Expert Systems with Applications 39 (2012) 3000–3011

Contents lists available at SciVerse ScienceDirect

Expert Systems with Applications

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

A novel hybrid MCDM approach based on fuzzy DEMATEL, fuzzy ANP and fuzzy
TOPSIS to evaluate green suppliers
Gülçin Büyüközkan ⇑, Gizem Çifçi
_
Industrial Engineering Department, Faculty of Engineering and Technology, Galatasaray University, 34357 Ortaköy, Istanbul, Turkey

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

Keywords: It is well known that ‘‘green’’ principles and strategies have become vital for companies as the public
Green supply chain awareness increased against their environmental impacts. A company’s environmental performance is
Supplier selection not only related to the company’s inner environmental efforts, but also it is affected by the suppliers’
Fuzzy ANP environmental performance and image. For industries, environmentally responsible manufacturing,
Fuzzy DEMATEL
return flows, and related processes require green supply chain (GSC) and accompanying suppliers with
Fuzzy TOPSIS
environmental/green competencies. During recent years, how to determine suitable and green suppliers
in the supply chain has become a key strategic consideration. Therefore this paper examines GSC man-
agement (GSCM) and GSCM capability dimensions to propose an evaluation framework for green suppli-
ers. However, the nature of supplier selection is a complex multi-criteria problem including both
quantitative and qualitative factors which may be in conflict and may also be uncertain. The identified
components are integrated into a novel hybrid fuzzy multiple criteria decision making (MCDM) model
combines the fuzzy Decision Making Trial and Evaluation Laboratory Model (DEMATEL), the Analytical
Network Process (ANP), and Technique for Order Performance by Similarity to Ideal Solution (TOPSIS)
in a fuzzy context. A case study is proposed for green supplier evaluation in a specific company, namely
Ford Otosan.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction improving quality and flexibility to meet the needs of the custom-
ers, etc. For this reason, the aim of this study is to propose an
Money, components, processes and information flows might evaluation model to judge the appropriateness of suppliers for an
establish a supply chain management system but simultaneously, organization which has environmental goals and measure the
due to government legislation and increasing awareness among validity of the model with a real case study.
the people to protect the environment; firms today cannot ignore There are various mathematical techniques for evaluation of
environmental issues if they want to survive in the global market. suppliers, such as data envelopment analysis (DEA) (Wu, 2009),
In this sense, green supply chain management (GSCM) has emerged heuristics (He, Chaudhry, Lei, & Baohua, 2009; Sen, Basßligil, S
ß en, &
as a way for firms to achieve profit and market share objectives by Baraçli, 2007), analytic hierarchy process (AHP) (Sevkli, Koh, Zaim,
lowering environmental impacts and increasing ecological effi- Demirbag, & Tatoglu, 2007), fuzzy AHP (Chan & Kumar, 2007; Lee
ciency (van Hock & Erasmus, 2000). In response to demands, com- et al., 2009; Rao & Holt, 2005), fuzzy goal programming (Kumar,
panies have to find ways to incorporate environmental and social Vrat, & Shankar, 2006; Tsai & Hung, 2009), fuzzy analytic network
aspects into their supply chain management. process (ANP) (Lin, 2009; Tuzkaya & Önüt, 2008) in literature. For
In order to reap the greatest benefits from environmental man- the purpose of evaluating and selecting green suppliers, both qual-
agement, firms must integrate all members in the green supply itative and quantitative factors must be considered. Thus, green
chain (GSC) (Lee, Kang, Hsu, & Hung, 2009). Among these expecta- supplier selection is a kind of multiple criteria decision making
tions, increasing attention is devoted to suppliers’ social responsi- (MCDM) problem and we need to employ MCDM methods to han-
bility with a particular focus on fair and legal use of natural dle it appropriately. Here emphasis is placed on the relationships of
resources. Hence, strategic partnership with environmentally, factors which can be handled by ANP (Saaty, 1996) effectively. The
socially and economically powerful suppliers should be integrated ANP can deal with the dependence in feedback systematically. In
within the GSC for improving the performance in many directions this study also Decision Making Trial and Evaluation Laboratory
including reducing costs and lead time, eliminating wastages, (DEMATEL) method (Gabus & Fontela, 1972) is used to extract the
mutual relationships of interdependencies within criteria and the
⇑ Corresponding author. Tel.: +90 212 227 4480. strength of interdependence. Lastly to choose the alternative for
E-mail address: [email protected] (G. Büyüközkan). ideal solution of this problem, Technique for Order Performance

0957-4174/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.eswa.2011.08.162
 G. Büyüközkan, G. Çifçi / Expert Systems with Applications 39 (2012) 3000–3011 3001

by Similarity to Ideal Solution (TOPSIS) is used. However it should The past few years have led researchers to investigate the envi-
not be ignored that the fuzzy nature of human life makes these ronmental concepts in management and supply chains. Lu, Wu,
kinds of MCDM analysis more difficult. Yet for human being’s sub- and Kuo (2007) proposed environmental principles applicable to
jective judgment, a theory needed in measuring the ambiguity of green supplier evaluation by using multi-objective decision analy-
these concepts. Therefore, fuzzy logic (Zadeh, 1965) is used in eval- sis. According to current environmental regulations, companies’
uations that allows for uncertainty among factors. environmental policies, and nongovernmental organizations’ envi-
Briefly, fuzzy DEMATEL (Chen, Tseng, & Lin, 2008; Tseng, 2009a; ronmental guidelines; the main environmental criteria were deter-
Wu & Lee, 2007); fuzzy ANP (Liu & Lai, 2009; Tuzkaya, Ozgen, mined as materials, energy use, solid residue, liquid residue,
Ozgen, & Tuzkaya, 2009; Yüksel & Dağdeviren, 2010); and fuzzy gaseous residue. And this framework was evaluated using a fuzzy
TOPSIS (Salehi & Tavakkoli-Moghaddam, 2008; Yong, 2006; Iç _ & AHP methodology. Ozgen, Önüt, Gülsün, Tuzkaya, and Tuzkaya
Yurdakul, 2010) approaches used by several authors are workable. (2008) presented a two-phase possibilistic linear programming
Because by applying these theories, it can be easy to discover methodology for multi-objective supplier evaluation and order
things inside the complex problem. In the literature there are some allocation problems. The required dimensions for evaluating sup-
works on these methods, but there is not any research that com- pliers were indicated as delivery reliability, flexibility and respon-
bines these three methods together. Thereby, this study proposes siveness, cost, assets and environmental responsiveness. Tuzkaya,
a new integrated approach that could cope with the interdepen- Ozgen, Ozgen, and Tuzkaya (2009) evaluated the environmental
dencies among various criteria in fuzzy environment. Ford Otosan performance of suppliers with a hybrid fuzzy multi-criteria deci-
is selected as a case company in this study for the evaluation of sion approach: fuzzy ANP and Fuzzy Preference Ranking Organiza-
green supplier alternatives. The supplied case study provides addi- tion Method for Enrichment Evaluations (PROMETHEE)
tional insights for research and practical applications. methodology. In their study, evaluation criteria are determined
The organization of the paper is then as follows. The paper as pollution control, green process management, environmental
begins with the literature survey of GSCM. Then, after a brief and legislative management, environmental costs, green product,
review of methodologies, various main components of the GSCM and green image. Gumus (2009) introduced evaluation of hazard-
are examined to structure a framework for green supplier evalua- ous waste transportation firms by using a two step fuzzy-AHP
tion. The next section includes the illustration of the proposed and TOPSIS methodology. The determined criteria were hygiene
green supplier methodology through the case of Ford Otosan. The and safety, quality of service, complementary service, economic
paper concludes with future directions. factors, service time, taking care of the human health and environ-
mental protection standards, problem solving ability, and the
owned vehicle fleet. Lee et al. (2009) presented a green supplier
2. Literature survey selection model for high-tech industry. The required dimensions
for evaluating green suppliers were indicated as quality, technol-
Industrial production can have a great impact and damage on ogy capability, pollution control, environment management, green
the sustainability of the natural environment and human life such product, and green competencies/green image.
as the impacts include depletive resource use, global environmen- Recently, Bai and Sarkis (2010) proposed a study for green sup-
tal impacts, local environmental impacts, health impacts, and plier development and performed an analytical evaluation using
safety risks. These environmental issues have received more and rough set theory. The methodology generates decision rules relat-
more attention in recent years and supply chain operation with ing the various attributes to the performance outcomes (environ-
sustainable consideration has become an increasingly important mental, business, and joint performance). Kuo, Wang, and Tien
issue. Thereby, these growing interest and importance to the sup- (2010) integrated artificial neural network (ANN) and two multi-
ply function raise the importance of the environmental perfor- attribute decision analysis (MADA) methods: DEA and ANP for
mance of suppliers (Faruk, Lamming, Cousins, & Bowen, 2002; green supplier selection. Their green supplier selection structure
Hall, 2000; Sarkis, 2003; Simpson & Power, 2005). The benefits to contains quality, cost, delivery, service, environment, and corpo-
the firm arising from advanced environmental management prac- rate social responsibility. Punniyamoorthy, Mathiyalagan, and Par-
tice can include: cost reduction (efficient use of raw materials, thiban (2011) introduced a strategic model using structural
reduction in fines, risks or insurance costs); quality improvement; equation modeling and fuzzy logic in supplier selection. Their cri-
early adoption of new regulations; and better human resource teria of supplier selection are management and organization, qual-
management practice (Simpson & Power, 2005; Theyel, 2001). ity, technical capability, production facilities and capacities,
GSCs are gaining increasing interest among researchers and financial position, delivery, services, relationships, safety and envi-
practitioners. GSC is a broad concept that refers to a variety of ronmental concern, and cost. Awasthi, Chauhan, and Goyal (2010)
methods by which companies work with their suppliers to proposed a fuzzy multi-criteria approach for evaluating environ-
improve and maintain the performance of their products or manu- mental performance of suppliers. They used fuzzy TOPSIS for eval-
facturing processes of the suppliers, customers or both. The emer- uation and their criteria were usage of environment friendly
gence of GSC is one of the most significant developments in the technology, environment friendly materials, green market share,
past decade, offering the opportunity for companies to align their partnership with green organizations, management commitment
supply chains in accordance with environmental and sustainability to green practices, adherence to environmental policies, involve-
goals. ment in green projects, staff training, lean process planning, design
The most common GSCM practices involve organizations for environment, environmental certification, and pollution control
assessing the environmental performance of their suppliers, initiatives.
requiring suppliers to undertake measures that ensure environ-
mental quality of their products, and evaluating the cost of waste
in their operating systems (Handfield, Walton, Sroufe, & Melnyk, 3. Proposed green supplier evaluation framework
2002). A high level of environmental performance achieved by a
firm may be broken down by a poor level of environmental man- This study proposes a novel hybrid analytic approach based on
agement by its suppliers. Therefore, green suppliers and their the fuzzy DEMATEL, fuzzy ANP, and fuzzy TOPSIS methodologies to
selection, evaluation, etc. processes are vital in a green supply assist in GSCM strategic decisions. The general view of the pro-
chain. posed green supplier evaluation methodology is shown in Fig. 1.
3002 G. Büyüközkan, G. Çifçi / Expert Systems with Applications 39 (2012) 3000–3011

Based on these steps, we firstly mention about the proposed tech- part of this study because of its integrated methodology), in which
niques, and then we identify the green supplier evaluation criteria the numeral represents the strength of influence.
and present the proposed evaluation model in the following sub Although DEMATEL is a good technique for evaluating prob-
sections. lems, the relationships of systems are generally given by crisp val-
ues in establishing a structural model. However, in this real world,
crisp values are inadequate. Many evaluation criteria are surely
3.1. Proposed methodologies for the green supplier evaluation imperfect and probably uncertain factors. Thus, fuzzy theory
framework (Zadeh, 1965) is applied to the DEMATEL method for solving such
a MCDM problem. Fuzzy DEMATEL method is used as many
3.1.1. Fuzzy DEMATEL researchers in the literature (Chang, Chang, & Wu, 2011; Chen
The DEMATEL method, originated from the Geneva Research et al., 2008; Lin & Wu, 2008; Liou, Yen, & Tzeng, 2008; Tseng,
Centre of the Battelle Memorial Institute (Gabus & Fontela, 2009b; Wu & Lee, 2007), considering the fact that human judgment
1973), is especially pragmatic to visualize the structure of compli- about preferences are often unclear and hard to estimate by exact
cated causal relationships. DEMATEL is a comprehensive method numerical values.
for building and analyzing a structural model involving causal
relationships between complex factors. It can clearly see the cause-
effect relationship of criteria when measuring a problem (Chen-Yi, 3.1.2. Fuzzy ANP
Ke-Ting, & Gwo-Hshiung, 2007). It portrays a basic concept of con- ANP is a general form of the analytical hierarchy process (AHP)
textual relation among the elements of the system (which is not a first introduced by Saaty (1996). While the AHP employs a

Create a list of GSCM alternatives

Literature review
Define strategies, factor and sub-factors for evaluation of the
GSCM activities to build a framework

Expert opinion Establish interdependences between elements Expert opinion

Making of fuzzy DEMATEL-based calculations

Form fuzzy direct-relation matrix Construct fuzzy pairwise comparison matrices using
within factors using fuzzy DEMATEL triangular fuzzy numbers between factors using fuzzy ANP

Making of fuzzy ANP-based calculations

Acquire fuzzy normalized direct- Calculate fuzzy relative importance weights of matrices
relation matrix

Measure the consistency of the matrices

Acquire fuzzy total direction matrix
No

Obtain inner dependence matrix CR 0.10 Necessary

Yes modifications

Form a supermatrix by entering the vectors obtained from

fuzzy DEMATEL and fuzzy ANP evaluations into the
Integration of the inner appropriate columns
dependence matrix into the
related parts of ANP Normalize this unweighted supermatrix so that the
supermatrix
numbers in every column sum to one

Raise the unweighted supermatrix to the power 2n+1

Expert opinion

Obtain the green supplier evaluation criteria weights

Evaluate the alternatives by fuzzy TOPSIS

Calculate negative and positive ideal solutions and separation measures

Rank the preference order for green supplier alternatives

Fig. 1. Proposed green supplier evaluation methodology.

 G. Büyüközkan, G. Çifçi / Expert Systems with Applications 39 (2012) 3000–3011 3003

unidirectional hierarchical relationship among decision levels, the fuzzy ANP steps as mentioned. There are various fuzzy TOPSIS
ANP enables interrelationships among the decision levels and attri- studies in various areas as clean agent selection (Aiello, 2009),
butes in a more general form. Instead of a hierarchy, the ANP-based firms’ competence evaluation (Amiri, Zandieh, Soltani, & Vahdani,
system is a network that replaces single direction relationships with 2009), assessing thermal-energy storage in concentrated solar
dependence and feedback (Saaty, 1996). The ANP uses ratio scale power (CSP) systems (Cavallaro, 2010), development of a quick
measurements based on pair wise comparisons; however, it does credibility scoring decision support system (Iç & Yurdakul, 2010),
not impose a strict hierarchical structure as in AHP, and models a personnel selection (Kelemenis & Askounis, 2010), supplier selec-
decision problem using a systems-with-feedback approach. The tion (Roghanian, Rahimi, & Ansari, 2010), assessment of traffic
ANP refers then to the systems of which a level may both dominate police centers performance (Sadi-Nezhad & Damghani, 2010), eval-
and be dominated, directly or indirectly, by other decision attributes uating the competitive advantages of shopping websites (Sun and
and levels. Fig. 2 depicts the structure difference of a hierarchy and a Lin; 2009), virtual enterprise partner selection (Ye, 2010), etc.
network. In the literature, there are some realized studies that combine
The ANP approach is capable of handling interdependence ANP, TOPSIS and DEMATEL methodologies. Chen and Chen (2010)
among elements by obtaining the composite weights through the presented an innovation support system for Taiwanese higher edu-
development of a ‘‘supermatrix’’. A node represents a component cation using a novel conjunctive MCDM approach based on DEMA-
(or cluster) with elements inside it; a straight line/or an arc de- TEL, fuzzy ANP, and TOPSIS. Lin, Hsieh, and Tzeng (2010) evaluated
notes the interactions between two components; and a loop indi- vehicle telematics system by using DEMATEL, ANP, and TOPSIS
cates the inner dependence of elements within a component. For techniques with dependence and feedback. However, although
instance, when the elements of a component ‘‘Goal’’ depend on an- these kinds of combined works have increased in the recent years,
other component ‘‘Criteria’’, we represent this relation with an ar- there is not any study combines DEMATEL, ANP, and TOPSIS in fuz-
row from component ‘‘Goal’’ to ‘‘Criteria’’. The supermatrix zy environment.
development is defined in the next sub-section.
As indicated that human judgment about preferences are often
3.2. Criteria of green supplier evaluation framework
unclear and hard to estimate by exact numerical values, again fuz-
zy logic is necessary for handling problems characterized by
A detailed literature search with the concepts related to GSC is
vagueness and imprecision. Therefore human judgments, which
realized. We can find some concepts and elements which can be
are unclear, simultaneously address the issue of combining both
served as the foundation for a decision framework for prioritizing
fuzzy set theory and ANP for green supplier assessment. In the lit-
or selecting systems by the organization that would aid in selecting
erature many researchers such as Tuzkaya and Önüt (2008), Moh-
green suppliers. These are summarized as follows:
anty, Agarwal, Choudhury, and Tiwari (2005), Liu and Lai (2009),
Green logistics dimension: A more tactical set of organizational
Dağdeviren and Yüksel (2010), Luo, Zhou, Zheng, Mo, and He
elements that will influence how the supply chain is to be managed,
(2010), Liu and Wang (2010), Vinodh, Ramiya, and Gautham
either internally or externally, can be described by green logistics
(2011) applied fuzzy ANP to several research fields.
dimension of an organization. Major elements of the green logistics
dimension will typically include procurement, production, distribu-
3.1.3. Fuzzy TOPSIS tion, reverse logistics and packaging (Awasthi et al., 2010; Lee et al.,
TOPSIS is a multiple criteria method to identify solutions from a 2009; Punniyamoorthy et al., 2011; Rao & Holt, 2005; Sarkis, 2003;
finite set of alternatives and initially proposed by Chen and Hwang Sarkis, Meade, & Talluri, 2004; Tuzkaya et al., 2009; Zhu, Sarkis, &
(1992). The underlying logic of TOPSIS proposed by Hwang and Lai, 2007; Zhu, Sarkis, & Lai, 2008). It is estimated that 80% of all
Yoon (1981) is to define the ideal solution and negative ideal solu- product related environmental impacts are determined in the
tion. The optimal solution should have the shortest distance from design phase, so integrating environmental considerations early in
the positive ideal solution and the farthest from the negative ideal the product design development cycle is the most effective way of
solution. If to remind, human judgments are usually rely on impre- reducing their impact and the major elements of the design stage
cision, subjectivity and vagueness; so they address fuzzy logic. are the selection of the materials and production design (Goosey,
Here evaluations expressed by linguistic terms and then set into 2004). In an environmental friendly chain the first step is procure-
fuzzy numbers. ment and vendor selection. Production influences the green supply
Fuzzy TOPSIS methodology requires preliminarily information chain with the design and the production process. Within this func-
about the relative importance of the criteria. This importance is tion, environmental issues such as closed-loop manufacturing, total
expressed by attributing a weight to each considered criterion wj. quality environmental management, de-manufacturing and source
The weight of each criterion is evaluated by fuzzy DEMATEL and reduction make some form of value-adding contribution, even

A hierarchy A network

Component,
Goal … cluster, level etc. …C1…

Criteria … …C2…
Elements
…C4…
Alternatives
… …C3…
Loop indicates inner dependence of
the elements in that component with
respect to a common property.

Fig. 2. Structure of a hierarchy and a network.

3004 G. Büyüközkan, G. Çifçi / Expert Systems with Applications 39 (2012) 3000–3011

though some of them also influence other functional areas (Sarkis of strategic thinkers (Handfield & Nichols, 2002; Ketchen & Hult,
et al., 2004). For instance a well designed product should avoid the 2007; Vachon & Klassen, 2006). One characteristic of these perfor-
need for using hazardous or restricted materials during the manu- mance measures is that they are not static. They tend to change
facturing process and should minimize waste during the manufac- over time and will be greatly influenced by the product life cycle.
turing process (Jabbour & Jabbour, 2009; Kubokawa & Saito, 2000; That is, in the introduction phases, flexibility and time may be
Kurk & Eagan, 2008). Distribution is another operation that effect more important than cost. Whereas cost efficiencies tend to gain
green supply chain. The carrier, the capacity of the carrier, the type importance in more mature environments. These dynamical char-
of fuel that the carrier uses, the frequency of transportation and acteristics are incorporated into the decision framework.
the distance to the customers are some items that affect the perfor- Green supplier evaluation criteria: The major five evaluation cri-
mance of the green distribution. A significant trend in GSCM has teria for green suppliers are organization, financial performance,
been the recognition of the strategic importance of reverse logistics. service quality, technology, and green competencies (Awasthi
The definition of reverse logistics from an environmental perspec- et al., 2010; Bai & Sarkis, 2010; Humphreys et al., 2003; Kou
tive focuses primarily on the return of recyclable or reusable prod- et al., 2010; Lee et al., 2009; Punniyamoorthy et al., 2011; Rao &
ucts and materials into the forward supply chain. Designing Holt, 2005; Vachon & Klassen, 2006; Walker, Sisto, & McBain,
effective and efficient RL networks is a key driver for providing the 2008). Organization factor shows the supplier’s degree of compat-
economic benefits necessary to initiate and sustain GSCM initiatives ibility to the GSC. Here business structure, degree of cooperation
on a large scale (Srivastava, 2007). Packaging characteristics such as relationship closeness and attitudes are the critical factors for the
size, shape and materials have an impact on distribution due to their supplier to be appropriate to GSC. The compatible organization cul-
affect on the transport characteristics of the products. Better packag- tures and degree of fitness are some of the desired attributes.
ing, along with rearranged loading patterns can reduce material Financial performance shows the performance and control of the
usage, increase space utilization in warehouses and in trucks, and re- supplier economically. Financial position, economical stability
duce the amount of handling. Systems that encourage and adopt and price/cost can take part in financial performance. There is no
returnable packaging will require a strong customer-supplier rela- doubt that financial position of the supplier and the stability of
tionship and an effective reverse logistics channel. the finance is fundamental for the continuity of the supplier firms.
Green organizational activities dimension: The major five green Service quality contains the factors that can improve the quality of
organization activities dimensions are reduce, reuse, remanufac- suppliers so GSC. The quality certificates that the supplier has as
ture, recycle and disposal (Awasthi et al., 2010; Humphreys, Wong, ISO 9000, etc., information quality, capability of on time delivery
& Chan, 2003; Lee et al., 2009; Meade & Sarkis, 2002; Rao & Holt, and on time response to request are the important factors for qual-
2005; Sarkis, 2003; Simpson & Power, 2005; Tuzkaya et al., ified suppliers. With these factors, they can improve their quality,
2009). Reduction is viewed as an in-process, relatively proactive, responsiveness and efficiency which are essential for a supplier’s
measure that can be taken by organizations. The waste manage- continuity. Technology is the factor that can facilitate innovations
ment hierarchy can typically be depicted by an inverted triangle and flexibility to the supplier and SSC. Capacity, R&D capability,
with reuse at the top which has maximum width, signifying max- and capability to manage environmental technologies, reverse
imum preference to this management option. Reuse is ‘the use of a flows, etc. are the contents of the technology factor. By this way,
product or component part in its same form for the same use with- suppliers can be more innovative, flexible and environmentally
out remanufacturing. The reuse of product may be the reuse of the friendly. Lastly, green competencies show the competencies of
entire product, for example the selling of second hand cars or com- supplier in improving GSC management. It contains social respon-
puters, or it may be the reuse of components of a product, for sibility, cleaner/environmental production and technologies envi-
spares for example. Remanufacturing product involves bringing ronmental management system. The supplier organization should
used products up to quality standards which are as rigorous as also be capable of environmental management competencies and
those for new products. Recycling is the process by which products environmental image.
otherwise destined for disposal are processed to recover base Green supplier alternatives: Some green supplier alternatives are
materials, for example, precious metals from computer chips. For identified for improving the environmental performance of the
minimization of environmental impact the ideal scenario would supply chains of the organizations. Fig. 3 presents the network
be maximum possible reuse and disposal in a landfill only when structure of this evaluation framework.
it cannot be reused or recycled. As recycling is preferred over dis-
posal in a landfill for the objectives of minimization of environ- 4. Case Study
mental impact and perceived risk, recycling of the waste would
be preferred even after it is no longer economically attractive than 4.1. Application of the evaluation framework in Ford Otosan
disposal. This would mean a delay in shift from recycling stage of
hierarchy to disposal, as compared to the scenario of priority to Ford Motor Company is a multinational corporation and the
minimization of cost. However, when the objective shifts to mini- world’s third largest automaker based on worldwide vehicle sales.
mization of cost, reuse will be preferred only if it is economically In 2006, Ford was the second-ranked automaker in the US with a
more attractive than recycle and recycling would be continued
only if it is economically more attractive than disposal in a landfill.
Organizational performance dimension: There are four widely
Goal
accepted manufacturing performance indicators: cost, quality, Organizational
delivery and flexibility (Jabbour & Jabbour, 2009; Kuo et al., Performance
2010; Lee et al., 2009; Tuzkaya et al., 2009; Punniyamoorthy
Green Logistics
et al. (2011). These generic strategic performance requirements,
which may not be environmentally based, are necessary to help
identify how well various alternatives can perform on these fac-
Green Organizational
tors. They are necessary because the alternative that is selected Green Supplier Activities
should not only best support the green supply chain, but also Evaluation Criteria
makes sense from a business perspective. The use of these organi-
zational performance measures have been supported by a number Fig. 3. Network structure of the evaluation framework.
 G. Büyüközkan, G. Çifçi / Expert Systems with Applications 39 (2012) 3000–3011 3005

17.5% market share, behind General Motors (24.6%) but ahead of e ;can be acquired by using the following formulas, in which the I is
T
Toyota (15.4%) and Daimler Chrysler (14.4%). In the 2007 Fortune denoted as the identity matrix.
500 list, Ford was the seventh-ranked American-based company Let ~
xij ¼ ðlij ; mij ; uij Þ and define three crisp matrices, whose ele-
according to global revenues ($160.1 Billion). In 2006, Ford pro- ments are extracted from X e as follows [35].
duced approximately 6.6 million automobiles, and employed about 2 3 2 3
0 l12 . . . l1n 0 m12 . . . m1n
280,000 employees at 100 plants and facilities worldwide (http:// 6 7 6 7
www.ford.com). 6 l21 0 . . . l2n 7 6 m21 0 . . . m2n 7
6 7 6 7
Otosan started production in 1965 and since then has occupied 6 7 6 7
6 : : 7 6 : : 7
a major role in the development of Turkish automotive industry. In 6 7 6 7
X1 ¼ 6
6
7;
7 X2 ¼ 6
6
7;
7
1997, Ford Motor Company and Koç Holding signed an agreement 6 : : 7 6 : : 7
and created Ford Otosan as a joint venture. Each company holds 6 7 6 7
6 7 6 7
41% share in the venture. Today, Ford Otosan’s capital is 500 Mil- 6 : : 7 6 : : 7
4 5 4 5
lion TL. Ford Otosan has three facilities in Turkey and is employing
ln1 ln2 ... 0 mn1 mn2 ... 0
8008 people. In 2006, Ford Otosan sold 113.857 vehicles just in
2 3
Turkey and has been the market leader last 5 years. In 2006, Ford 0 u12 . . . u1n
Otosan was the market leader with a 17.1% market share (http:// 6 7
6 u21 0 . . . u2n 7
www.ford.com.tr). Ford Otosan’s plant, located in Kocaeli, is named 6 7
6 7
as ‘‘Best Plant in the World’’ having the best scores in 2002, 2003, 6 : : 7
6 7
2004 and 2005 among European Ford Plants. X3 ¼ 6
6
7:
7
Today, Ford Otosan is one of the biggest and most technologi- 6 : : 7
6 7
cally advanced automotive plants in the world and green practices
6 7
6 : : 7
4 5
are implemented at all stages of the manufacturing process. For
these reasons Ford Otosan is selected as a case company in this un1 un2 ... 0
study to evaluate green supply chain management initiatives.
According to the crisp case, we define the total-relation fuzzy
Decision makers were Vedat Okyar (Senior Purchasing Manager- e through (2):
matrix T
Trim Parts in Gölcük Plant) and Serdar Aydın (New Project Chief
in Gölcük Plant). There were five possible green suppliers that Te ¼ X
e ðI  X
e Þ1 : ð2Þ
are thought they have specific green competencies.
2~ ~t 12 3
t 11 . . . ~t 1n
4.2. The computational steps of the proposed integrated framework 6 ~t ~t 22 . . . ~t 2n 7
6 21 7
6 7
6 : : 7
Step 1: Determination of the evaluation model. After setting the Let Te ¼ 6
6 :
7;
decision goal, construct a committee of experts with E members 6 : 77
6 7
and determine the alternatives and sets of criteria for evaluation. 4 : : 5
The evaluation criteria have already been discussed in Section 3.2 et n1 et n2 . . . ~t nn
and the evaluation model can be seen in Fig. 4. 0
Step 2: Design fuzzy linguistic scale for evaluations. In this step, where ~tij ¼ ðlij ; m0ij ; u0ij Þ then
h i
development of relationships within and among the attributes 0
Matrix lij ¼ X l ðI  X l Þ1 ; ð3Þ
using experts’ opinion through paired comparison analysis is
needed. Firstly, for the purpose of measuring the relationships, it h i
is required to design the comparison scale as shown in Table 1. Matrix m0ij ¼ X m ðI  X m Þ1 ; ð4Þ
The different degrees of influence are expressed with eleven lin-
guistic terms and the equivalent fuzzy membership functions for h i
linguistic values are shown in Fig. 5. Consensus of opinions exists Matrix u0ij ¼ X u ðI  X u Þ1 : ð5Þ
among experts in the evaluation process.
By applying these formulas, the total-relation matrix acquired is
Step 2: Establish casual relations using the fuzzy DEMATEL.
given in Table 4.
Step 2.1: Acquire fuzzy direct-relation matrix. Experts make
Step 2.4: Obtain the inner dependence matrix. In this step, after
sets of the pairwise comparisons in terms of influence and direc- e by using Eq. (6), the
e in which defuzzification of the total-relation matrix T
tion within necessary criteria that is a n  n matrix A,
~ij ¼ ðlij ; mij ; uij Þ is denoted as the degree to which the criterion i sum of each column in total-relation matrix became equal to 1
a
by the normalization method.
affects the criterion j for experts. Table 2 gives an example of fuzzy
direct-relation matrix for organization performance dimension. Z 1  
Step 2.2: Acquire normalized fuzzy direct-relation matrix. After Fð~t ij Þ ¼ 1=2 inf ~t aij þ sup ~t aij da: ð6Þ
0 x2R x2R
producing the direct-relation matrix as the first step, we can con-
tinue with normalizing the direct-direction matrix as in DEMATEL Then the inner dependence matrix can be acquired to put in the un-
method. On the base of the direct-relation matrix A, e the normal- weighted supermatrix of ANP later. Table 5 shows the inner depen-
e
ized direct-relation matrix X can be obtained through Eq. (1). In dence matrix of organizational performance dimension and can be
Table 3, normalized direct-relation matrix can be seen seen in Fig. 6 as matrix B of the supermatrix.
Step 3: Establish remaining relations using the fuzzy ANP. In
X
n
~ij ¼ ðlij ; mij ; uij Þ and s ¼ 1=max16i6n
Let a uij ; then ANP, like AHP, pair wise comparisons of the elements in each level
j¼1 are conducted with respect to their relative importance towards
their control criterion. By using triangular fuzzy numbers again,
e
e ¼ s  A: the relative strength of each pair of elements and the preferences
X ð1Þ
of the decision maker in the same hierarchy are indicated. Via
Step 2.3: Acquire fuzzy total-relation matrix. As soon as the normal- pair-wise comparison, the fuzzy judgment matrix A e 0 is constructed
ized direct-relation matrix Xe is obtained, the total-relation matrix as:
3006 G. Büyüközkan, G. Çifçi / Expert Systems with Applications 39 (2012) 3000–3011

GOAL
Organizational
Performance Dimension Improve green supply chain
management activities
Quality Cost Time Flexibility
(OP1) (OP2) (OP3) (OP4) (B)

(A)
(D) (C)
Green Logistics
Dimension
(E)
Procurement Production Distribution Reverse L. Packaging
(GL1) (GL2) (GL3) (GL4) (GL5)

(F)
Green Organizational
Activities Dimension

Reduce Recycle Remanufacture Reuse Disposal

(GOA1) (GOA2) (GOA3) (GOA4) (GOA5)

(G) (H) (I) Green Supplier

Evaluation Criteria

Organization Financial Performance Service Quality Technology Green Competencies

(EC1) (EC2) (EC3) (EC4) (EC5)

S1 S2 S3 S4 S5

Fig. 4. Detailed evaluation model.

Table 1
Corresponding linguistic terms for evaluation.

Linguistic term Abbrev. Fuzzy scales

None N (0, 0, 1)
Very Low VL (0, 0.1, 0.2)
Low L (0.1, 0.2, 0.3)
Fairly Low FL (0.2, 0.3, 0.4)
More or less Low ML (0.3, 0.4, 0.5)
Medium M (0.4, 0.5, 0.6)
More or less Good MG (0.5, 0.6, 0.7)
Fairly Good FG (0.6, 0.7, 0.8)
Good G (0.7, 0.8, 0.9)
Very Good VG (0.8, 0.9, 1)
Excellent E (0.9, 1, 1) Fig. 5. Fuzzy membership functions for linguistic values.

these weights (Onüt, Kara, & Isßik, 2009; Ramik, 2007; Tuzkaya
2 ~0 ~012 ~01n 3 et al., 2009; Tuzkaya & Önüt, 2008):
a11 a ... a
6a 0 ~022 ~02n 7
6 ~21 a ... a 7 ~ k ¼ ðwl k; wm
w u
6 7 k ; wk Þ k ¼ 1; 2; . . . ; n where;
6 : 7
e0 ¼ 6 :
A 7; ð7Þ
6 : : 7
6 7 Qn
6 7 ð askj Þ1=n
4 : : 5 wsk ¼ Pn
i¼1
Qn m 1=n ; s 2 fl; m; ug ð8Þ
~0n1
a ~0n2
a ~0nn
... a i¼1 ð i¼1 aij Þ

 
where a ~0ij ¼ l0ij ; m0ij ; u0ij indicates the importance among the com- for 0 < a 6 1 and all i, j, where i = 1, 2, . . . , n, j = 1, 2 . . . , n. In order to
pared criteria (importance of i over j) where i = j = 1,2, . . . , n. Table control the result of the method, the consistency ratio for each of
6 gives examples of linguistic and fuzzy evaluations between green the matrices and the overall inconsistency for the hierarchy are cal-
logistics dimensions and goal. This evaluation can be seen in Fig. 6 culated. The Consistency Ratio (CR) is used to directly estimate the
as matrix A of supermatrix. Other evaluations are populated in the consistency of the pair-wise comparisons and should be less than
same way. 0.10. Then it can be said the comparisons are acceptable, otherwise
Step 3.1: Calculate the relative importance weights. The priority they are not acceptable. In this study, the inconsistency ratios for all
vectors for each pairwise comparison matrix will be needed to the comparison matrices were calculated for the mean values of the
complete the various supermatrix submatrices. Estimate triangular fuzzy numbers. Because the lower and upper values provide flexi-
fuzzy priorities w ~ k where k = 1, 2 . . . , n from the judgment matrix. bility for human judgments, they are not expected to have rigid
The logarithmic least-squares method can be used for calculating consistency.
 G. Büyüközkan, G. Çifçi / Expert Systems with Applications 39 (2012) 3000–3011 3007

Table 2
Fuzzy direct relation matrix of organizational performance dimension.

Quality (OP1) Cost (OP2) Time (OP3) Flexibility (OP4)

Quality (OP1) ⁄ (0.8, 0.9, 1) (0.4, 0.5, 0.6) (0, 0.1, 0.2)
Cost (OP2) (0.5, 0.6, 0.7) ⁄ (0.4, 0.5, 0.6) (0, 0, 0.1)
Time (OP3) (0.4, 0.5, 0.6) (0.6, 0.7, 0.8) ⁄ (0, 0.1, 0.2)
Flexibility (OP4) (0.3, 0.4, 0.5) (0.8, 0.9, 1) (0.5, 0.6, 0.7) ⁄

Step 3.2: Defuzzify the weights obtained from fuzzy matrices. In 4.1: Solve the supermatrix. To complete this task, firstly each of
this step, defuzzification of the weights is done in the same way as the columns may either be normalized by dividing each weight in
Eq. (6). the column by the sum of that column. Then, the final step in the
Such an example, priority calculation of the production – one of process is to obtain a priority ranking for each of the alternatives.
the green logistics dimensions – with respect to goal (from Table 6) To derive the overall priorities of elements, the normalized
is as follows. By applying Eq. (8), the fuzzy weight is obtained as, supermatrix is raised to limiting powers to calculate the overall

1=5
ð0:2  0:1  0:6  1=1  1Þ
wlk ¼ 1=5 1=5
¼ 0:3048:
ð1  0:9  1=0:4  1=0:3  1=0:3Þ þ ð1=0:9  1  1=0:4  1=0:2  1=0:2Þ þ ð0:4  0:4  1  1=0:7  1=0:7Þ1=5 þ ð0:3  0:2  0:7  1  1Þ1=5 þ ð0:3  0:2  0:7  1=1  1Þ1=5

1=5
ð0:3  0:2  0:7  1=1  1Þ
wm
k ¼ ¼ 0:3828:
ð1  0:9  1=0:4  1=0:3  1=0:3Þ1=5 þ ð1=0:9  1  1=0:4  1=0:2  1=0:2Þ1=5 þ ð0:4  0:4  1  1=0:7  1=0:7Þ1=5 þ ð0:3  0:2  0:7  1  1Þ1=5 þ ð0:3  0:2  0:7  1=1  1Þ1=5

1=5
ð0:4  0:3  0:8  1=0:9  1Þ
wuk ¼ 1=5 1=5
¼ 0:5478:
ð1  0:9  1=0:4  1=0:3  1=0:3Þ þ ð1=0:9  1  1=0:4  1=0:2  1=0:2Þ þ ð0:4  0:4  1  1=0:7  1=0:7Þ1=5 þ ð0:3  0:2  0:7  1  1Þ1=5 þ ð0:3  0:2  0:7  1=1  1Þ1=5

Then using this fuzzy vector and applying Eq. (6), defuzzified priorities, and thus the cumulative influence of each element on
weight 0.4046 is obtained. every other element with which it interacts is obtained. In this
Step 4: Form a supermatrix. ANP uses the formation of a superm- case, the supermatrix is raised to the power 25. This weighted
atrix to allow for the resolution of the effects of the interdepen- supermatrix is shown in Table 8.
dence that exists between the clusters within the decision According to this weighted supermatrix, weights of the criteria
network hierarchy. The supermatrix is a partitioned matrix, where on the objective of green supplier selection are shown in the ‘‘Goal’’
each submatrix is composed of a set of relationships between two column to use in fuzzy TOPSIS steps later.
clusters in the graphical model. A generic supermatrix is shown Step 5: Evaluate the alternatives by using fuzzy TOPSIS steps.
in Fig. 6, with the notation representing the various relationships The technique is adapted from Chen (2000) and the steps of the
from Fig. 4; for instance, ‘‘A’’ is the submatrix representing the influ- methodology are as follows.
ence relationship between green logistics dimension elements’ and Step 5.1: Establish fuzzy decision matrix for evaluation of the
control factor of the goal of selecting a green supplier. green supplier alternatives. With m alternatives and n criteria, fuz-
By entering the priorities found by fuzzy DEMATEL and fuzzy zy MCDM problem can be expressed as:
ANP into the appropriate columns, initial supermatrix can be con-
structed. Table 7 presents the initial supermatrix of the study.

Table 5
Inner dependence matrix of organizational performance dimension.
Table 3
Fuzzy normalized direct relation matrix of organizational performance dimension. (OP1) (OP2) (OP3) (OP4)

(OP1) (OP2) (OP3) (OP4) (OP1) 0.19 0.28 0.27 0.29

(OP2) 0.25 0.14 0.24 0.20
(OP1) ⁄ (0.36, 0.40, 45) (0.18, 0.22, 0.27) (0, 0.04, 0.09)
(OP3) 0.26 0.25 0.16 0.28
(OP2) (0.22, 0.27, 0.31) ⁄ (0.18, 0.22, 0.27) (0, 0, 0.04)
(OP4) 0.30 0.33 0.33 0.23
(OP3) (0.18, 0.22, 0.27) (0.27, 0.31, 0.36) ⁄ (0, 0.04, 0.09)
(OP4) (0.13, 0.18, 0.22) (0.36, 0.40, 0.45) (0.22, 0.27, 0.31) ⁄

Goal GL OP GOA EC
Table 4 Goal 0 0 0 0 0
Fuzzy total direct relation matrix of organizational performance dimension. Green Logistics Dimensions (GL) A E D 0 0
(OP1) (OP2) (OP3) (OP4) Organizational Performance (OP) 0 C B 0 0
(OP1) (0.17, 0.33, 0.70) (0.51, 0.73, 1.23) (0.31, 0.49, 0.89) (0, 0.08, 0.29) Green Organizational Activities (GOA) 0 F 0 0 0
(OP2) (0.32, 0.47, 0.82) (0.19, 0.34, 0.76) (0.28, 0.42, 0.78) (0, 0.04, 0.23)
Green Supplier Evaluation Criteria (EC) 0 H G I I
(OP3) (0.30, 0.48, 0.86) (0.42, 0.63, 1.10) (0.13, 0.27, 0.62) (0, 0.08, 0.28)
(OP4) (0.34, 0.56, 1.03) (0.60, 0.86, 1.43) (0.40, 0.61, 1.07) (0, 0.05, 0.26)
Fig. 6. General submatrix notation for supermatrix. Note: I is the identity matrix.
3008 G. Büyüközkan, G. Çifçi / Expert Systems with Applications 39 (2012) 3000–3011

Table 6
Linguistic and fuzzy evaluation matrices of green logistics with respect to goal.

Linguistic terms Fuzzy terms

GL1 GL2 GL3 GL4 GL5 GL1 GL2 GL3 GL4 GL5
1 VG 1 (0.8, 0.9, 1) (1/0.5, 1/0.4, 1/0, 3) (1/0.4, 1/0.3, 1/0, 2) (1/0.4, 1/0.3, 1/0, 2)
1 (1/1, 1/0.9, 1/0, 8) 1 (1/0.5, 1/0.4, 1/0, 3) (1/0.3, 1/0.2, 1/0, 1) (1/0.3, 1/0.2, 1/0, 1)
ML ML 1 (0.3, 0.4, 0.5) (0.3, 0.4, 0.5) 1 (1/0.8, 1/0.7, 1/0, 6) (1/0.8, 1/0.7, 1/0, 6)
FL L FG 1 E (0.2, 0.3, 0.4) (0.1, 0.2, 0.3) (0.6, 0.7, 0.8) 1 (0.9, 1, 1)
FL L FG 1 (0.2, 0.3, 0.4) (0.1, 0.2, 0.3) (0.6, 0.7, 0.8) (1/1, 1/1, 1/0, 9) 1

Table 7
Initial supermatrix of green supplier selection for the improvement of GSC.

Goal GL1 GL2 GL3 GL4 GL5 OP1 OP2 OP3 OP4 GOA1 GOA2 GOA3 GOA4 GOA5 C1 C2 C3 C4 C5
Goal 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
GL1 0.32 0.18 0.30 0.22 0.25 0.29 0.43 0.44 0.08 0.06 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
GL2 0.40 0.34 0.21 0.26 0.30 0.28 0.29 0.20 0.34 0.16 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
GL3 0.13 0.13 0.10 0.09 0.14 0.16 0.14 0.20 0.34 0.44 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
GL4 0.08 0.15 0.21 0.17 0.11 0.13 0.06 0.07 0.17 0.26 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
GL5 0.09 0.20 0.18 0.26 0.19 0.14 0.07 0.07 0.09 0.08 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
OP1 0.00 0.61 0.12 0.24 0.13 0.52 0.19 0.28 0.27 0.29 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
OP2 0.00 0.26 0.06 0.09 0.59 0.26 0.25 0.14 0.24 0.20 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
OP3 0.00 0.08 0.25 0.53 0.23 0.16 0.26 0.25 0.16 0.28 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
OP4 0.00 0.06 0.57 0.14 0.06 0.06 0.30 0.33 0.33 0.23 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
GOA1 0.00 0.46 0.46 0.46 0.46 0.46 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
GOA2 0.00 0.10 0.10 0.10 0.10 0.10 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
GOA3 0.00 0.15 0.15 0.15 0.15 0.15 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
GOA4 0.00 0.26 0.26 0.26 0.26 0.26 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
GOA5 0.00 0.05 0.05 0.05 0.05 0.05 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
C1 0.00 0.20 0.15 0.20 0.25 0.10 0.19 0.11 0.20 0.25 0.10 0.10 0.10 0.15 0.20 1.00 0.00 0.00 0.00 0.00
C2 0.00 0.25 0.25 0.25 0.10 0.20 0.20 0.30 0.25 0.25 0.15 0.10 0.15 0.15 0.20 0.00 1.00 0.00 0.00 0.00
C3 0.00 0.25 0.25 0.15 0.20 0.20 0.30 0.19 0.15 0.19 0.15 0.10 0.15 0.15 0.10 0.00 0.00 1.00 0.00 0.00
C4 0.00 0.20 0.25 0.30 0.15 0.20 0.20 0.20 0.30 0.26 0.24 0.35 0.24 0.20 0.20 0.00 0.00 0.00 1.00 0.00
C5 0.00 0.10 0.10 0.10 0.30 0.30 0.11 0.20 0.10 0.05 0.36 0.35 0.36 0.35 0.30 0.00 0.00 0.00 0.00 1.00

Table 8
Weighted supermatrix of green supplier selection for the improvement of GSC.

GOAL GL1 GL2 GL3 GL4 GL5 OP1 OP2 OP3 OP4 GOA1 GOA2 GOA3 GOA4 GOA5
C1 0.16 0.16 0.16 0.16 0.17 0.14 0.18 0.15 0.18 0.20 0.11 0.10 0.11 0.15 0.20
C2 0.20 0.20 0.20 0.20 0.17 0.19 0.21 0.24 0.22 0.22 0.15 0.10 0.15 0.15 0.20
C3 0.19 0.20 0.19 0.17 0.18 0.18 0.23 0.19 0.18 0.19 0.15 0.10 0.15 0.15 0.10
C4 0.23 0.22 0.24 0.25 0.21 0.23 0.22 0.22 0.26 0.25 0.24 0.35 0.24 0.20 0.20
C5 0.22 0.22 0.21 0.22 0.27 0.26 0.16 0.20 0.16 0.14 0.35 0.35 0.35 0.35 0.30

Table 9
Linguistic and fuzzy decision matrix for green supplier alternative evaluation.

Linguistic terms Fuzzy terms

C1 C2 C3 C4 C5 C1 C2 C3 C4 C5
S1 FG G FG MG ML (0.6, 0.7, 0.8) (0.7, 0.8, 0.9) (0.6, 0.7, 0.8) (0.5, 0.6, 0.7) (0.3, 0.4, 0.5)
S2 VG G G VG VG (0.8, 0.9, 1) (0.7, 0.8, 0.9) (0.7, 0.8, 0.9) (0.8, 0.9, 1) (0.8, 0.9, 1)
S3 VG E E G G (0.8, 0.9, 1) (0.9, 1, 1) (0.9, 1, 1) (0.7, 0.8, 0.9) (0.7, 0.8, 0.9)
S4 MG VG G VG MG (0.5, 0.6, 0.7) (0.8, 0.9, 1) (0.7, 0.8, 0.9) (0.8, 0.9, 1) (0.5, 0.6, 0.7)
S5 M MG FG G G (0.4, 0.5, 0.6) (0.5, 0.6, 0.7) (0.6, 0.7, 0.8) (0.7, 0.8, 0.9) (0.7, 0.8, 0.9)

Table 10
Weighted decision matrix for green supplier alternative evaluation.

C1 C2 C3 C4 C5
S1 (0.10, 0.11, 0.13) (0.14, 0.16, 0.18) (0.11, 0.13, 0.15) (0.12, 0.14, 0.16) (0.07, 0.09, 0.11)
S2 (0.13, 0.14, 0.16) (0.14, 0.16, 0.18) (0.13, 0.15, 0.17) (0.18, 0.21, 0.23) (0.18, 0.20, 0.22)
S3 (0.13, 0.14, 0.16) (0.18, 0.20, 0.20) (0.17, 0.19, 0.19) (0.16, 0.18, 0.21) (0.15, 0.18, 0.20)
S4 (0.08, 0.10, 0.11) (0.16, 0.18, 0.20) (0.13, 0.15, 0.17) (0.18, 0.21, 0.23) (0.11, 0.13, 0.15)
S5 (0.06, 0.08, 0.10) (0.10, 0.12, 0.14) (0.11, 0.13, 0.15) (0.16, 0.18, 0.21) (0.15, 0.18, 0.20)
 G. Büyüközkan, G. Çifçi / Expert Systems with Applications 39 (2012) 3000–3011 3009

Table 11
Positive–negative distances and final performance indices of green supplier alternatives.

Positive Negative
d1 d2 d3 d4 d5 dTOT d1 d2 d3 d4 d5 dTOT
S1 0.89 0.84 0.87 0.86 0.91 4.37 0.11 0.16 0.13 0.14 0.09 0.64
S2 0.86 0.84 0.85 0.79 0.80 4.14 0.14 0.16 0.15 0.21 0.20 0.86
S3 0.86 0.81 0.82 0.82 0.82 4.12 0.14 0.19 0.18 0.18 0.18 0.88
S4 0.90 0.82 0.85 0.79 0.87 4.23 0.10 0.18 0.15 0.21 0.13 0.77
S5 0.48 0.88 0.87 0.82 0.82 3.87 0.08 0.12 0.13 0.18 0.18 0.70

þ
X
n
Table 12 di ¼ dðv~ ij ; v~ þj Þ; i ¼ 1; 2; . . . ; m; j ¼ 1; 2 . . . n; ð12Þ
Final performance indices of green supplier
j¼1
alternatives.

Performance index 
X
n
di ¼ dðv~ ij ; v~ j Þ; i ¼ 1; 2 . . . m; j ¼ 1; 2; . . . ; n; ð13Þ
S1 0.1272 j¼1
S2 0.1728
S3 0.1767 rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
S4 0.1541
e BÞ
e ¼ 1h i
S5 0.1528 dð A; ða1  b1 Þ2 þ ða2  b2 Þ2 þ ða3  b3 Þ2 : ð14Þ
3
Positive and negative distances of the green supplier alternatives
can be seen in Table 11.
Step 5.5: Rank the alternatives. The performance indices are
computed to rank the alternatives. Performance indices are sorted
in a decreasing order. Table 12 shows the final ranking and accord-
ing to this hybrid methodology, the best possible green supplier is
S3 with a score of 0.1767.

5. Conclusion
e represents the fuzzy decision matrix with alternatives A and cri-
D
teria C, and can be seen with linguistic and fuzzy terms in Table 9. This study suggests a novel hybrid MCDM approach to evaluate
Step 5.2: Normalize the decision matrix. Normalized fuzzy deci- green suppliers for the need of improving GSCM initiatives. Based
sion matrix Re is calculated as: on the literature survey and with the validation of industrial ex-
perts, possible green supplier evaluation criteria were defined
e ¼ ½~r ij   ;
R i ¼ 1; 2 . . . ; m; j ¼ 1; 2 . . . n; and a new evaluation model was formulated. The proposed model
m n
was implemented in Ford Otosan, one of the pioneering companies
! about environmental subjects in Turkey.
aij bij cij
~r ij ¼ ; ; ; ð9Þ The combined fuzzy ANP and fuzzy DEMATEL approaches used
C þj C þj C þj in this study offered a more precise and accurate analysis by inte-
grating interdependent relationships within and among a set of cri-
where C þ j ¼ maxi C ij .To avoid the complicated normalization for- teria. Moreover, fuzzy TOPSIS method helped to choose the
mula used in the classical TOPSIS, the linear scale transformation alternative for ideal solution of this problem efficiently.
is used to transform the various criteria scales into a comparable While it is believed that the presented model provides value,
scale (Chen, 2000). Linear scale transformation for normalization there are also further points that can be included. To our knowl-
is also employed by Kuo et al. (2007) and Celik et al. (2009). Here edge, no previous work investigated such a problem by an inte-
normalized decision matrix remains the same because max Cij = 1. grated method with DEMATEL, ANP, and TOPSIS in fuzzy
Step 5.3: Compute weighted decision matrix. Weighted normal- environment. As the proposed approach is novel, it might be
ized fuzzy decision matrix that is shown in Table 10 is computed applied to other MCDM problems.
by using Eq. (11), where wj is the weight for the criterion j obtained
from supermatrix
Acknowledgements

v~ ij ¼ ~rij  w~ j ; ð10Þ The authors would like to express their deep gratitude towards
the industrial experts of Ford Otosan, especially to Vedat Okyar
where v~ ¼ ½v~ ij m n ; i ¼ 1; 2; . . . ; m; j ¼ 1; 2; . . . ; n.
(Senior Purchasing Manager-Trim Parts in Gölcük Plant) and Serdar
Step 5.4: Calculate the distances from positive and negative
Aydın (New Project Chief in Gölcük Plant). The authors acknowl-
ideal points. Since the triangular fuzzy numbers are included in
edge Alisßan Çapan for his contribution in realizing the application
[0, 1] range, positive and negative ideal reference points (FPIRP,
part and R.Ufuk Bilsel for his help in improving the linguistic qual-
FNIRP) are as follows:
ity of the paper.
Aþ ¼ fv~ þ1 ; v~ þ2 ; . . . v~ þn g; A ¼ fv~ 1 ; v~ 2 ; . . . v~ n g; ð11Þ
References
where v~ þ
j ¼ ð1; 1; 1Þ, v j ¼ ð0; 0; 0Þ.
~
The next step is to calculate the distance of alternatives from Aiello, G. (2009). Clean agent selection approached by fuzzy TOPSIS decision-
FPIRP and FNIRP. making method. Fire Technology, 45, 405–418.
3010 G. Büyüközkan, G. Çifçi / Expert Systems with Applications 39 (2012) 3000–3011

Amiri, M., Zandieh, M., Soltani, R., & Vahdani, B. (2009). A hybrid multi-criteria Lin, C.-L., Hsieh, M.-S., & Tzeng, G.-H. (2010). Evaluating vehicle telematics system
decision-making model for firms competence evaluation. Expert Systems with by using a novel MCDM techniques with dependence and feedback. Expert
Applications, 36, 12314–12322. Systems with Applications, 37(10), 6723–6736.
Awasthi, A., Chauhan, S. S., & Goyal, S. K. (2010). A fuzzy multicriteria approach for Lin, C.-J., & Wu, W.-W. (2008). A causal analytical method for group decision-
evaluating environmental performance of suppliers. International Journal of making under fuzzy environment. Expert Systems with Applications, 34(1),
Production Economics, 126(2), 370–378. 205–213.
Bai, C., & Sarkis, J. (2010). Green supplier development: Analytical evaluation using Lin, R.-H. (2009). An integrated FANP–MOLP for supplier evaluation and order
rough set theory. Journal of Cleaner Production, 18(12), 1200–1210. allocation. Applied Mathematical Modelling, 33, 2730–2736.
Cavallaro, F. (2010). Fuzzy TOPSIS approach for assessing thermal-energy storage in Liou, J. J. H., Yen, L., & Tzeng, G.-H. (2008). Building an effective safety management
concentrated solar power (CSP) systems. Applied Energy, 87(2), 496–503. system for airlines. Journal of Air Transport Management, 14(1), 20–26.
Chan, F. T. S., & Kumar, N. (2007). Global supplier development considering risk Liu, K. F. R., & Lai, J.-H. (2009). Decision-support for environmental impact
factors using fuzzy extended AHP-based approach. Omega, 35, 417–431. assessment: A hybrid approach using fuzzy logic and fuzzy analytic network
Chang, B., Chang, C.-W., & Wu, C.-H. (2011). Fuzzy DEMATEL method for developing process. Expert Systems with Applications, 36, 5119–5136.
supplier selection criteria. Expert Systems with Applications, 38(3), 1850–1858. Liu, H.-T., & Wang, C.-H. (2010). An advanced quality function deployment model
Chen, C.C., Tseng, M.L., & Lin, Y.H. (2008). Using fuzzy DEMATEL to develop a causal using fuzzy analytic network process. Applied Mathematical Modelling, 34(11),
and effect model of hot spring service quality expectation. In IEEE international 3333–3351.
conference on industrial engineering and engineering management (pp. 1004– Lu, L. Y. Y., Wu, C. H., & Kuo, T.-C. (2007). Environmental principles applicable to
1008). green supplier evaluation by using multiobjective decision analysis.
Chen, J.-K., & Chen, I-S. (2010). Using a novel conjunctive MCDM approach based on International Journal of Production Research, 45, 4317–4331.
DEMATEL, fuzzy ANP, and TOPSIS as an innovation support system for Luo, Z.-M., Zhou, J.-Z., Zheng, L.-P., Mo, L., & He, Y.-Y. (2010). A TFN–ANP based
Taiwanese higher education. Expert Systems with Applications, 37, 1981–1990. approach to evaluate Virtual Research Center comprehensive performance.
Chen, S. J., & Hwang, C. L. (1992). Fuzzy multiple attribute decision-making methods Expert Systems with Applications, 37(12), 8379–8386.
and application. In Lecture notes in economics and mathematical systems. New Meade, L., & Sarkis, J. (2002). A conceptual model for selecting and evaluating third-
York: Springer. party reverse logistics providers. Supply Chain Management: An International
Chen-Yi, H., Ke-Ting, C., & Gwo-Hshiung, T. (2007). FMCDM with fuzzy DEMATEL Journal, 7(5), 283–295.
approach for customers’ choice behavior model. International Journal of Fuzzy Mohanty, R. P., Agarwal, R., Choudhury, A. K., & Tiwari, M. K. (2005). A fuzzy ANP-
Systems, 9(4), 236–246. based approach to R&D project selection: A case study. International Journal of
Dağdeviren, M., & Yüksel, I. _ (2010). A fuzzy analytic network process (ANP) model Production Research, 43, 5199–5216.
for measurement of the sectoral competition level (SCL). Expert Systems with Ozgen, D., Önüt, S., Gülsün, B., Tuzkaya, U. R., & Tuzkaya, G. (2008). A two-phase
Applications, 37(2), 1005–1014. possibilistic linear programming methodology for multi-objective supplier
Faruk, A. C., Lamming, R. C., Cousins, P. D., & Bowen, F. E. (2002). Analyzing, evaluation and order allocation problems. Information Sciences, 178, 485–500.
mapping, and managing environmental impacts along the supply chain. Journal Onüt, S., Kara, S. S., & Isßik, E. (2009). Long term supplier selection using a combined
of Industrial Ecology, 5(2), 13–36. fuzzy MCDM approach: A case study for a telecommunication company. Expert
Gabus, A., & Fontela, E. (1972). World problems, an invitation to further thought within Systems with Applications, 36, 3887–3895.
the framework of DEMATEL. Switzerland, Geneva: Battelle Geneva Research Punniyamoorthy, M., Mathiyalagan, P., & Parthiban, P. (2011). A strategic model
Centre. using structural equation modeling and fuzzy logic in supplier selection. Expert
Gabus, A., & Fontela, E. (1973). Perceptions of the world problematique: Systems with Applications, 38(1), 458–474.
Communication procedure, communicating with those bearing collective Rao, P., & Holt, D. (2005). Do green supply chains lead to competitiveness and
responsibility. Switzerland Geneva: Battelle Geneva Research Centre. no. 1.. economic performance? International Journal of Operations & Production
Goosey, M. (2004). End-of-life electronics legislation – An industry perspective. Management, 25(9), 898–916.
Circuit World, 30(2), 41–45. Ramik, J. (2007). A decision system using ANP and fuzzy inputs. International Journal
Gumus, A. T. (2009). Evaluation of hazardous waste transportation firms by using a of Innovative Computing, Information and Control, 3(4), 825–837.
two step fuzzy-AHP and TOPSIS methodology. Expert Systems with Applications, Roghanian, E., Rahimi, J., & Ansari, A. (2010). Comparison of first aggregation and
36, 4067–4074. last aggregation in fuzzy group TOPSIS. Applied Mathematical Modelling, 34(12),
Hall, J. (2000). Environmental supply chain dynamics. Journal of Cleaner Production, 3754–3766.
8(6), 206–225. Saaty, T. L. (1996). The analytic network process. Pittsburgh: RWS Publications.
Handfield, R. B., & Nichols, E. L. (2002). Supply chain redesign. Upper Saddle River, NJ: Sadi-Nezhad, S., & Damghani, K. K. (2010). Application of a fuzzy TOPSIS method
Financial Times Prentice Hall.. base on modified preference ratio and fuzzy distance measurement in
Handfield, R., Walton, S., Sroufe, R., & Melnyk, S. (2002). Applying environmental assessment of traffic police centers performance. Applied Soft Computing,
criteria to supplier assessment: A study of the application of the analytical 10(4), 1028–1039.
hierarchy process. European Journal of Operational Research, 141, 70–87. Salehi, M., & Tavakkoli-Moghaddam, R. (2008). Project selection by using a fuzzy
He, S., Chaudhry, S. S., Lei, Z., & Baohua, W. (2009). Stochastic vendor selection TOPSIS technique. World Academy of Science, Engineering and Technology, 40,
problem: Chance-constrained model and genetic algorithms. Annals of 85–90.
Operations Research, 168, 169–179. Sarkis, J. (2003). A strategic decision framework for green supply chain
Humphreys, P. K., Wong, Y. K., & Chan, F. T. S. (2003). Integrating environmental management. Journal of Cleaner Production, 11(4), 397–409.
criteria into the supplier selection process. Journal of Materials Processing Sarkis, J., Meade, L. M., & Talluri, S. (2004). E-logistics and the natural environment.
Technology, 138, 349–356. Supply Chain Management: An International Journal, 9(4), 303–312.
Hwang, C. L., & Yoon, K. (1981). Multiple attribute decision-making: Methods and Sen, S., Basßligil, H., Sßen, C. G., & Baraçli, H. (2007). A framework for defining both
application. New York: Springer. qualitative and quantitative supplier selection criteria considering the buyer-
Iç, Y. T., & Yurdakul, M. (2010). Development of a quick credibility scoring decision supplier integration strategies. International Journal of Production Research,
support system using fuzzy TOPSIS. Expert Systems with Applications, 37, 46(7), 1825–1845.
567–574. Sevkli, M., Koh, S. C. L., Zaim, S., Demirbag, M., & Tatoglu, E. (2007). An application of
Jabbour, A. B. L. S., & Jabbour, C. J. C. (2009). Are supplier selection criteria going data envelopment analytic hierarchy process for supplier selection: A case
green? Case studies of companies in Brazil. Industrial Management & Data study of BEKO in Turkey. International Journal of Production Research, 45,
Systems, 109, 477–495. 1973–2003.
Kelemenis, A., & Askounis, D. (2010). A new TOPSIS-based multi-criteria approach to Simpson, D. F., & Power, D. J. (2005). Use the supply relationship to develop lean
personnel selection. Expert Systems with Applications, 37(7), 4999–5008. and green suppliers. Supply Chain Management: An International Journal, 10,
Ketchen, D. J., Jr., & Hult, G. T. M. (2007). Bridging organization theory and supply 60–68.
chain management: The case of best value supply chains. Journal of Operations Srivastava, S. K. (2007). Green supply chain management: A state-of-the-art
Management, 25(2), 573–580. literature review. International Journal of Management Reviews, 9(1), 53–80.
Kubokawa, S., & Saito, I. (2000). Manufacturing management strategies for Sun, C.-C., & Lin, G. T. R. (2009). Using fuzzy TOPSIS method for evaluating the
environmental protection: Toward the environmental upgrading of competitive advantages of shopping websites. Expert Systems with Applications,
management and manufacturing systems to cope with environmental laws. 36(9), 11764–11771.
Production Planning & Control, 11(2), 107–112. Theyel, G. (2001). Customer and supplier relations for environmental performance.
Kumar, M., Vrat, P., & Shankar, R. (2006). A fuzzy goal programming approach for Greener Management International, 35, 61–69.
vendor selection problem in a supply chain. International Journal of Production Tsai, W.-H., & Hung, S.-J. (2009). A fuzzy goal programming approach for green
Economics, 101, 273–285. supply chain optimization under activity-based costing and performance
Kuo, R. J., Wang, Y. C., & Tien, F. C. (2010). Integration of artificial neural network evaluation with a value-chain structure. International Journal of Production
and MADA methods for green supplier selection. Journal of Cleaner Production, Research, 47(18), 4991–5017.
18(12), 1161–1170. Tseng, M.-L. (2009a). A causal and effect decision making model of service quality
Kurk, F., & Eagan, P. (2008). The value of adding design-for the-environment to expectation using grey-fuzzy DEMATEL approach. Expert Systems with
pollution prevention assistance options. Journal of Cleaner Production, 16(6), Applications, 36, 7738–7748.
722–726. Tseng, M.-L. (2009b). Using the extension of DEMATEL to integrate hotel service
Lee, A. H. I., Kang, Y., Hsu, H. C-F., & Hung, H.-C. (2009). A green supplier selection quality perceptions into a cause-effect model in uncertainty. Expert Systems
model for high-tech industry. Expert Systems with Applications, 36, 7917–7927. with Applications, 36, 9015–9023.
 G. Büyüközkan, G. Çifçi / Expert Systems with Applications 39 (2012) 3000–3011 3011

Tuzkaya, U. R., & Önüt, S. (2008). A fuzzy analytic network process based approach Wu, W.-W., & Lee, Y.-T. (2007). Developing global managers’ competencies using
to transportation-mode selection between Turkey and Germany: A case study. the fuzzy DEMATEL method. Expert Systems with Applications, 32, 499–507.
Information Sciences, 178, 3133–3146. Ye, F. (2010). An extended TOPSIS method with interval-valued intuitionistic fuzzy
Tuzkaya, G., Ozgen, A., Ozgen, D., & Tuzkaya, U. R. (2009). Environmental numbers for virtual enterprise partner selection. Expert Systems with
performance evaluation of suppliers: A hybrid fuzzy multi-criteria decision Applications, 37(10), 7050–7055.
approach. International Journal of Environmental Science and Technology, 6, Yong, D. (2006). Plant location selection based on fuzzy TOPSIS. International Journal
477–490. of Advanced Manufacturing Technology, 28, 839–844.
Vinodh, S., Ramiya, R. A., & Gautham, S. G. (2011). Application of fuzzy analytic _ & Dağdeviren, M. (2010). Using the fuzzy analytic network process (ANP)
Yüksel, I.,
network process for supplier selection in a manufacturing organisation. Expert for balanced scorecard (BSC): A case study for a manufacturing firm. Expert
Systems with Applications, 38(1), 272–280. Systems with Applications, 37(2), 1270–1278.
Vachon, S., & Klassen, R. D. (2006). Green project partnership in the supply chain: Zadeh, L. A. (1965). Fuzzy sets. Information & Control, 8, 338–353.
The case of the package printing industry. Journal of Cleaner Production, 14, Zhu, Q., Sarkis, J., & Lai, K.-H. (2007). Green supply chain management: pressures,
661–671. practices and performance within the Chinese automobile industry. Journal of
van Hock, R. I., & Erasmus (2000). From reversed logistics to green supply chains. Cleaner Production, 15(11–12), 1041–1052.
Logistics Solutions, 2, 28–33. Zhu, Q., Sarkis, J., & Lai, K.-H. (2008). Green supply chain management implications
Walker, H., Sisto, L. D., & McBain, D. (2008). Drivers and barriers to environmental for closing the loop. Transportation Research Part E: Logistics and Transportation
supply chain management practices: Lessons from the public and private Review, 44(1), 1–18.
sectors. Journal of Purchasing & Supply Management, 14, 69–85.
Wu, D. (2009). Supplier selection: A hybrid model using DEA, decision tree and
neural network. Expert Systems with Applications, 36, 9105–9112.