. [15] Ghodsypour SH, OBrien C. The total cost of logistics in supplier, under conditions of multiple sourcing, multiple criteria and capacity constraint. International Journal of Production Economics 2001;73:1527. [16] Feng CX, Wang J, Wang JS. An optimization model for concurrent selection of tolerances and suppliers. Computers & Industrial Engineering 2001;40:1533. [17] Dyer JS, Fishburn PC, Steuer RE, Wallenius J, Zionts S. Multiple criteria decision making, multi attribute utility theory: the next ten years. Management Science 1992;38(5):64554. [18] Khorramshahgol R, Azani H, Gousty Y. An integrated approach to project evaluation and selection. IEEE Transactions on Engineering Management 1988;35:26570. [19] Choy KL, Lee WB. On the development of a case based supplier management tool for multinational manufacturers. Measuring Business Excellence 2002;6(1):1522. [20] Verma R, Pullman ME. An analysis of the supplier selection process. Omega, International Journal of Management Science 1998;26(6):73950. [21] Liu J, Ding FY, Lall V. Using data envelopment analysis to compare suppliers for supplier selection and performance improvement. Supply Chain Management: An International Journal 2000;5(3):14350. [22] OBrien C, Ghodsypour SH. A decision support system for supplier selection using an integrated analytic hierarchy process and linear programming. International Journal of Production Economics 1998;56:199212. [23] Wang G, Samuel HH, Dismukes JP. Product-driven supply chain selection using integrated multi-criteria decision-making methodology. International Journal of Production Economics 2004;91:115.

suppliers selection keeping in mind the view of new business scenario emerged. Moreover, the risk factors discussed in this study have not carried much weight at the end, but the paper has made a signicant contribution to identify and emphasize the important criteria and paves the way for future studies to consider these practically relevant and interesting issues like political stability, economy and terrorism in the global supplier selection problems. These risk factors are biased and subjective and it depends on the type of products, area of supply and strategy of the manufacturing rm. The FEAHP model discussed in this paper is proved to be simple, less time taking and having less computational expense as compared to other existing decisionmaking systems [32]. The use of FEAHP does not involve cumbersome mathematical operation and so making it of general use for solving practical multiattribute decision making problems. The FEAHP has the ability to capture the vagueness of human thinking style and effectively solve multi-attribute decision making problems. The illustrative example has demonstrated the thoughtfulness, exibility, and efciency of the proposed model to directly tap the subjectivity and preferences of the decision makers. Despite the aforementioned various advantages of the proposed approach for the global supplier selection, this research work can be extended to add more supplier alternatives which encompass both domestic and international suppliers but this can increase the computational complexities. The multi-objective treatment of global supplier selection and order splitting among the chosen suppliers should also be added in the supplier selection problems. The paper can be extended to incorporate the suppliers capacity constraints and the buyers aggregate quality and service limitations in the supplier selection process. As the policies of supplier partnership varies in different strategies, a exible method which can reect the corporate strategy in the supplier selection process and consider the suppliers and buyers limitations, is required. The global supplier selection decision may include environmental and ethical guidelines set by the manufacturing rm. References
[1] Saaty TL. The analytic hierarchy process. New York: McGrawHill Book Co; 1980. [2] Chang DY. Extent analysis and synthetic decision. Optimization techniques and applications. vol. 1. Singapore: World Scientic; 1992. p. 352. [3] Weber CA, Current JR, Benton WC. Vendor selection criteria and methods. European Journal of Operational Research 1991;50(1):218.

F.T.S. Chan, N. Kumar / Omega 35 (2007) 417 431 [24] Ghodsypour SH, OBrien C. A decision support system for supplier selection using an integrated analytic hierarchy process and linear programming. International Journal of Production Economics 1998;5657:119212. [25] Lee EK, Ha S, Kim SK. Supplier selection and management system considering relationships in supply chain management. IEEE Transactions on Engineering Management 2001;48(3): 30718. [26] Min H. International supplier selection: a multi-attribute utility approach. International Journal of Physical Distribution & Logistics Management 1994;24(5):2433. [27] Van laarhoven PJM, Pedrycz W. A fuzzy extension of saatys priority theory. Fuzzy Sets and Systems 1983;11:22941. [28] Buckley J. Fuzzy hierarchical analysis. Fuzzy Sets and Systems 1985;17:23347. [29] Bortolan G, Degani R. A review of some methods for ranking fuzzy subsets. Fuzzy Sets and Systems 1985;15:119. [30] Mikhailov L. Deriving priorities from fuzzy pairwise comparison judgements. Fuzzy Sets and Systems 2003;134: 36585. [31] Chiou HK, Tzeng GH, Cheng DC. Evaluating sustainable shing development strategies using fuzzy MCDM approach. Omega, International Journal of Management Science 2005;33:22334. [32] Chang DY. Applications of the extent analysis method on fuzzy AHP. European Journal of Operational Research 1996;95: 64955.

431

[33] Kwong CK, Bai H. Determining the importance weights for the customer requirements in QFD using a fuzzy AHP with an extent analysis approach. IIE Transactions 2003;35(7):61926. [34] Kahraman C, Cebeci U, Ulukan Z. Multi criteria supplier selection using fuzzy AHP. Logistics Information Management 2003;16(6):38294. [35] Ching-Hsue C. Evaluating naval tactical missile systems by fuzzy AHP based on the grade value of membership function. European Journal of Operational Research 1997;96(2):34350. [36] Weck M, Klocke F, Schell H, Ruenauver E. Evaluating alternative production cycles using the extended fuzzy AHP method. European Journal of Operational Research 1997;100(2):35166. [37] Cheng CH, Yang KL, Hwang CL. Evaluating attack helicopters by AHP based on linguistic variable weight. European Journal of Operational Research 1999;116(2):42343. [38] Miller GA. The magical number seven plus or minus two: some limitations on our capacity for processing information. Psychological Review 1956;63:8197. [39] Ross TJ. Fuzzy logic with engineering applications. New York: McGraw-Hill Book Co; 1997. [40] Zadeh LA. Fuzzy sets. Information and Control 1965;8: 33853.