Sei sulla pagina 1di 21

The current issue and full text archive of this journal is available at

www.emeraldinsight.com/0263-5577.htm

IMDS
108,1 Hybrid analytical hierarchy
process model for supplier
selection
122
Mehmet Sevkli
Department of Industrial Engineering, Fatih University, Istanbul, Turkey
Received 24 January 2007
Revised 2 August 2007 S.C. Lenny Koh
Accepted 5 September 2007
Logistics and Supply Chain Management Research Group,
Management School, University of Sheffield, Sheffield, UK
Selim Zaim
Department of Management, Fatih University, Istanbul, Turkey
Mehmet Demirbag
Management School, University of Sheffield, Sheffield, UK, and
Ekrem Tatoglu
Faculty of Business Administration, Bahcesehir University, Istanbul, Turkey

Abstract
Purpose – This paper aims to propose a new approach called “analytical hierarchy process weighted
fuzzy linear programming model (AHP-FLP)” for supplier selection.
Design/methodology/approach – A hybrid method of supplier selection, AHP-FLP is applied to a
real industry case. The weights of the various criteria, taken as local weights from a given judgment
matrix, are calculated using analytical hierarchy process (AHP) that are also considered as the weights
of the fuzzy linear programming model. This new model is compared with the classical AHP method.
Findings – This study concluded that the AHP-FLP method outperforms the AHP method for supplier
selection with respect to restricted supplier selection criteria. Drawing on a real case, Supplier 1 was
identified to be the best supplier through the AHP model under no restrictions, which contradicts the
finding that Supplier 2 was selected as the best supplier by the AHP-FLP model subject to constraints.
Research limitations/implications – More research is definitely called for within the context of
studying a more complex supply chain with multiple supply network and nodes. There is also a crucial
need for investigating other hybrid methods to find the optimum supplier.
Practical implications – The findings of this study indicate that the weights of supplier selection
criteria calculated by the AHP-FLP model are in line with the actual supplier selection decision of
purchasing managers. Since the AHP-FLP model is relatively more difficult to implement compared
with the crisp AHP, its application will be more appropriate for high-value components where
stringent purchasing criteria are required. In contrast, AHP remains an appropriate approach for
relatively lower value components (C class).
Originality/value – The novelty of this study lies in the application of a hybrid approach to a real
Industrial Management & Data
industry case. This study has dealt with one of the most important subjects in supply chain management,
Systems providing a better decision for supplier selection using appropriate quantitative techniques.
Vol. 108 No. 1, 2008
pp. 122-142 Keywords Analytical hierarchy process, Fuzzy logic, Decision making, Supply chain management,
q Emerald Group Publishing Limited Supplier evaluation
0263-5577
DOI 10.1108/02635570810844124 Paper type Research paper
1. Introduction Analytical
Supplier selection and evaluation have an important role in the supply chain process hierarchy
and are crucial to the success of a manufacturing firm (Hartley and Choi, 1996;
Deagraeve et al., 2000). This is because the cost and quality of good and services sold process
are directly related to the cost and quality of goods and services purchased.
Traditionally, vendors are selected from among many suppliers on their ability to meet
the quality requirements, delivery schedule, and the price offered. 123
In this approach, suppliers aggressively compete with each other. The relationship
between buyer and seller is usually adversarial. This traditional purchasing approach
places special emphasis on the commercial transaction between supplier and customer.
The main purchasing objective in this approach is to obtain the lowest possible price
by creating strong competition between suppliers, and negotiating with them.
However, in the modern business world, many firms prefer a strategy of few suppliers
(Chandra and Kumar, 2000). This strategy implies that a buyer wants to have a
long-term relationship and the cooperation of a few dedicated suppliers. Using few
suppliers can create value to the buyer and yield both lower transaction and production
costs (Koh et al., 2007). Cooperation between buyer and supplier is the starting point to
establish a successful supply chain management and a necessary, but insufficient
condition. The next level requires coordination and collaboration between buyer and
suppliers. Collaboration between buyer and supplier includes specified work-flow,
sharing information through electronic data interchange and the internet, and joint
planning and other mechanisms that permit to carry out the just in time ( JIT) system
and total quality management in the company (Spekman et al., 1998).
After implementing advanced concepts in material management, quality
management, logistics, and achieving JIT objectives, a company needs to work with
specialized suppliers in producing the right quality product. Therefore, the supplier
selection process is a multi-objective decision, encompassing many tangible and
intangible factors in a hierarchical manner.
From a viewpoint of company or a member of a supply chain, there exists
competition with other supply chains in order to secure suitable supplies and
deliveries. This competition necessitates selecting carefully suitable suppliers for
collaboration. Various factors have been used as criteria for supplier selection
including price, delivery performance, reputation in the industry, size of enterprise,
geographical location, quality, environmental compliance, capacity, services, lead-time,
packaging, transportation storage, and product development. The applicability of
these criteria depends on the product or service produced and the market for which
these products or services is targeted. To this end, extensive research focuses on
developing methods to assist in supplier selection. Nevertheless, relatively little work
has been undertaken on rationalizing the real industrial applications of many of the
supplier selection methods. To rectify this imbalance, this paper applies a hybrid
method of supplier selection, analytical hierarchy process weighted fuzzy linear model
(AHP-FLP) to a well-known Turkish company operating in appliance industry.
The remainder of the paper is organized as follows. Next section provides a review
of relevant literature. Section 3 briefly discusses the methodologies of AHP and
AHP-FLP. This is then followed by the application of AHP-FLP method to a real
industry case. Conclusions and managerial implications are in the final section.
IMDS 2. Background literature
108,1 There exists a plethora of research on the supplier selection process. Traditional
methodologies of the supplier selection process in the extant literature range from single
objective techniques such as the cost-ratio method, linear or mixed integer programming
to goal and multi-objective linear programming models (Ghodsypour and O’Brien, 1998;
Yan et al., 2003; Oliveria and Lourenço, 2002). Although these methods have been widely
124 used in the area of supplier selection, there are certain drawbacks associated with the
implementation of these methods. One serious problem is that both single and multiple
objective programming applied to supplier selection have shortcomings due to inclusion
of qualitative criteria that are in fact highly important in supplier selection decisions.
Apart from these traditional methods for supplier selection, recently fuzzy systems
theory has been successfully applied to supplier selection problems (Kahraman et al.,
2003, 2004; Chan and Kumar, 2007). Two kinds of justification can be provided for
increasing popularity of the application of fuzzy systems theory into supply chain
management. First, fuzziness must be introduced so as to obtain a reasonable model to
solve the complex problems of supplier selection. Second, there is a need to formulate
human knowledge in a systematic manner and put it into mathematical models.
While several supplier selection methods have been identified and widely applied in
the industry, industrialists and academics differ in their approach to the study of
methods for supplier selection. Industrialists take a relatively more practical approach
than academics.
The use of hybrid method for supplier selection is not new. Wang et al. (2005) have
developed a decision-based methodology for supply chain design that a plant manager
can use to select suppliers. This methodology derived from the techniques of analytical
hierarchy process (AHP) and pre-emptive goal programming. AHP is a widely adopted
decision support technique in management research. For example, the applications of
AHP can be found in evaluating risk factors in enterprise resource planning
implementation and in translating knowledge of supply chain uncertainty (Koh and
Tan, 2006). Recently, Haq and Kannan (2006) developed an integrated supplier
selection and multi-echelon distribution inventory model for the original equipment
manufacturing company in a built-to-order supply chain environment using fuzzy
AHP and a genetic algorithm. Owing to its advantages of implementing a hybrid
method, researchers attempted to address the question of paucity of research with real
industrial applications through undertaking a survey on supplier evaluation within a
multinational telecommunications company. They indicated that the proposed supplier
selection methodology would indeed assist in reducing the product development
timeframe as it automates the evaluation process and provides the procurement team
with a flexible and responsive tool for assessing prospective suppliers. The assessment
tool includes four types of indices to measure supplier involvement in design, namely
satisfaction index, flexibility index, risk index, and confidence index. These indices,
nonetheless, measure the extent to which both the customer requirements and the
supplier capabilities match or mismatch and therefore reflect the potential or risk of
signing a project contract. It may be noted that these indices are limited in
measurement nature, and such supplier selection method was not conducted using
established quantitative approaches. In a more recent work, Sevkli et al. (2007) applied
the data envelopment AHP methodology developed by Ramanathan (2006) in a real
industry application.
It is generally noted that the practical methods used in industry have not been Analytical
evaluated, while at the same time the theoretical methods developed by academics hierarchy
have not been applied in industry. To rectify this imbalance, a hybrid method –
analytical hierarchy process weighted fuzzy multi-objective linear programming model process
(AHP-FLP) has been employed in this study to solve supplier selection problems of an
appliance manufacturer based in Turkey. Supplier selection is a multi-objective
decision-making problem, in which selection criteria should have different weights. 125
Uncertainty of the information in this problem along with inherent difficulties related
to human knowledge make the decision making highly complicated. Assigning
different weights to various criteria, a fuzzy multi-objective model enables the decision
makers to consider the vagueness of information (Fu et al., 2006). Although the purpose
of crisp AHP is to capture the expert’s knowledge, the traditional AHP still may not
reflect the human thinking style (Kahraman et al., 2003). Therefore, AHP methodology
integrated with the fuzzy multi-objective linear programming model has been adopted
as an alternative to the conventional and singular methods of weight derivation in
AHP. This paper attempts to address the shortfalls identified earlier based on a real
case application of AHP-FLP method for supplier selection.

3. Methodology
3.1 The analytic hierarchy process (AHP) methodology
The AHP methodology, which was developed by Saaty (1980), is a powerful tool in
solving complex decision problems. The AHP helps the analysts to organize the
critical aspects of a problem into a hierarchical structure similar to a family tree. By
reducing complex decisions to a series of simple comparisons and rankings, then
synthesizing the results, the AHP not only helps the analysts to arrive at the
best decision, but also provides them with a clear rationale for the choices made (Chin
et al., 1999).
In the AHP approach, the decision problem is structured hierarchically at different
levels with each level consisting of a finite number of decision elements. The upper
level of the hierarchy represents the overall goal, while the lower level consists of all
possible alternatives. One or more intermediate levels embody the decision criteria and
sub-criteria (Partovi, 1994).
The weights of the criteria and the scores of the alternatives, which are called local
priorities, are considered as decision elements in the second step of the decision
process. The decision maker is required to provide his preferences by pairwise
comparisons, with respect to the weights and scores. The values of the weights vi and
scores rij are elicited from these comparisons and represented in a decision table. The
last step of the AHP aggregates all local priorities from the decision table by a
weighted sum of the type, as shown below:
X
Rj ¼ vi £ r ij
i

The global priorities Rj thus obtained are finally used for ranking of the alternatives
and selection of the best alternative. The first and the last steps of the AHP are
relatively simple and straightforward, while the assessment of local priorities based on
pairwise comparisons is the main constituent of this method. The pairwise comparison
in the AHP assumes that the decision maker can compare any two elements Ei, Ej at the
IMDS same level of the hierarchy and provide a numerical value aij of the ratio of
their importance. If the element Ei is preferred to Ej then aij . 1. Correspondingly, the
108,1 reciprocal property aji ¼ 1/aij, j ¼ 1, 2, 3,. . . n and i ¼ 1, 2, 3,. . . n always holds.
Each set of comparisons for a level with n elements requires [n £ (n 2 1)]/2
judgments. The second half of the comparison matrix is the reciprocals of those
judgments lying above the diagonals and are usually omitted. Judgments are provided
126 by means of a nine-point ratio scale that ranges from two factors being equally
important to one of the factors being absolutely more important than the others. After
the expert evaluates the supplier selection criteria, local priorities of each element are
calculated (Tung and Tang, 1998). A local priority vector w ¼ (w1, w2, w3, . . . wn)T may
be obtained from the comparison matrix by applying some prioritization techniques,
such as the eigenvalue method or the logarithmic least squares method. The set of n
relative priorities should be normalized to sum of one:

X
n
wi ¼ 1 wi . 1 and i ¼ 1; 2; 3; . . .n
i¼1

So the number of independent local priorities is (n 2 1).


When the decision maker is perfectly consistent in his answers to pairwise
comparison questions then all elements aij have perfect values, aij ¼ wi/wj. In this case,
aij ¼ aikakj for all i, j, k ¼ 1, 2, 3, . . . n.
In most practical situations, the decision maker’s evaluations aij are not consistent,
since they are only estimations of the exact but unknown ratios wi/wj. However, when
there is a substantial consistency of the decision maker’s preferences, the eigenvalue
method provides satisfactory solution.
Saaty (1980) states that in many practical cases the pairwise judgments of decision
makers will contain some degree of uncertainty. It is usually the case that the
decision maker is certain about the ranking order of the comparison elements but
uncertain about the precise numerical values of his judgments. The classical AHP
attempts to overcome this problem by introducing a discrete linguistic set of
comparison judgments. Instead of directly assigning numerical values to
the comparison ratios, the decision maker chooses an appropriate linguistic phrase,
best corresponding to his comparison preferences.

3.2 Analytical hierarchy process weighted fuzzy linear model (AHP-FLP)


For supplier selection problems, the collected data does not behave crisply and they are
typically fuzzy in nature. Bellman and Zadeh (1970) suggested a fuzzy programming
model for decision making in fuzzy environment. Later, their method was first used by
Zimmermann (1978) to solve fuzzy multi-objective linear programming problems. In
this subsection, the general fuzzy multi-objective model for supplier selection is
presented in the following manner (Tiwari et al., 1987; Amid et al., 2006).
Find a vector X, X ¼ [x1, x2, x3, . . . xn] which maximizes the supplier performance
using objective function Z~ k with for number of m criteria:

X
n
max Z~ k ¼ ðcki £ xi Þ $, Z 0k k ¼ 1; 2; 3; . . . ; m ð1Þ
i¼1
and constraints: Analytical
X
n hierarchy
ari £ xi # br ð2Þ process
i¼1

where cki, ari, and br are crisp values.


In this model, the sign (, ) indicates the fuzzy environment. The symbol ($ , ) 127
denotes the fuzzified version of $ and has linguistic interpretation “essentially greater
than or equal to.” Z 0k is the aspiration level that the decision maker wants to reach.
Every objective function value, Z~ k , changes linearly from Z kmin to Z kmax . So it may be
considered as a fuzzy number with the linear membership function as shown in
Figure 1.
Z kmin and Z kmax are obtained through solving the multi-objective problem as a single
objective.
Based on linear membership function, maximization goals ðZ~ k Þ are given as follows:
8
>
> 1 for Z k $ Z max k
>
>
< max
  max min

mzk ðxÞ ¼ Z k 2 Z k ðxÞ = Z k 2 Z k for Z k , Z k , Z kmax
min
ð3Þ
>
>
>
>
:0 for Z k # Z kmin

The model formulated in equations (1) and (2) can be solved using weighted additive
model which is widely used in vector-objective optimization problems; the basic
concept is to use a single utility function to express the overall preference of decision
maker draw out the relative importance of criteria (Lai and Hawang, 1994). In this
approach, multiplying each membership function of fuzzy goals by their
corresponding weights and then adding the results together obtain a linear
weighted utility function. The weighted additive model proposed by Sakawa (1993)
is equivalent to solving the following crisp single objective programming model
(Zimmermann, 1978):
X
m
Max ðwk £ lk Þ ð4Þ
k¼1

mzk(x)

Figure 1.
Maximizing objective
Z kmin Z kmax function as fuzzy number
IMDS subject to:
108,1 lk # mzk ðxÞ ð5Þ
lk [ ½0; 1 and k ¼ 1; 2; 3; . . . ; m ð6Þ
X
m

128 wk ¼ 1; wk $ 0 ð7Þ
k¼1

xi $ 0; i ¼ 1; 2; 3; :::; n ð8Þ
where wk and mzk ðxÞ represent the solution of membership function, weighting
coefficients that present the relative importance among the fuzzy goals and
membership function of objective function.
Overall, formulation of this model can be expressed in the following stages:
(1) Stage 1. Supplier selection criteria are determined and the hierarchical structure
of the supplier selection is developed.
(2) Stage 2. Involves the weight calculation of each level to obtain the overall score
of each supplier with respect to all criteria and pairwise comparisons of the
main selection criteria.
(3) Stage 3. Constructing the supplier selection model according to the criteria,
constraint and suppliers.
(4) Stage 4. Finding the lower bound ðZ kmin Þ and upper bound ðZ kmax Þ to solve
multi-objective supplier selection problem as a single objective linear
programming model.
(5) Stage 5. Using Z kmin and Z kmax values to find the membership function for each
criterion in equation (3).
(6) Stage 6. Based on AHP weighted additive model, we formulate the equivalent
crisp model of the fuzzy optimization problem according to equations (4)-(8).
(7) Stage 7. Finding the optimal solution vector X, where X is the efficient solution
of the original multi-objective supplier selection problem.

4. Application of the AHP-FLP model


The objective of this study is to develop a hybrid model, which will help to solve the
supplier selection problems of a major Turkish appliance manufacturer, BEKO, which is
the second leading company in Europe in terms of TV set manufacturing and the major
appliance subsidiary of Turkey’s largest conglomerate, Koç Group. In 2005, being the
only Turkish group listed in Fortune’s Global 500, Koç operates in the automotive,
durable goods, food, retailing, energy, financial services, tourism, construction, and IT
industries. Koç Group has consolidated 118 companies, 87,000 employees and 12,000
dealers as well as agencies and after-sales services, generating $34.4 billion in revenues
as of 2006. BEKO has a large portfolio of appliances including nearly 400 product types
ranging from white goods, electronics, and vacuum cleaners to mobile phones and air
conditioners. BEKO, the first brand in Turkey ever to export its products under its own
brand, began this journey with the objective to become a major player in the global white
goods industry. Having achieved a great deal in the domestic market first, BEKO has
now managed to introduce its brand to millions of consumers in more than 100 countries
worldwide reaching consolidated sales volume of e3 billion as of 2005 (www.beko.com. Analytical
tr). Specifically, the study was undertaken to solve the supplier selection problem of hierarchy
BEKO for TV tube purchasing. Global TV tube manufacturing is a highly concentrated
industry where there are only few global suppliers accounting for nearly 90 percent process
of world wide sales. This study was undertaken on three major TV tube suppliers of
BEKO. The main reasons for selecting TV tube suppliers are twofold. First, TV tube is of
high-value component (A class) nature and also the most expensive supply item within 129
this category. Second, it has the longest lead time in all supply items for TV set
manufacturing. In order to maintain the confidentiality of the supplier companies, they
will be numbered as suppliers 1, 2, 3. The model presented in this study utilizes the AHP
and AHP-FLP approaches comparatively.

4.1 Definition of supplier selection criteria


Supplier selection criteria were determined based on the review of prior literature
(Barbarosoglu and Yazgac, 1997; Braglia and Petroni, 2000; Tam and Tummala, 2001;
Masella and Rangone, 2000) and semi-structured interviews undertaken with
22 managers from relevant departments including purchasing, manufacturing,
quality assurance, and R&D. Figure 2 shows the structuring of the hierarchy of
supplier selection problem, which includes four levels. The top level of the hierarchy
represents the ultimate goal of the problem, while the second level of the hierarchy
consists of six main supplier selection criteria, which are namely performance
assessment, human resources, quality system assessment, manufacturing, business,
and information technology. At the third level, these criteria are decomposed into
various sub-criteria that may affect the buyer’s choice for a particular supplier. Finally,
the bottom level of the hierarchy represents the alternative suppliers.
Supplier Evaluation

Performance Assessment Human Resources Quality System Assessment Manufacturing Business Criteria Information Technology

Shipment Number of Emplyees Management Commitment Production Capacity Reputation RFID

Supplier 1 Supplier 1 Supplier 1 Supplier 1 Supplier 1 Supplier 1


Supplier 2 Supplier 2 Supplier 2 Supplier 2 Supplier 2 Supplier 2
Supplier 3 Supplier 3 Supplier 3 Supplier 3 Supplier 3 Supplier 3
Delivery Organizational Structure Inspection and Control Maintenance Location EDI

Supplier 1 Supplier 1 Supplier 1 Supplier 1 Supplier 1 Supplier 1


Supplier 2 Supplier 2 Supplier 2 Supplier 2 Supplier 2 Supplier 2
Supplier 3 Supplier 3 Supplier 3 Supplier 3 Supplier 3 Supplier 3
Cost Training Quality Planning Lead- Time Price Internet

Supplier 1 Supplier 1 Supplier 1 Supplier 1 Supplier 1 Supplier 1


Supplier 2 Supplier 2 Supplier 2 Supplier 2 Supplier 2 Supplier 2
Supplier 3 Supplier 3 Supplier 3 Supplier 3 Supplier 3 Supplier 3
Number of Technical Staff Quality Assurance Up to Date Patent

Supplier 1 Supplier 1 Supplier 1 Supplier 1


Supplier 2 Supplier 2 Supplier 2 Supplier 2
Supplier 3 Supplier 3 Supplier 3 Supplier 3
Storage Technical Capability

Supplier 1 Supplier 1
Supplier 2 Supplier 2
Supplier 3 Supplier 3
Development Figure 2.
Supplier 1 Structure of the supplier
Supplier 2 selection problem
Supplier 3
IMDS 4.2 Calculation of the weights of the criteria
First, the hierarchical structure of the supplier selection has been identified based on
108,1 the evaluations of our responding managers from the buyer company. They also
indicated their degree of preference between and within the criteria at each level of the
hierarchy in a pairwise form using Saaty’s scales ranging from 1 – equally preferred to
9 – extremely preferred. Next step involves the weight calculation of each level to
130 obtain the overall score of each supplier with respect to all 25 sub-criteria and pairwise
comparisons of the main selection criteria. Table I presents the local weights of each
supplier with respect to main supplier selection criteria. Each supplier’s local
weight related to the respective selection criterion, as shown in Table I, is taken as an
objective function coefficient in multi-objective linear programming model. The
detailed calculation of the local weights using AHP methodology is provided in the
Appendix.

4.3 Constructing multi-objective linear programming model


This stage involves construction of multi-objective linear programming model as a
single objective supplier selection problem using each time only one objective.
The multi-objective linear programming of our application is presented as max
Z1 to Z6 :
Max Z 1 ¼ 0:482x1 þ 0:279x2 þ 0:239x3
Max Z 2 ¼ 0:587x1 þ 0:285x2 þ 0:128x3

Max Z 3 ¼ 0:738x1 þ 0:179x2 þ 0:083x3


Max Z 4 ¼ 0:311x1 þ 0:507x2 þ 0:181x3

Max Z 5 ¼ 0:231x1 þ 0:411x2 þ 0:358x3


Max Z 6 ¼ 0:429x1 þ 0:429x2 þ 0:143x3
subject to:
x1 þ x2 þ x3 ¼ 1
x1 ; x2 ; x3 $ 0

Performance Human Quality system Manufacturing Business Information


assessment resources assessment criteria criteria technology

Supplier
1 (x1) 0.482 0.587 0.738 0.311 0.231 0.429
Supplier
2 (x2) 0.279 0.285 0.179 0.507 0.411 0.429
Supplier
Table I. 3 (x3) 0.239 0.128 0.083 0.181 0.358 0.143
Input data for supplier Row
selection averages 0.244 0.055 0.096 0.153 0.420 0.033
4.4 Finding the lower and upper bound for each main criterion Analytical
Then, the linear membership function is used for fuzzifying the objective functions and hierarchy
the constraint for the above problem. The data set for the values of the lower bounds
ðZ kmin Þ and upper bounds ðZ kmax Þ of the objective functions are provided in Table II. process
4.5 Finding fuzzy multi-objective model
In this stage, the membership functions for six objective functions and the constraint 131
are provided by which to maximize the performance of suppliers related to each main
supplier selection criterion. To exemplify, we take the performance assessment criteria
to show the membership function of Z1.
Membership function of Z1 (performance assessment), which is shown in Figure 3, is
computed according to the equation (3):
8
>
> 1 for Z 1 $ 0:482
<
mz1 ðxÞ ¼ ð0:482 2 Z 1 ðxÞÞ=ð0:482 2 0:239Þ for 0:239 , Z 1 ðxÞ , 0:482 ð9Þ
>
>
: 0 for Z 1 # 0:239
The fuzzy multi-objective formulation of the application is shown below.
Find a vector X, X ¼ ½x1 ; x2 ; x3  to satisfy:

max Z~ 1 ¼ 0:482x1 þ 0:279x2 þ 0:239x3 $, Z 01

max Z~ 2 ¼ 0:587x1 þ 0:285x2 þ 0:128x3 $, Z 02

Z kmin (m ¼ 0) Z kmax (m ¼ 1)

Z1 – performance assessment 0.239 0.482


Z2 – human resources 0.128 0.587
Z3 – quality system assessment 0.083 0.738
Z4 – manufacturing criteria 0.181 0.507 Table II.
Z5 – business criteria 0.231 0.411 Data set for the
Z6 – information technology 0.143 0.429 membership functions

1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
Figure 3.
0.1
Membership function of
0 performance assessment
0.239 0.482
IMDS max Z~ 3 ¼ 0:738x1 þ 0:179x2 þ 0:083x3 $, Z 03
108,1
max Z~ 4 ¼ 0:311x1 þ 0:507x2 þ 0:181x3 $, Z 04

max Z~ 5 ¼ 0:231x1 þ 0:411x2 þ 0:358x3 $, Z 05


132 max Z~ 6 ¼ 0:429x1 þ 0:429x2 þ 0:143x3 $, Z 06
subject to:
x1 þ x2 þ x3 ¼ 1
x1 ; x2 ; x3 $ 0

4.6 Developing AHP-FLP model


In our application, weights (wk) associated with the kth objective are taken from
pairwise comparisons of the main selection criteria using AHP which are provided in
Table I, as “row averages.” It can be noted from Table I, that the total weights are equal
to 1. These weights can also be called as “decision maker’s preference.”
Based on AHP weighted additive model (4)-(8), the crisp single objective programming
model, equivalent to the defined fuzzy model above, can be stated as follows:
Max 0:244l1 þ 0:055l2 þ 0:096l3 þ 0:153l4 þ 0:420l5 þ 0:033l6
subject to:
0:482 2 ð0:482x1 þ 0:279x2 þ 0:239x3 Þ
l1 #
0:482 2 0:239
0:587 2 ð0:587x1 þ 0:285x2 þ 0:128x3 Þ
l2 #
0:587 2 0:128
0:738 2 ð0:738x1 þ 0:179x2 þ 0:083x3 Þ
l3 #
0:738 2 0:083
0:507 2 ð0:311x1 þ 0:507x2 þ 0:181x3 Þ
l4 #
0:507 2 0:181
0:411 2 ð0:231x1 þ 0:411x2 þ 0:358x3 Þ
l5 #
0:411 2 0:231
0:429 2 ð0:429x1 þ 0:429x2 þ 0:143x3 Þ
l6 #
0:429 2 0:143

4.7 Solving AHP-FLP model


The linear programming software Xpress MP is used to solve this problem. The
optimal solution is obtained as follows:
x1 ¼ 0, x2 ¼ 1 and x3 ¼ 0, suggesting that Supplier 2 is the best choice according to
decision maker’s preferences.
Objective (Zk) and membership ðmzk ðxÞ or lk Þ function values are obtained as
follows:
Z 1 ¼ 0:279; Z 2 ¼ 0:285; Z 3 ¼ 0:179; Z 4 ¼ 0:507; Z 5 ¼ 0:411; Z 6 ¼ 0:429
mz1 ðxÞ ¼ l1 ¼ 0:165; mz2 ðxÞ ¼ l2 ¼ 0:342; mz3 ðxÞ ¼ l3 ¼ 0:147; mz4 ðxÞ ¼ l4 ¼ 1:000, Analytical
mz5 ðxÞ ¼ l5 ¼ 1:000; mz6 ðxÞ ¼ l6 ¼ 1:000, Membership functions values represent hierarchy
that the achievement levels of Z4 – manufacturing criteria, Z5 – business criteria, and
Z6 – information technology are more than Z1 – performance assessment, Z2 – human process
resources and Z3 – quality system assessment. In other words, the achievement level of
the objective functions correspond with the priority of supplier selection criteria (based
on decision maker’s preferences) indicating that Supplier 2 is selected as the optimum 133
supplier.

4.8 Comparing the AHP and AHP-FLP results


Table III shows the overall scores of the each supplier using AHP and AHP-FLP. As
noted earlier, Supplier 1 was identified to be the best supplier using the crisp AHP
approach under no restrictions. In this approach, business criteria which include
reputation, location, price, patent and technical capability were found as the most
important criteria for supplier selection, whereas the information technology criteria
which are composed of RFID, EDI and internet were identified as the least important
supplier selection criteria. When AHP-FLP approach that is subject to constraints was
employed, Supplier 2 has been identified as the most suitable supplier. In this approach,
manufacturing criteria which contain production capacity, maintenance, lead time, up to
date technology and transportation were found to be the most important supplier
selection criteria, while the quality system assessment criteria which are comprised of
management commitment, inspection and control, quality planning and quality
assurance, were noted as the least important criteria. The finding that Supplier 2 has
been identified as the most appropriate supplier under AHP-FLP approach also tends to
confirm the views of purchasing managers in our case, supporting our argument that
AHP-FLP approach is somewhat superior to AHP approach.

5. Conclusion and managerial implications


Most supplier selection decisions are made today in increasingly complex
environments where the theory of fuzzy decision making can be of significant use.
Since, there is naturally fuzziness and ambiguity in human judgments, conventional
comparison ratio scale with the crisp number as proposed in Saaty’s AHP fails to
address the problem of fuzziness. In this study, the integration of AHP methodology
with the fuzzy multi-objective linear programming model has been employed as an
alternative to the conventional and singular methods of weight derivation in AHP.
Drawing on a real industry case, this study concludes that AHP-FLP approach
outperforms AHP method for supplier selection with respect to restricted supplier
selection criteria. Of the three suppliers in our industry case, Supplier 1 was identified
to be the best supplier through AHP model under no restrictions, which contradicts the
finding that Supplier 2 was selected as the best supplier by AHP-FLP model subject

AHP approach AHP-FLP approach

Supplier 1 0.379 0.000 Table III.


Supplier 2 0.365 1.000 Comparing the AHP and
Supplier 3 0.256 0.000 AHP-FLP results
IMDS to constraints. These findings also indicate that the weights of supplier selection
108,1 criteria calculated by AHP-FLP model are in line with the actual supplier selection
decision of purchasing managers.
It should also be noted that as AHP-FLP model is relatively more difficult to implement
as compared to crisp AHP, its application will be more appropriate for high-value
components where stringent purchasing criteria are required. In contrast, AHP remains to
134 be an appropriate approach for relatively lower value components (C class).

5.1 Managerial implications


This study offers a number of managerial implications. First, supplier selection
decisions have long become an important component of production and logistics
management. As purchases from outside suppliers may constitute a large proportion of
a product’s costs, suppliers should be carefully compared with each other to determine
their relative strengths and weaknesses. In real industry applications, supplier selection
criteria which may inter alia include quality assessment, price, delivery, capacity, and
flexibility cannot be quantitatively and precisely measured using traditional decision
making tools such as crisp AHP. To overcome this deficiency, fuzzy numbers which
enable to capture decision maker’s subjective evaluation related to supplier selection
criteria are applied to provide accurate and consistent supplier selection decisions.
Secondly, crisp AHP approach is implemented under non-restricted situations in
selecting the best supplier. Given the intensity of global competition, however,
companies are more exposed to different resource constraints, e.g. human capital,
capacity, and manufacturing. Therefore, an optimum supplier selection decision must
focus on finding the best supplier which would most contribute to the well being of the
buyer company under resource constraints. To this end, a fuzzy linear programming
model is integrated to crisp AHP to address the problems of fuzziness in human
judgment and also to take into account resource constraints.
Another implication emerging from this study is related to the differences in the
relative weights of supplier selection criteria found in AHP and AHP-FLP approaches.
Purchasing managers are increasingly realizing that their suppliers have a major
influence on customers’ satisfaction level. Therefore, they should not make their
supplier selection decision-based solely on the price and reputation, which was
essentially the case under the traditional AHP approach. Instead, manufacturing
related supplier selection criteria including lead time, up to date technology,
transportation-storage, new product development and production capacity emerge as
more critical criteria under AHP-FLP approach. In order for buyer company to gain
competitive edge over its rivals, one needs to provide the required amount of product at
the minimal cost at the desired quality level and at the right time to its customers,
which could only be achieved by implementing the most appropriate supplier selection
tool in terms of better responding to uncertainty and resource constraints. In this
regard, hybrid methods such as AHP-FLP would prove more useful than traditional
singular multi-criterion methods.

5.2 Future research agenda


The novelty of this research lies in the application of a hybrid approach to a real
industry case – AHP-FLP method for supplier selection, where none or little has been
done on this subject. This study has dealt with one of the most important subjects in
supply chain management providing better decision for supplier selection using Analytical
appropriate quantitative approaches. Despite the mentioned advantages of the hierarchy
proposed approach for the supplier selection problem, more research is certainly called
for within the context of studying a more complex supply chain with multiple supply process
network and nodes as well as investigating other hybrid methods to find the optimum
supplier selection solution. This research can also be extended by incorporating
additional selection criteria such as risk factors and environmental concerns. Different 135
alternative methodologies such as fuzzy analytic network process, fuzzy TOPSIS and
fuzzy electre can also be implemented to solve supplier selection problems. Finally,
adding more alternative suppliers which encompass both domestic and international
suppliers may serve another avenue for future research, though it may increase
computational difficulties.

References
Amid, A., Ghodsypour, S.H. and O’Brien, C. (2006), “Fuzzy multiobjective linear model for the
supplier selection in a supply chain”, International Journal of Production Economics,
Vol. 104, pp. 394-407.
Barbarosoglu, G. and Yazgac, T. (1997), “An application of the analytic hierarchy process to the
supplier selection problem”, Production & Inventory Management Journal, pp. 14-21,
First quarter.
Bellman, R.G. and Zadeh, L.A. (1970), “Decision making in a fuzzy environment”, Management
Science, Vol. 17 No. 2, pp. 141-64.
Braglia, M. and Petroni, A. (2000), “A quality assurance-oriented methodology for handling
trade-offs in supplier selection”, International Journal of Physical Distribution & Logistics
Management, Vol. 30 No. 2, pp. 96-111.
Chan, F.T.S. and Kumar, N. (2007), “Global supplier development considering risk factors using
fuzzy extended AHP-based approach”, Omega, Vol. 35, pp. 417-31.
Chandra, C. and Kumar, S. (2000), ““Supply chain management in theory and practice: a passing
fad or a fundamental change?”, Industrial Management & Data Systems, Vol. 100 No. 3,
pp. 100-13.
Chin, K.S., Chiu, S. and Tummala, V.M.R. (1999), “An evaluation of success factors using the
AHP to implement ISO 14001- based ESM”, International Journal of Quality & Reliability
Management, Vol. 16 No. 4, pp. 341-61.
Deagraeve, Z., Labro, E. and Roodhooft, F. (2000), “An evaluation of vendor selection models
from a total cost of ownership perspective”, European Journal of Operational Research,
Vol. 125, pp. 34-58.
Fu, H.P., Ho, Y.C., Chen, R.C.Y., Chang, T.H. and Chien, P.H. (2006), “Factors affecting the
adoption of electronic marketplaces”, International Journal of Operations & Production
Management, Vol. 26 No. 12, pp. 1301-24.
Ghodsypour, S.H. and O’Brien, C. (1998), “A decision support system for supplier selection using
an integrated analytical hierarchy process and linear programming”, International Journal
of Production Economics, Vol. 56/57, pp. 199-212.
Haq, A.N. and Kannan, G. (2006), “Design of an integrated supplier selection and multi-echelon
distribution inventory model in a built-to-order supply chain environment”, International
Journal of Production Research, Vol. 44 No. 10, pp. 1963-85.
Hartley, J.L. and Choi, T.Y. (1996), “Supplier development: customer as a catalyst of process
change”, Business Horizons, Vol. 39 No. 4, pp. 37-40.
IMDS Kahraman, C., Cebeci, U. and Ruan, D. (2004), “Multi-attribute comparison of catering service
companies using fuzzy AHP: the case of Turkey”, International Journal of Production
108,1 Economics, Vol. 87, pp. 171-84.
Kahraman, C., Cebeci, U. and Ulukan, Z. (2003), “Multi criteria supplier selection using fuzzy
AHP”, Logistics Information Management, Vol. 16 No. 6, pp. 382-94.
Koh, S.C.L. and Tan, K.H. (2006), “Translating knowledge of supply chain uncertainty into
136 business strategy and actions”, Journal of Manufacturing Technology Management, Vol. 17
No. 4, pp. 472-85.
Koh, S.C.L., Demirbag, M., Bayraktar, E., Tatoglu, E. and Zaim, S. (2007), “The impact of supply
chain management practices on performance of SMEs”, Industrial Management & Data
Systems, Vol. 107 No. 1, pp. 103-24.
Lai, Y.J. and Hawang, C.L. (1994), Fuzzy Multiple Objective Decision Making, Methods and
Applications, Springer, Berlin.
Masella, C. and Rangone, A. (2000), “A contingent approach to the design of vendor selection
systems for different types of co-operative customer/supplier relationship”, International
Journal of Operations & Production Management, Vol. 20 No. 1, pp. 70-84.
Oliveria, R.C. and Lourenço, J.C. (2002), “A multicriteria model for assigning new orders to
service suppliers”, European Journal of Operational Research, Vol. 139, pp. 390-9.
Partovi, F.Y. (1994), “Determining what to benchmark: an analytic hierarchy process approach”,
International Journal of Operations & Production Management., Vol. 14 No. 6, pp. 25-39.
Ramanathan, R. (2006), “Data envelopment analysis for weight derivation and aggregation in the
analytic hierarchy process”, Computers & Operations Research, Vol. 33, pp. 1289-307.
Saaty, T.L. (1980), The Analytical Hierarchy Process, McGraw-Hill, New York, NY.
Sakawa, M. (1993), Fuzzy Sets and Interactive Multiobjective Optimization, Plenum Press,
New York, NY.
Sevkli, M., Koh, S.C.L., Zaim, S., Demirbag, M. and Tatoglu, E. (2007), “An application of data
envelopment analytic hierarchy process for supplier selection: a case study of BEKO in
Turkey”, International Journal of Production Research, Vol. 45 No. 9.
Spekman, E.R., Kamauff, J.W. Jr and Myhr, N. (1998), ““An empirical investigation into supply
chain management”, International Journal of Physical Distribution & Logistics
Management, Vol. 28 No. 8, pp. 630-50.
Tam, M.C.Y. and Tummala, V.M.R. (2001), “An application of the AHP in vendor selection of a
telecommunication system”, Omega, Vol. 29, pp. 171-82.
Tiwari, R.N., Dharmahr, S. and Rao, J.R. (1987), “Fuzzy goal programming-an additive model”,
Fuzzy Sets and Systems, Vol. 24, pp. 27-34.
Tung, S.L. and Tang, S.L. (1998), “A comparison of the Saaty’s AHP and modified AHP for right
and left eigenvector inconsistency”, European Journal of Operational Research, Vol. 116,
pp. 123-8.
Wang, G., Huang, S.H. and Dismukes, J.P. (2005), “Manufacturing supply chain design and
evaluation”, International Journal of Advanced Manufacturing Technology, Vol. 25 Nos 1/2,
pp. 93-100.
Yan, H., Yu, Z. and Cheng, T.C.E. (2003), “A strategic model for supply chain design with logical
constraints: formulation and solution”, Computers & Operations Research, Vol. 30,
pp. 2135-55.
Zimmermann, H.J. (1978), “Fuzzy programming and linear programming with several objective
functions”, Fuzzy Sets and Systems, Vol. 1, pp. 45-55.
Appendix Analytical
hierarchy
A. Comparison of criteria with respect to performance assessment process
DMU Shipment Delivery Cost AHP
Shipment 1 5 6 0.723
Delivery 1/5 1 2 0.174
Cost 1/6 1/2 1 0.103 137
Consistency ratio ¼ 0.0374
DMU Supplier 1 Supplier 2 Supplier 3 AHP
B. Comparison of suppliers with respect to shipment
Supplier 1 1 3 5 0.633
Supplier 2 1/3 1 3 0.260
Supplier 3 1/5 1/3 1 0.106
Consistency ratio ¼ 0.0419
C. Comparison of suppliers with respect to delivery
Supplier 1 1 1/5 1/6 0.082
Supplier 2 5 1 1/2 0.343
Supplier 3 6 2 1 0.575
Consistency ratio ¼ 0.0299
D. Comparison of suppliers with respect to cost
Supplier 1 1 1/3 1/6 0.100
Supplier 2 3 1 1/2 0.300
Supplier 3 6 2 1 0.600 Table AI.
Consistency ratio ¼ 0.0000 Performance assessment

(See overleaf for Table AII)


138
108,1
IMDS

Table AII.
Human resources
A. Comparison of criteria with respect to human resources
DMU No. of employees Organizational structure Training No. of technical staff AHP
No. of employees 1 1/5 1/6 1/3 0.063
Organization structure 5 1 1/2 3 0.309
Training 6 2 1 4 0.492
Number of technical staff 3 1/3 1/4 1 0.136
Consistency ratio ¼ 0.0367
DMU Supplier 1 Supplier 2 Supplier 3 AHP
B. Comparison of suppliers with respect to number of employees
Supplier 1 1 1/3 1/5 0.110
Supplier 2 3 1 1/2 0.309
Supplier 3 5 2 1 0.581
Consistency ratio ¼ 0.0037
C. Comparison of suppliers with respect to organization structure
Supplier 1 1 4 6 0.685
Supplier 2 1/4 1 3 0.221
Supplier 3 1/6 1/3 1 0.093
Consistency ratio ¼ 0.0644
D. Comparison of suppliers with respect to number of training
Supplier 1 1 4 6 0.685
Supplier 2 1/4 1 3 0.221
Supplier 3 1/6 1/3 1 0.093
Consistency ratio ¼ 0.0644
E. Comparison of suppliers with respect to number of technical staff
Supplier 1 1 1/3 2 0.230
Supplier 2 3 1 5 0.648
Supplier 3 1/2 1/5 1 0.122
Consistency ratio ¼ 0.0041
A. Comparison of criteria with respect to quality system assessment
DMU Management commitment Inspection and control Quality planning Quality assurance AHP
Management commitment 1 14 1/4 1/6 0.062
Inspection and control 4 1 1/3 1/3 0.165
Quality planning 4 3 1 1/3 0.270
Quality assurance 6 3 3 1 0.503
Consistency ratio ¼ 0.0944
DMU Supplier 1 Supplier 2 Supplier 3 AHP
B. Comparison of suppliers with respect to management commitment
Supplier 1 1 4 5 0.665
Supplier 2 1/4 1 3 0.231
Supplier 3 1/5 1/3 1 0.104
Consistency ratio ¼ 0.0996
C. Comparison of suppliers with respect to inspection and control
Supplier 1 1 5 7 0.724
Supplier 2 1/5 1 3 0.193
Supplier 3 1/7 1/3 1 0.083
Consistency ratio ¼ 0.0844
D. Comparison of suppliers with respect to quality planning
Supplier 1 1 7 9 0.790
Supplier 2 1/7 1 2 0.133
Supplier 3 1/9 1/2 1 0.077
Consistency ratio ¼ 0.0322
E. Comparison of suppliers with respect to quality assurance
Supplier 1 1 5 7 0.724
Supplier 2 1/5 1 3 0.193
Supplier 3 1/7 1/3 1 0.083
Consistency ratio ¼ 0.0844
Analytical

assessment
process
hierarchy

Quality system
Table AIII.
139
140
108,1
IMDS

Table AIV.
Manufacturing criteria
A. Comparison of criteria with respect to manufacturing criteria
DMU Production capacity Maintenance Lead time Up to date technology Transportation-storage NPD AHP
Production capacity 1 2 3 1/2 4 1/3 0.151
Maintenance 1/2 1 1/2 1/4 2 1/7 0.062
Lead time 1/3 2 1 1/3 2 1/6 0.079
Up to date technology 2 4 3 1 6 1/2 0.241
Transportation-storage 1/4 1/2 1/2 1/6 1 1/9 0.039
NPD 3 7 6 2 9 1 0.428
Consistency ratio ¼ 0.0183
DMU Supplier 1 Supplier 2 Supplier 3 AHP
B. Comparison of suppliers with respect to production capacity
Supplier 1 1 1/5 1/4 0.096
Supplier 2 5 1 3 0.619
Supplier 3 4 1/3 1 0.284
Consistency ratio ¼ 0.0923
C. Comparison of suppliers with respect to maintenance
Supplier 1 1 3 5 0.633
Supplier 2 1/3 1 3 0.260
Supplier 3 1/5 1/3 1 0.106
Consistency ratio ¼ 0.0419
D. Comparison of suppliers with respect to production planning
Supplier 1 1 2 1/4 0.201
Supplier 2 1/2 1 1/5 0.118
Supplier 3 4 5 1 0.681
Consistency ratio ¼ 0.0290
E. Comparison of suppliers with respect to up to date
Supplier 1 1 1 4 0.444
Supplier 2 1 1 4 0.444
Supplier 3 1/4 1/4 1 0.111
Consistency ratio ¼ 0.0000
F. Comparison of suppliers with respect to storage
Supplier 1 1 3 1/3 0.272
Supplier 2 1/3 1 1/4 0.120
Supplier 3 3 4 1 0.608
Consistency ratio ¼ 0.0767
G. Comparison of suppliers with respect to development
Supplier 1 1 1/3 6 0.290
Supplier 2 3 1 8 0.646
Supplier 3 1/6 1/8 1 0.064
Consistency ratio ¼ 0.0840
A. Comparison of criteria with respect to business criteria
DMU Reputation Location Price Patent Technical capability AHP
Reputation 1 2 1/8 1/3 1/7 0.054
Location 1/2 1 1/9 1/5 1/8 0.036
Price 8 9 1 5 2 0.473
Patent 3 5 1/5 1 1/3 0.131
Technical capability 7 8 1/2 3 1 0.306
Consistency ratio ¼ 0.0452
DMU Supplier 1 Supplier 2 Supplier 3 AHP
B. Comparison of suppliers with respect to reputation
Supplier 1 1 3 8 0.646
Supplier 2 1/3 1 6 0.290
Supplier 3 1/8 1/6 1 0.064
Consistency ratio ¼ 0.0840
C. Comparison of suppliers with respect to location
Supplier 1 1 4 5 0.665
Supplier 2 1/4 1 3 0.231
Supplier 3 1/5 1/3 1 0.104
Consistency ratio ¼ 0.0996
D. Comparison of suppliers with respect to price
Supplier 1 1 1/3 1/4 0.120
Supplier 2 3 1 1/3 0.272
Supplier 3 4 3 1 0.608
Consistency ratio ¼ 0.0767
E. Comparison of suppliers with respect to patent
Supplier 1 1 1 1 0.333
Supplier 2 1 1 1 0.333
Supplier 3 1 1 1 0.333
Consistency ratio ¼ 0.0000
F. Comparison of suppliers with respect to technical capability
Supplier 1 1 1/4 5 0.236
Supplier 2 4 1 9 0.701
Supplier 3 1/5 1/9 1 0.062
Consistency ratio ¼ 0.0893
Analytical

Business criteria
process
hierarchy

Table AV.
141
IMDS A. Comparison of criteria with respect to information technology
108,1 DMU RFID EDI Internet AHP
RFID 1 1/6 1/7 0.070
EDI 6 1 1/2 0.350
Internet 7 2 1 0.580
Consistency ratio ¼ 0.0374
DMU Supplier 1 Supplier 2 Supplier 3 AHP
142 B. Comparison of suppliers with respect to RFID
Supplier 1 1 1 3 0.429
Supplier 2 1 1 3 0.429
Supplier 3 1/3 1/3 1 0.143
Consistency ratio ¼ 0.0000
C. Comparison of suppliers with respect to EDI
Supplier 1 1 1 3 0.429
Supplier 2 1 1 3 0.429
Supplier 3 1/3 1/3 1 0.143
Consistency ratio ¼ 0.0000
D. Comparison of suppliers with respect to internet
Supplier 1 1 1 3 0.429
Supplier 2 1 1 3 0.429
Table AVI. Supplier 3 1/3 1/3 1 0.143
Information technology Consistency ratio ¼ 0.0000

Corresponding author
S.C. Lenny Koh can be contacted at: S.C.L.Koh@Sheffield.ac.uk

To purchase reprints of this article please e-mail: reprints@emeraldinsight.com


Or visit our web site for further details: www.emeraldinsight.com/reprints

Potrebbero piacerti anche