Professional Documents
Culture Documents
com
Expert Systems
with Applications
Expert Systems with Applications 34 (2008) 2241–2253
www.elsevier.com/locate/eswa
Abstract
This study presents a strategy-aligned fuzzy simple multiattribute rating technique (SMART) approach for solving the supplier/ven-
dor selection problem from the perspective of strategic management of the supply chain (SC). The majority of supplier rating systems
obtained their optimal solutions without considering firm operations management (OM)/SC strategy. The proposed system utilizes OM/
SC strategy to identify supplier selection criteria. A fuzzy SMART is applied to evaluate the alternative suppliers, and deals with the
ratings of both qualitative and quantitative criteria. The final decision-maker incorporates the supply risks of individual suppliers into
final decision making. Finally, an empirical study is conducted to demonstrate the procedure of the proposed system and identify the
suitable supplier(s).
2007 Elsevier Ltd. All rights reserved.
Keywords: Supplier selection; Simple multiattribute rating technique; Fuzzy set theory; Multiple attributes decision-making; Supply chain management
0957-4174/$ - see front matter 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.eswa.2007.03.001
2242 S.-Y. Chou, Y.-H. Chang / Expert Systems with Applications 34 (2008) 2241–2253
Supplier selection in industry is a cross-functional, perception and judgment that cannot be quantified pre-
group decision-making (GDM) problem, frequently solved cisely. Consequently, supplier selection problems typically
by a non-programmed decision-making process, and in- involve the vagueness inherent in linguistic assessment
volving long-term commitment for a company. A decision and multiple attributes/criteria decision-making (termed
group includes many decision-makers/experts (DMs) such MADM in this study). Approaches employing only exact
as the research and development (R&D), engineering, qual- numerical (crisp) values cannot support the decision-
ity assurance, and financial personnel. In reality, the making procedures required to solve such evaluation
importance (or reliability) of individual DMs may not be problems. Fuzzy set theory (FST), which allows for vague
equal or uniform. The final outcome of the GDM process boundaries, provides a mechanism to utilize fuzziness in
may be significantly influenced by the degree of importance subjective or imprecise determinations of preferences,
of such individual DMs. Pearson and Ellram (1995) found constraints, and goals (Yager, 1982; Kahraman, Ruan, &
that the growing importance of cross-functional team Dogan, 2003).
involvement in supplier selection and evaluation signifi- Harrison and Pelletier (2001) noted that successful stra-
cantly influences overall buying firm performance. There- tegic outcomes are a joint function of managerial attitudes
fore, an effective method of aggregating the various towards the decision-making process and towards the
influences of individual opinions, evaluations, and ratings decision itself. Their study concluded that the best type
from multiple DMs must be identified and applied to sup- of decision-making is characterized by attainable strategic
plier selection problems. objectives pursued through a judgmental rather than a
There are four stages in the purchasing and supply computational process.
literature, namely, defining the problem, formulation of This study is interested in the supplier selection prob-
criteria, qualification, and final selection. Defining the lem in a SC. The proposed solution methodology aims to
problem in the decision-making processes is the first step achieve the following characteristics.
in the method that supports the DMs in carefully question-
ing the need for a decision and identifying available alter- • The formulations of multiple attributes/criteria must
natives. During the criteria formulation stage, the main match firm operations management (OM)/SC strategy.
task for buying firms is assessing the key competitive fac- • The problem should be simplified to save subsequent
tors in their industry and translating these dimensions into implementation planning and evaluation costs while
supplier selection criteria. Strategic management decisions making decisions effectively without incurring significant
influence the relative importance of the various criteria loss of quality in dynamic and uncertain decision-mak-
involved in the supplier selection process (Weber, Current, ing environments.
& Desai, 2000). The choice and the number of criteria to be • DMs are better off pursuing satisfactory outcomes
included in the supplier selection process must be carefully obtained via a judgmental process rather than pursuing
determined to represent the competitive strategies of buy- a maximized outcome obtained through a single compu-
ing firm (Sarkis & Talluri, 2002; Talluri & Narasimhan, tational process.
2003). No publication that treats the stages of problem def- • Expertise, experience, authority, and the responsibilities
inition and criteria formulation can be found for supplier of different DMs regarding specific attributes/
selection processes (de Boer, Labro, & Morlacchi, 2001). criteria and alternative ratings need not be considered
The majority of supplier selection models in the existing lit- uniform.
erature ignore the fact that evaluation criteria must be • The proposed approach must consider both quantitative
aligned with firm strategy. The combined data envelop- and qualitative criteria.
ment analysis (DEA) method proposed by Talluri and • The supplier selection model must consider the supply
Narasimhan (2004) and the framework based on the risks for the individual potential suppliers.
analytic network process (ANP) employed by Sarkis and
Talluri (2002) for discriminating vendor strategic capabili- The supplier rating system (SRS) is the core of the case
ties and performance metrics are somewhat extent relevant based supplier selection module (CSSM), which is primarily
in this strategy-aligned context. In fact, many existing deci- intended to rate supplier performance level using the pre-
sion models only consider quantitative criteria in supplier defined criteria (Choy, Lee, & Lo, 2003). This study thus
selection. Several influences are often ignored in the proposes a pragmatic SRS, strategy-aligned fuzzy simple
decision-making process, including additional qualitative multiattributes rating technique (SMART) approach, for
criteria, incomplete information and imprecise preferences solving the supplier selection problem in a SC. The remain-
(Chen, Lin, & Huang, 2006). Additionally, from a long- der of this paper is organized as follows. Section 2 discusses
term perspective, supplier selection is a decision-making the related literature. Section 3 then describes a review of
problem at the strategic management level that involves a FST. Next, Section 4 introduces the strategy-aligned fuzzy
semi-structured process. The inherent imprecision of the SMART approach in the GDM situation. Subsequently,
relevant information and decision process from such Section 5 illustrates the procedures involved in the proposed
problems is broad, foresight, non-recurring and external. system using an empirical example. Finally, conclusions are
The majority of these attributes are evaluated by human drawn in Section 6.
S.-Y. Chou, Y.-H. Chang / Expert Systems with Applications 34 (2008) 2241–2253 2243
2. Literature review (or SC) based on planning and then identifies five classes
of SCs by varying the position of the decoupling point
Firms achieve missions in three conceptual ways: (1) as follows: Buy-to-Order, Make-to-Order, Assemble-to-
cost leadership, (2) differentiation and (3) quick response Order, Make-to-Stock, and Ship-to-Stock. Furthermore,
(Hambrick & Fredrickson, 2001; Porter, 1980). Functional Naylor, Naim, and Berry (1999) related the decoupling
managers translate these strategic concepts into tangible point directly to the characteristics of the lean and agile
tasks. In defining the link between corporate and func- paradigms. The lean paradigm can be applied to the SC
tional strategies, Buffa (1984) proposed four basic compet- upstream of the decoupling point and the agile paradigm
itive elements of strategy: cost, quality, dependability, and can be applied to the SC downstream of the decoupling
flexibility. Watts, Kim, and Hahn (1992) developed a con- point. Combining agility and leanness in one SC via the
ceptual framework for connecting procurement to firm strategic use of a decoupling point has been termed ‘‘le-
competitive strategy, and to the strategies of other func- agility’’. Lean SC works well where demand is relatively
tions. These strategic competitive elements are prioritized stable and thus predictable and where variety is low. Con-
to satisfy firm level strategies. Depending on the firm strat- versely, in contexts in which demand is volatile and the cus-
egy, functional managers must identify their strategic pri- tomer requirement for variety is high, agility needs to be
orities and the procurement strategy should be consistent significantly higher (Agarwal, Shankar, & Tiwari, 2006).
with the competitive strategy. Vonderembse, Uppal, Huang, and Dismukes (2006) devel-
Monczka and Morgan (1992) pointed out that the inte- oped a framework for classifying supply chain types
gration of procurement strategy with business unit level according to product characteristics and the product life
strategy comprises five main stages: (1) definition of corpo- cycle (PLC) stage. Required SC strategy may change
rate objectives and strategy, (2) procurement strategy according to product type and stage of PLC (Wang,
development, (3) commodities/people strategy develop- Huang, & Dismuke, 2004; Vonderembse et al., 2006).
ment, (4) laying out the linkages and accountabilities and Table 2 summarizes the SC classification based on product
(5) performance measurement. Saunders (1994) proposed type and PLC. Meanwhile, Tables 3 and 4 compare differ-
that it is possible to move beyond the conventional func- ent types of SCs.
tional differentiation between purchasing and manufactur- de Boer et al. (2001) developed a prescriptive framework
ing and consider a more integrated perspective which links for classifying the available models in the literature. The
the two processes to produce a single framework for SC diversity of the buying situation in terms of complexity
strategy. As firms strive to increase their competitiveness and importance of the purchasing practice, including new
by pursuing high quality, reduced cost, product customiza- task purchase, modified re-buy, and straight re-buy, are
tion, and speed-to-market, they increasingly emphasize the shown on one axis of the framework. The other axis covers
SC. The key to effective SCM is to make the suppliers to the four different stages in the supplier evaluation process:
help the firm satisfy ever-changing market demand in the (1) problem definition, (2) criteria formulation, (3) supplier
firm strategy. To ensure that the SC supports the firm’s qualification and (4) final selection. Most the decision-
strategy, it is necessary to consider the SC issues listed in making methods are applied to the qualification and final
Table 1. Just as the OM function supports overall the firm selection stage of the buying process, namely supplier selec-
strategy, the SC is designed to support the OM strategy tion. Thorough supplier selection is essential to partnership
(Heizer & Render, 2004). strategy. Early in the supplier development, the market is
Hockstra and Romme (1992) noted that the decoupling scanned for the alternative suppliers. These suppliers are
point separates the part of organization (or SC) oriented screened based on their R&D and engineering capabilities,
towards customer orders from the part of organization financial stability, production capacity, present R&D
Table 1
Supply chain issues for supporting the firm’s strategy (source: Heizer & Render, 2004)
Low cost strategy Response strategy Differentiation strategy
1. Supplier’s goal Supply demand at lowest possible Respond quickly to changing requirements Share market research; jointly develop
cost and demand to minimize stockouts products and options
2. Primary Select primarily for cost Select primarily for capacity, speed, and Select primarily for product development
selection criteria flexibility skills
3. Process Maintain high average utilization Invest in excess capacity and flexible Modular processes that lend themselves to
characteristics processes mass customization
4. Inventory Minimize inventory throughout Develop responsive system, with buffer Minimize inventory in the chain to avoid
characteristics the chain to hold down costs stocks positioned to ensure supply obsolescence
5. Lead time Shorten lead time as long as it does Invest aggressively to reduce production Invest aggressively to reduce development
characteristics not increase costs lead time lead time
6. Product design Maximize performance and Use product designs that lead to low setup Use modular design to postpone product
characteristics minimize cost. time and rapid production ramp-up differentiation for as long as possible
2244 S.-Y. Chou, Y.-H. Chang / Expert Systems with Applications 34 (2008) 2241–2253
Table 3
Comparison of lean, agile and Leagile supply chains (I) (source: Wang et al., 2004; Vonderembse et al., 2006)
Category Lean SC Agile SC Leagile/hybrid SC
1. Definition A lean SC employs continuous Agility relates to the interface between aA hybrid SC generally involves ‘‘ assemble
improvement to focus on the elimination company and the market. Agile SCs profit to order’’ products where demand can be
of waste or non-value added steps in the by responding to rapidly changing, accurately forecasted. The SC helps to
SC. It is supported by the reduction of continually fragmenting global markets by achieve some degree of customization by
setup times to allow for the economic being dynamic and context-specific, postponing product differentiation until
production of small quantities; thereby aggressively changing, and growth final assembly. Lean or agile SCs are
achieving cost reduction, flexibility and oriented. They are driven by customer ultilized for component production. The
internal responsiveness designed products and services agile part of the chain establishes an
interface to understand and satisfy
customer requirements by being
responsive and innovative
2. Purpose Focus on cost reduction and flexibility for Understands customer requirements by Employ lean production methods
already available products. Employs a interfacing with customers and market manufacturing. Interfaces with the market
continuous improvement process to focus and being adaptable to future changes. to understand customer requirements.
on the elimination of waste or non-value Aims to produce in any volume and Achieve a degree of customization by
added activities across the chain. deliver to a wide variety of market niches postponing product differentiation until
Primarily aims at cost cutting, flexibility simultaneously. Provides customized final assembly and adding innovative
and incremental improvements in products at short lead times components to the existing products
products (responsiveness) by reducing the cost of
variation
3. Approach to Supplier attributes involve low cost and Supplier attributes involve speed, Supplier attributes involve low cost and
choosing suppliers high quality. flexibility, and quality. high quality, along with the capability for
speed and flexibility, as and when required
S.-Y. Chou, Y.-H. Chang / Expert Systems with Applications 34 (2008) 2241–2253 2245
Table 4
Comparison of lean, agile and Leagile supply chains (II) (source: Wang et al., 2004; Vonderembse et al., 2006)
Distinguishing attributes Lean SC Agile SC Leagile/hybrid SC
1. Market demand Predictable/stable Volatile Volatile and unpredictable
2. Product variety Low High Medium
3. PLC Long Short Short
4. Customer drivers Cost Lead time and availability Service level
5. Profit margin Low High Moderate
6. Purchasing policy Buy goods Assign capacity Vendor managed inventory
7. Typical products Commodities Fashion goods Product as per customer demand
8. Quality Market qualifier Market qualifier Market qualifier
9. Cost Market winner Market qualifier Market winner
10. Lead time Market qualifier Market qualifier Market qualifier
11. Service level Market qualifier Market winner Market winner
Table 5
Supplier selection criteria adapted from Ellram (1990)
Criteria Sub-criteria
1. Financial Economic performance, financial stability
2. Performance Delivery, quality, price
3. Technology Manufacturing capability, design capabilities, ability to cope with technology changes, capacity utilization
4. Organizational culture and Top management capability, feeling of trust, internal and external integration of supplier, management attitude/outlook
strategy for the future, compatibility across levels and functions of buyer and supplier, strategic fit between buyer and supplier
5. Other factors Safety record of the supplier, business references, supplier’s customer base
Table 6
Buyer–supplier integration strategies and their important criteria (source: Ghodsypour & O’Brien, 1998)
Levels of integration Important criteria for supplier selection
1. No integration Price and quality
2. Logistical integration Besides the price and quality, the operational logistics elements such as reliability, flexibility, supply lots, lead time
and so on
3. Operational integration The suppliers’ process capability such as set up time, lot size, lead time, etc. should be considered in the supplier
selection process
4. Integrating process and products Besides the criteria of the preceding levels, the supplier’s human resource from several points of view such as design
with supplier involvement, management ability, culture, etc. should be considered
5. Business partnership Besides technological and human resource point of view, the strategic directions of the suppliers should be
considered
technique for order preference by similarity to ideal solu- & Elhag, 2006). No study has yet integrated FST and
tion (TOPSIS); elimination and (et) choice translating real- SMART to solve fuzzy MADM problems in GDM. The
ity (ELECTRE); and four analytical hierarchy processes judgmental decision-making system is necessary and suffi-
(AHPs). The rank-reversal dimension indices in the simula- cient for strategic MADM. Firms encounter supply risk,
tion identified SAW and MEW as the best approaches. whether it is explicitly understood and assessed, or reac-
Additionally, Chang and Yeh (2001) proposed using the tively managed. The approach used by Deng (1999),
SAW method in accordance with the result of empirical namely using uniform rather than the different judgmental
validation for the three methods (the SAW method, the coefficients to compute the total values of individual sup-
weighted product method, the TOPSIS). These findings pliers and then obtain the ranking order, is impractical.
of these studies indicate that simpler evaluation techniques This study proposes a simple and pragmatic fuzzy SMART
do not necessarily have worse performance. The SMART with a strategy-aligned framework and a judgmental deci-
proposed in this study is a modified SAW and thus inherits sion-making process to solve supplier selection problems
its simplicity. in the GDM situation. The proposed system has several
As indicated by the literature review, the FST is incor- advantages over traditional methods used to solve strategic
porated into many concepts and procedures when enhanc- supplier selection problem. First, the proposed system
ing their capability to handle MADM problems in fuzzy obtains the optimal solution by considering its fit with firm
environments (Chen et al., 2006; Chang & Wang, 2006; strategy. This intention is achieved by linking the criteria of
Kuo & Chen, 2004; Lee, Lau, Liu, & Tam, 2001; Wang supplier selection and firm OM/SC strategy. Additionally,
2246 S.-Y. Chou, Y.-H. Chang / Expert Systems with Applications 34 (2008) 2241–2253
(1) definition of corporate objectives and strategy, (2) OM/ e jp , denoted as dð~
weights. The defuzzification of W wjp Þ, is
SC strategy development and (3) identification of supplier therefore given by
selection criteria. In the final stage, the DMs identify the 1
important supplier selection criteria according to OM/SC wjp Þ ¼ ðajp þ bjp þ cjp Þ;
dð~ j ¼ 1; 2; . . . ; n; p ¼ 1; 2; . . . ; u:
3
strategy, buyer–supplier integration strategy, and contin-
ð3Þ
gency situation, etc.
Assume that there is a committee of k DMs (Dt, t = The crisp value of normalized weight for sub-criterion Cjp,
1, 2, . . . , k) who are responsible for assessing the importance denoted as Wjp, is given by
weights of the u sub-criteria and n criteria (Cj, j =
1, 2, . . . , n) as well as the scores of m alternatives (Ai, i = dð~wjp Þ
W jp ¼ Pu ; j ¼ 1; 2; . . . ; n; p ¼ 1; 2; . . . u; ð4Þ
1, 2, . . . , m) under each of the u sub-criteria (Cjp, j = 1, wjp Þ
p¼1 dð~
2, . . . , n, p = 1, 2,. . . , u). Pu
where p¼1 W jp ¼ 1; j ¼ 1; 2; . . . n.
4.2. Assessment and aggregation of the fuzzy weights e j , denoted as dð~
The defuzzification of W wj Þ, is therefore
given by
Introduce linguistic weighting variables for DMs to
assess sub-criteria and criteria importance, and then com- 1
wj Þ ¼ ðaj þ bj þ cj Þ;
dð~ j ¼ 1; 2; . . . ; n: ð5Þ
pute aggregated fuzzy weight of sub-criterion and each 3
criterion. The crisp value of normalized weight for criterion Cj, de-
noted as Wj, is given by
4.2.1. Fuzzy sub-criteria weights
Let It, t = 1, 2,. . . , k, where It 2 I and It 2 [1, 10], be the dð~wj Þ
e jpt ¼ W j ¼ Pn ; j ¼ 1; 2; . . . ; n; ð6Þ
weight of importance given to individual DMs Dt, W wj Þ
j¼1 dð~
ðajpt ; bjpt ; cjpt Þ; j ¼ 1; 2; . . . ; n; p ¼ 1; 2; . . . ; u; t ¼ 1; 2; . . . ; P
k; be the linguistic weight given to sub-criteria Cjp, where nj¼1 W j ¼ 1. The weight vector W = [w1, w2, . . . , wn]
j = 1, 2,. . . , n, p = 1, 2,. . . , u, by DM Dt. The aggregated can therefore be formed.
fuzzy sub-criterion weight with respect to individual crite-
ria Cj, W e jp ¼ ðajp ; bjp ; cjp Þ; j ¼ 1; 2; . . . ; n; p ¼ 1; 2; . . . u, 4.4. Computation of the aggregated fuzzy ratings of criteria
of sub-criteria Cjp assessed by the committee of k DMs is
defined as: Uses linguistic rating variables for DMs to assess fuzzy
.X k (or crisp) ratings of alternatives with respect to individual
We jp ¼ ðI 1 W e jp1 I 2 W e jp2 I k W jpk Þ I t; sub-criteria, and then compute aggregated fuzzy ratings
t¼1 and construct the fuzzy rating matrix.
ð1Þ
Pk Pk Pk Pk 4.4.1. Aggregation of fuzzy ratings of individual sub-criteria
where ajp ¼ t¼1 I t ajpt = t¼1 I t ; bjp ¼ t¼1 I t bjpt = t¼1 I t ;
Pk P Let ~xijpt ¼ ðaijpt ; bijpt ; cijpt Þ; i ¼ 1; 2; . . . ; m, j = 1,2,. . . , h,
cjp ¼ t¼1 I t cjpt = kt¼1 I t .
p = 1,2,. . .,u, t = 1,2,. . .,k, be the linguistic suitability
If I 1 ¼ I 2 ¼ ¼ I k , the importance of individual DMs’
rating assigned to alternative supplier Ai for qualitative/
opinions influence the outcome of decision-making are
subjective sub-criteria Cjp by DM Dt. Let ~vijpt ¼ ðd ijpt ;
equal.
eijpt ; fijpt Þ, i = 1, 2,. . . , m, j = h + 1, h + 2,. . . , n, p = 1, 2,
. . . , u, t = 1, 2,. . . , k, be the fuzzy (or crisp) cost or benefit
4.2.2. Fuzzy criteria weights
e jt ¼ ðajt ; bjt ; cjt Þ; j ¼ 1; 2; . . . ; n; t ¼ 1; 2; . . . ; k; be assessed to alternative supplier Ai for quantitative/objec-
Let W
tive sub-criteria Cjp by DM Dt. Eqs. (7) and (8) shown in
the linguistic weight given to criteria Cj = 1,2, . . . , n, by
ej ¼ the following are applied to normalize the quantitative
DM Dt. The aggregated fuzzy criterion weight, W
value.
ðaj ; bj ; cj Þ; j ¼ 1; 2; . . . ; n; of criteria Cj assessed by the
committee of k DMs is defined as: ~vijpt mini fd ijpt g
~xijpt ¼ 100; ð7Þ
.X k maxi ffijpt g mini fd ijpt g
We j ¼ ðI 1 W e l1 I 2 W e j2 I k W jk Þ I t ; ð2Þ where ~xijpt denotes the normalized fuzzy rating of fuzzy
t¼1
P P P P benefit ~vijpt .
where
Pk aj P¼ kt¼1 I t ajt = kt¼1 I t ; bj ¼ kt¼1 I t bjt = kt¼1 I t ; cj ¼ maxi ffijpt g ~vijpt
k
t¼1 I t cjt = t¼1 I t .
~xijpt ¼ 100; ð8Þ
maxi ffijpt g mini fd ijpt g
4.3. Defuzzification of the fuzzy weights of criteria where ~xijpt denotes the normalized fuzzy rating of fuzzy cost
~vijpt .
The simple and popular method, centroid method, is Let us further defined ~xijp as the aggregated fuzzy rating
adopted to defuzzify the fuzzy sub-criteria and criteria of alternative Ai for sub-criterion Cjp, such that
2248 S.-Y. Chou, Y.-H. Chang / Expert Systems with Applications 34 (2008) 2241–2253
,
X
k Step 3: Compute the normalized values of individual
~xijp ¼ ðI 1 ~xijp1 I 2 ~xijp2 I k ~xijpk Þ I l; ð9Þ quantitative sub-criteria and the aggregated fuzzy
l¼1
rating of each qualitative sub-criterion.
which can subsequently be represented and computed as Step 4: Compute the aggregated fuzzy rating of each sup-
~xijp ¼ ðaijp ; bijp ; cijp Þ; i ¼ 1; 2; . . . ; m; j ¼ 1; 2; . . . ; n; plier-criterion combination, and then, form the
fuzzy rating matrix.
p ¼ 1; 2; . . . ; u;
P P P P Step 5: Compute the total fuzzy values of individual sup-
where aijp ¼ kt¼1 I t aijpt = kt¼1 I t ; bijp ¼ kt¼1 I t bijpt = kt¼1 I t ; pliers, and then, rank order of suppliers based on
P P
cijp ¼ kt¼1 I t cijpt = kt¼1 I t : the maximum TCV.
• Because of applying the lean SC, the supplier • Strategic fit (C42): Strategic fit represents the fit
must be capable of providing low cost and high between firm strategy and supplier strategy as
quality. judged by the DMs.
• Low cost is imperative, because cost is critical (5) Technical capacity (C5)
for a market winner in a SC. • Innovation (C51): The DMs impress on supplier
• Supplier organizational culture and strategy ability to develop new products or techniques
should be considered to select the desired sup- and improve existing products.
plier, especially given their contribution to the • Technical problem-solving (C52): A supplier
overall cost and quality appeal of products in spends how many days solving unexpected prob-
long-term continuous improvement. lems happening in Company G, such as post-sale
After comprehensive discussion based on the above profile, support.
all the selection criteria for USB cables supplier are identi- The committee then continues the sequent procedures to
fied accordingly. The five DMs collectively set up five crite- identify the best alternative supplier(s) as the business
ria and establish two critical sub-criteria for each criterion. partner(s).
The detail descriptions of criteria and sub-criteria are listed Step 2: After applying the compromise mechanism of the
below: committee discussion, the importance weights of
(1) Cost (C1) individual DMs It, (I1 = 5, I2 = 3, I3 = 2, I4 = 6,
• Unit price (C11): The percent discount a supplier I5 = 4), are assigned by the final DM. Linguistic
can provide for a given unit the relative market price weighting variables (Table 7) and their respective
at a particular time. Here, purchasing price, trans- fuzzy numbers (Fig. 1) for DMs are then used to
portation cost and taxes are considered together. assess the importance weights of the sub-criteria
• Cost reduction plan (C12): The percentage dis- (Table 8) and criteria (Table 9). Based on the
count margins that a supplier can provide for a assessment values, the fuzzy weights of individual
given unit on an annual basis. sub-criteria and criteria can then be computed by
(2) Quality (C2) using Eqs. (1), (2), (4) and (6), respectively (Tables
• Interval rejection rate (C21): A supplier checks the 8 and 9).
ratio of defective units when inspecting the USB Step 3: For quantitative sub-criteria, company G collects
cables. real historical data that has not been assessed by
• Customer rejection rate (C22): Company G checks the DM regarding the three suppliers in Table 10
the ratio of defective units when detecting the through short-term collaboration. In accordance
USB cables. with real historical data, the normalized values
(3) Delivery (C3) of individual quantitative sub-criteria can be
• Lead time (C31): The number of days from order computed by using Eqs. (7) or (8) (Table 10).
placement to the receipt of the goods ordered. Meanwhile, the linguistic rating variables (Table
• Flexibility (C32): A supplier can fulfill how many 7) and their respective fuzzy numbers for DMs
percentages of USB cables company G unexpect- are used to assess the fuzzy ratings of the
edly request. three alternatives with respect to each qualitative
(4) Organizational culture and strategy (C4) sub-criterion, and then compute the aggregated
• Management capability (C41): The DMs impress fuzzy rating of each qualitative sub-criterion can
on the management capability image of a supplier be computed by Eq. (9). Table 11 lists the
according to mutual interaction. results.
Table 8
Linguistic and aggregated importance weights of the sub-criteria
Sub-criteria DMs’ linguistic weights Aggregated weights
D1 D2 D3 D4 D5 Fuzzy Defuzzified Normalized
C11 VH H VH H VH (8.1, 9.55, 10) 9.22 0.49
C12 VH VH H VH VH (8.8, 9.9, 10) 9.57 0.51
C21 MH H M ML MH (3.9, 5.9, 7.75) 5.85 0.39
C22 VH H H VH H (8.1, 9.55, 10) 9.22 0.61
C31 VH M MH MH H (6.1, 7.85, 9.15) 7.70 0.49
C32 MH H H MH MH (6.6, 8.25, 9.45) 8.10 0.51
C41 MH VH H MH VH (6.1, 8.25, 9.45) 8.10 0.57
C42 MH H MH ML MH (4.1, 6.1, 7.95) 6.05 0.43
C51 MH ML M MH M (3.8, 5.8, 7.8) 5.80 0.47
C52 M MH MH H M (4.7, 6.7, 8.4) 6.60 0.53
2250 S.-Y. Chou, Y.-H. Chang / Expert Systems with Applications 34 (2008) 2241–2253
Table 9
Linguistic and aggregated importance weights of the criteria
Criteria DMs’ linguistic weights Aggregated weights
D1 D2 D3 D4 D5 Fuzzy Defuzzified Normalized
C1 VH VH H H VH (8.2, 9.6, 10) 9.27 0.25
C2 H H MH H MH (6.4, 8.4, 9.7) 8.17 0.22
C3 H MH H MH MH (5.7, 7.7, 9.35) 7.58 0.20
C4 MH H H MH MH (5.5, 7.5, 9.25) 7.42 0.20
C5 ML MH M M ML (2.6, 4.6, 6.6) 4.60 0.13
Table 10
Real data and normalized fuzzy rating (NFR) of each supplier with respect to all quantitative criteria
Sub-criteria Suppliers Real data (crisp values) Benefit/cost NFRs
C11 A1 90% Cost (61.9, 61.9, 61.9)
A2 96.5% (0, 0, 0)
A3 86% (100, 100, 100)
C12 A1 3.5% Benefit (40, 40, 40)
A2 2.5% (0, 0, 0)
A3 5% (100, 100, 100)
C21 A1 0.45% Cost (100, 100, 100)
A2 0.61% (42.86, 42.86, 42.86)
A3 0.73% (0, 0, 0)
C22 A1 0.26% Cost (57.89, 57.89, 57.89)
A2 0.18% (100, 100, 100)
A3 0.37% (0, 0, 0)
C31 A1 7 Cost (0, 0, 0)
A2 3 (100, 100, 100)
A3 4 (75, 75, 75)
C32 A1 56% Benefit (0, 0, 0)
A2 68% (57.14, 57.14, 57.14)
A3 77% (100, 100, 100)
C52 A1 1 Cost (0, 0, 0)
A2 1 (0, 0, 0)
A3 0.5 (100, 100, 100)
Table 11
Linguistic and aggregated fuzzy ratings of sub-criteria
Sub-criteria Suppliers DMs’ linguistic ratings Aggregated fuzzy ratings
D1 D2 D3 D4 D5
C41 A1 G G VG G G (72, 91, 100)
A2 MG G G G G (65, 85, 97.5)
A3 VG MG G MG VG (70, 85.5, 95.5)
C42 A1 MG G MG G MG (59, 79, 94.5)
A2 G MG MG MG G (59, 79, 94.5)
A3 G MG G VG MG (69, 86, 96.5)
C51 A1 F MG G G MG (53, 73, 89)
A2 MP F F F MG (29, 49, 69)
A3 VG G G M VG (73, 88.5, 97)
Step 4: Compute the aggregated fuzzy rating of each sup- individual TCV by Eq. (12). Suppliers are existing
plier-criterion combination using Eq. (10), and vendor of the company G. Based on the track
then, form the fuzzy rating matrix (Table 12). records, the final DM identifies the risk coefficients
Step 5: Compute the total fuzzy values of individual sup- a1, a2, a3 as 0.2, 0.1, 0.15, respectively. The ranking
pliers by Eq. (11) (Table 13). Then compute the order for the three alternative suppliers is A3, A2,
S.-Y. Chou, Y.-H. Chang / Expert Systems with Applications 34 (2008) 2241–2253 2251
Table 12
Fuzzy rating matrix
Suppliers Criteria
C1 C2 C3 C4 C5
A1 (50.73, 50.73, 50.73) (74.31, 74.31, 74.31) (0, 0, 0) (70.4, 85.84, 97.64) (24.91, 34.31, 38.27)
A2 (0, 0, 0) (77.72, 77.72, 77.72) (78.14, 78.14, 78.14) (61.56, 82.42, 96.21) (12.47, 21.07, 29.67)
A3 (100, 100, 100) (0, 0, 0) (87.75, 87.75, 87.75) (69.57, 85.72, 95.93) (87.31, 94.6, 98.59)
(SMART using swings) or SMARTER (SMART extended (7) Subtraction of any realk and a fuzzy number x
to ranking) suitable for solving the supplier selection e . k ¼ ða1 k; a2 k; a3 kÞ;
A
problem.
k.Ae ¼ ðk a3 ; k a2 ; k a1 Þ:
Appendix A
Two multiplication commutative properties are also estab-
A.1. Triangular fuzzy number (TFN)
lished to facilitate the computation in the proposed system:
A TFN Ae is described as any fuzzy subset of the real line
eB
A e¼B e
e A;
R with membership function leðxÞ that possesses the fol-
A
lowing basic features (Dubois & Prade, 1978; Keufmann e¼B
kB e k:
& Gupta, 1991):
The proofs of these two properties are straight-forward
(1) lA~ ðxÞ is a continuous mapping from R to the closed and thus omitted.
interval [0, 1],
(2) lA~ ðxÞ ¼ 0; for all x 2 [1, a], References
(3) lA~ ðxÞ is strictly increasing on [a, b],
(4) lA~ ðxÞ ¼ 1; for x = b, Agarwal, A., Shankar, R., & Tiwari, M. K. (2006). Modeling the metrics
of lean, agile and leagile supply chain: an ANP-based approach.
(5) lA~ ðxÞ is strictly decreasing on [b, c],
European Journal of Operational Research, 173, 211–225.
(6) lA~ ðxÞ ¼ 0; for all x 2 [c,1], Buffa, E. S. (1984). Meeting the competitive challenge. Illinois: Homewood.
Cakravastia, A., & Takahashi, K. (2004). Integrated model for supplier
where a, b, c are real numbers. A TFN Ae can be defined selection and negotiation in a make-to -order environment. Interna-
as a triplet (a, b, c), and the membership function lA~ ðxÞ is tional Journal of Production Research, 42(21), 4457–4474.
Chan, F. T. S. (2003). Interactive selection model for supplier selection
defined as:
process an AHP. International Journal of Production Research, 41,
8
> ðx aÞ 3549–3579.
>
> ; a 6 x 6 b; Chang, P. C., & Wang, Y. W. (2006). Fuzzy Delphi and back-propagation
>
> ðb aÞ
>
< model for sales forecasting in PCB industry. Expert Systems with
lA~ ðxÞ ¼ ðc xÞ Applications, 30(4), 715–726.
>
> ; b 6 x 6 c; Chang, Y. H., & Yeh, C. H. (2001). Evaluating airline competitiveness
> ðc bÞ
>
>
> using multiattribute decision-making. Omega, 29, 405–415.
: Chen, C. T. (2001). A fuzzy approach to select the location of the
0; otherwise:
distribution center. Fuzzy Sets and Systems, 118, 65–73.
Chen, S. J., & Hwang, C. L. (1992). Fuzzy multiple attribute decision-
making method and applications. Berlin, Heidelberg and New York:
A.2. Fuzzy operations for TFNs
Springer-Verlag.
Chen, C. T., Lin, C. T., & Huang, S. F. (2006). A fuzzy approach for
Given two TFNs A e ¼ ða1 ; a2 ; a3 Þ and B
e ¼ ðb1 ; b2 ; b3 Þ, supplier evaluation and selection in supply chain management.
seven primary operations of these two TFNs can be International Journal of Production Economics, 102, 289–301.
expressed as (Chen & Hwang, 1992; Chiou et al., 2005): Chiou, H. K., Tzeng, G. H., & Cheng, D. C. (2005). Evaluating
sustainable fishing development strategies using fuzzy MCDM
approach. Omega, 33, 223–234.
(1) Addition of two fuzzy numbers Choi, T. Y., & Hartley, J. L. (1996). An exploration of supplier selection
eB
e ¼ ða1 þ b1 ; a2 þ b2 ; a3 þ b3 Þ; practices across the supply chain. Journal of Operations Management,
A
14, 333–343.
(2) Multiplication of two fuzzy numbers Choy, K. L., Lee, W. B., & Lo, V. (2003). Design of a case based
intelligent supplier relationship management- the integration of
eB
A e ¼ ða1 b1 ; a2 b2 ; a3 b3 Þ; supplier rating system and product coding system. Expert Systems
with Applications, 25, 87–100.
(3) Multiplication of any real number k and a fuzzy de Boer, L., Labro, E., & Morlacchi, P. (2001). A review of methods
number supporting supplier selection. European Journal of Purchasing and
Supply Management, 7, 75–89.
e ¼ ðkb1 ; kb2 ; kb3 Þ;
kB Deng, H. (1999). Multicriteria analysis with fuzzy pairwise comparison.
International Journal of Approximate Reasoning, 21, 215–231.
(4) Division of two fuzzy numbers D Dickson, G. W. (1966). An analysis of vendor selection system and
e Be ¼ ða1 =b3 ; a2 =b2 ; a3 =b1 Þ; decisions. Journal of Purchasing, 2(1), 28–41.
AD Dubois, D., & Prade, H. (1978). Operations on fuzzy numbers. Interna-
(5) Division by any real number kD tional Journal of Systems Science, 9, 613–626.
Edwards, W. (1977). How to multiattribute utility measurement for social
e
ADk ¼ ða1 =k; a2 =k; a3 =kÞ; decision-making. IEEE Transactions on Systems, Man, and Cybernet-
ics, SMC-7, 326–340.
(6) Subtraction of two fuzzy numbers x Edwards, W., & Barron, F. H. (1994). SMARTS and SMARTER:
improved simple methods for multiattribute utility measurement.
e.B
A e ¼ ða1 b3 ; a2 b2 ; a3 b1 Þ; Organizational Behavior and Human Decision Processes, 60(3), 306–325.
S.-Y. Chou, Y.-H. Chang / Expert Systems with Applications 34 (2008) 2241–2253 2253
Ellram, L. M. (1990). The supplier selection decision in strategic partner- Porter, M. E., & Millar, V. E. (1985). How information gives you
ships. Journal of Purchasing and Material Management, 26(4), 8–12. competitive advantage. Harvard Business Review, 63(4), 149–160.
Ghobadian, A., Stainer, A., & Kiss, T. (1993). A computerized vendor Sarkis, J., & Talluri, S. (2002). A model for strategic supplier selection.
rating system. In Proceedings of the First International Symposium on Journal of Supply Chain Management, 38(1), 18–28.
Logistics (pp. 321–328). Nottingham, UK: The University of Saunders, M. (1994). Strategic purchasing and supply chain management.
Nottingham. London: Pitman Publishing.
Ghodsypour, S. H., & O’Brien, C. (1998). A decision support system for Simchi-Levi, D., Kaminsky, P., & Simchi-Levi, E. (2003). Designing and
supplier selection using an integrated analytic hierarchy process and managing the supply chain: concepts, strategies, and case studies. New
linear programming. International Journal of Production Economics., York: McGraw-Hill.
56–57(1–3), 199–212. Stevenson, W. J. (2005). Operations management. New York: McGraw-Hill.
Giunipero, L. C., & Brand, R. R. (1996). Purchasing’s role in supply chain Talluri, S., & Narasimhan, R. (2003). Vendor evaluation with perfor-
management. International Journal of Logistics management, 7(1), mance variability: a max-min approach. European Journal of Opera-
29–38. tional Research, 146(3), 543–552.
Hambrick, D. C., & Fredrickson, J. W. (2001). Are you sure you have a Talluri, S., & Narasimhan, R. (2004). A methodology for strategic
strategy? Academy of Management Executive, 15(4), 48–59. sourcing. European Journal of Operational Research, 154(1), 236–
Harrison, E. F., & Pelletier, M. A. (2001). Revisiting strategic decision 250.
success. Management Decision, 33(3), 169–179. Tang, M. T., Tzeng, G. H., & Wang, S. W. (1999). A hierarchy fuzzy
Heizer, J., & Render, B. (2004). Principles of operations management. New MCDM method for studying electronic marketing strategies in the
Jersey: Prentice Hall. information service industry. Journal of International Information
Hockstra, S., & Romme, J. (1992). Integral logistics structures: developing Management, 8(1), 1–22.
customer oriented goods flow. London: McGraw-Hill. Tsaur, S. H., Tzeng, G. H., & Wang, G. C. (1997). Evaluating tourist risks
Hong, G. H., Park, S. C., Jang, D. S., & Rho, H. M. (2005). An effective from fuzzy perspectives. Annals of Tourism Research, 24(4), 796–812.
supplier selection method for constructing a competitive supply- van Weele, A. J. (2000). Purchasing and supply chain management.
relationship. Expert Systems with Applications, 28, 629–639. London: Thomson Learning.
Kahraman, C., Ruan, D., & Dogan, I. (2003). Fuzzy group decision- Vonderembse, M. A., Uppal, M., Huang, S. H., & Dismukes, J. P. (2006).
making for facility location selection. Information Sciences, 157, Designing supply chains: towards theory development. International
135–153. Journal of Production Economics, 100, 223–238.
Keufmann, A., & Gupta, M. M. (1991). Introduction to fuzzy arithmetic: Wang, Y. M., & Elhag, T. M. S. (2006). Fuzzy TOPSIS method based on
theory and application. New York: Van Nostrand Reinhold. alpha level sets with an application to bridge risk assessment. Expert
Klic, G. I., & Yan, B. (1995). Fuzzy sets and fuzzy logic theory and Systems with Applications, 31(2), 309–319.
applications. London: Prentice-Hall. Wang, G., Huang, S. H., & Dismuke, J. P. (2004). Product-driven supply
Kumar, M., Vrat, P., & Shankar, R. (2004). A fuzzy goal programming chain selection using integrated multicriteria decision-making meth-
approach for vendor selection problem in a supply chain. Computer odology. International Journal of Production Economics, 91, 1–15.
and Industrial Engineering, 46, 69–85. Watts, C. A., Kim, K. Y., & Hahn, C. K. (1992). Linking purchasing to
Kuo, R. J., & Chen, J. A. (2004). A decision support system for order corporate competitive strategy. International Journal of Purchasing and
selection in electronic commerce based on fuzzy neural network Materials Management, 4–7.
supported by real-coded genetic algorithm. Expert Systems with Weber, C. A., Current, J. R., & Benton, W. C. (1991). Vendor selection
Applications, 26(2), 141–154. criteria and methods. European Journal of Operational Research, 50(1),
Lee, W. B., Lau, H., Liu, Z. Z., & Tam, S. (2001). A fuzzy analytic 2–18.
hierarchy process approach in modular product design. Expert Weber, C. A., Current, J. R., & Desai, A. (2000). An optimization
Systems, 18(1), 32–42. approach to determining the number of vendors to employ. Supply
Liang, G. S. (1999). Fuzzy MCDM based on ideal and anti-ideal concepts. Chain Management: An International Journal, 5(2), 90–98.
European Journal of Operational Research, 112, 682–691. Yager, R. R. (1982). Fuzzy sets and possibility theory. Oxford: Pergamon
Miller, G. A. (1965). The magic number seven, plus or minus seven. Press.
Psychological Review, 63, 81–97. Yao, J. S., & Chiang, J. (2003). Inventory without backorder with fuzzy
Monczka, J. P., & Morgan, J. P. (1992). Strategic sourcing management. total cost and fuzzy storing cost defuzzified by centroid and signed
Purchasing, 13, 70. distance. European Journal of Operational Research, 148, 401–
Naylor, J. B., Naim, M. M., & Berry, D. (1999). Leagility: Integrating the 409.
lean and agile manufacturing paradigms in the total supply chain. Yetton, P., & Botter, P. (1983). The relationships among group size,
International Journal of Production Economics, 62, 107–118. member ability, social decision schemes, and performance. Organiza-
Pearson, J. N., & Ellram, L. M. (1995). Supplier selection and evaluation tional Behavior and Human Performance(October), 145–159.
in small versus large electronics firm. Journal of Small Business Zadeh, L. A. (1965). Fuzzy sets. Information and Control, 8, 338–353.
Management, 33(4), 53–65. Zanakis, S. H., Solomon, A., Wishart, N., & Dublish, S. (1998).
Porter, M. E. (1980). Competitive strategy: techniques for analyzing Multiattribute decision-making: A simulation comparison of select
industries and competitors. New York: The Free Press. methods. European Journal of Operational Research, 107, 507–529.