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

www.emeraldinsight.com/1741-038X.htm

JMTM
17,7 Supplier selection using combined
analytical hierarchy process and
grey relational analysis
926
Ching-Chow Yang and Bai-Sheng Chen
Department of Industrial Engineering, Chung-Yuan Christian University,
Received January 2005
Revised September 2005 Chung-Li, Taiwan, Republic of China
Accepted November 2005

Abstract
Purpose – The purpose of this paper is to develop an evaluation model considering the qualitative
and quantitative criteria for supplier selection in an outsourcing manufacturing organization.
Design/methodology/approach – In this paper, the authors propose an integrated model by
combining the analytical hierarchy process (AHP) and grey relational analysis (GRA) into a single
evaluation model. The model is illustrated with a case study of a notebook computer manufacturer to
demonstrate the effectiveness of this integrated model.
Findings – Through the proposed model, it is possible to effectively integrate the specialized
knowledge and experience of each dispersed evaluator, and the quantitative data to select the best
suppliers for cooperation.
Originality/value – The decision-aiding software (AHP-GRA model) has been implemented in Excel
to automate the supplier selection process. This can widely apply the integrated model for the
industry.
Keywords Suppliers, Analytical hierarchy process
Paper type Research paper

1. Introduction
Supplier selection is increasingly recognized as a critical decision in supply chain
management in manufacturing industries (Choi and Hartley, 1996; Dahel, 2003). In
such industries, raw materials and outsourced components are usually the two largest
costs, and the procurement department often plays an important role in reducing
purchasing cost and selecting appropriate suppliers. In addition, manufacturers
increasingly practice just-in-time (JIT) and mass customisation manufacturing.
These require suppliers to supply materials and outsourced parts punctually – in the
appropriate quantity and with consistent quality. Manufacturers, therefore, require
suppliers to have effective systems in production management and quality control.
Because reliable suppliers enable manufacturers to reduce inventory costs and
improve product quality, it is understandable that manufacturers are increasingly
concerned about supplier selection (Braglia and Petroni, 2000). It is apparent that the
selection of appropriate suppliers and effective supplier relationship management are
Journal of Manufacturing Technology key factors in raising the competitiveness of firms (Davis, 1993; Hines, 1994;
Management Ghodsypour and O’Brien, 2001; Choy et al., 2003).
Vol. 17 No. 7, 2006
pp. 926-941
q Emerald Group Publishing Limited
1741-038X
The authors would like to thank the National Science Council of Taiwan, ROC, for supporting
DOI 10.1108/17410380610688241 this research (NSC-93-2213-E-033033).
 For these reasons, purchasing departments need to evaluate the performance of Supplier
suppliers periodically in terms of several critical criteria (Mummalaneni et al., 1996). selection
Several studies have discussed the issue of selecting appropriate supplier partners, and
most agree that it is a complicated matter because of the large number of criteria to be
considered (Choy et al., 2002). It is apparent that the selection of long-term suppliers
requires consideration of both quantitative criteria (for example, quality, cost, delivery)
and qualitative criteria (for example, flexibility, service, culture). From the viewpoints 927
of literatures and practices in supplier selection, several problems include:
.
Selecting suppliers only on the basis of evaluators’ personal experiences is
neither effective nor scientific – due to the inherent risks of subjective judgment
and a lack of systematic analysis (Choy et al., 2002).
.
Several evaluation models such as linear-weighting (LW), total cost of ownership
(TCO), and mathematical programming, it is apparent that LW models do not
include quantitative factors, whereas the mathematical-programming (MP)
models have the disadvantage of not including qualitative factors (De Boer et al.,
1998; Ghodsypour and O’Brien, 1998).
. For the practical application of supplier selection model, it lacks of a decision
support system in supporting company to evaluate and select suppliers, avoids
the complexity of the mathematical analysis.

The present paper, therefore, proposes an integrated model based on the integration of
analytical hierarchy process (AHP) and grey relational analysis (GRA). AHP can be
used to systematically integrate various judgments from different evaluators and
obtain the weights of qualitative criteria. By determining the relative weights of
qualitative criteria, then uses these weights as coefficients of GRA model. The
algorithm of GRA is then used to combine the qualitative and quantitative data when
considering its characteristics of “large-is-better” or “smaller-is-better” to perform the
evaluation of supplier selection. A program written in Excel (called AHP-GRA model)
has been developed to automatically calculate the mathematical analysis and obtain
the results. Finally, a real example of a notebook computer manufacturer in Taiwan is
provided to demonstrate the effectiveness of this integrated model.

2. Methods of supplier selection

Three evaluation models for supplier selection feature prominently in the literature:
LW models, total-cost models, and MP models (Table I). Each of these is discussed
below.

2.1 Linear-weighting models

LW models evaluate potential suppliers using several equally weighted factors, and
then allow the decision-maker to choose the supplier with the highest total score
(Timmerman, 1986). Although this method is simple, it depends heavily on subjective
judgment. In addition, these models weight the criteria equally, which rarely happens
in practice (Min, 1994; Ghodsypour and O’Brien, 1998). Min (1994) used multiple
attribute utility theory (MAUT) to deal with the problem of selection of international
suppliers. The criteria used were described as financial, quality, risks, service,
partnerships, cultural and communication, and trade restrictions.
JMTM
LW models Total-cost models MP models
17,7
Timmerman (1986) Monczka and Trecha (1988) Weber and Current (1993)
Smytka and Clemens (1993)
Bhutta and Huq (2002)
Roodhooft and Konings (1996)
928 Nydick and Hill (1992) Pan (1989)
Barabarosoglu and Yazgac (1997) Ghodsypour and O’Brien (1998)
Tam and Tummala (2001) Buffa and Jackson (1983)
Bhutta and Huq (2002) Karpark et al. (1999)
Table I. Handfield et al. (2002) Chaudhry et al. (1993)
Methods of supplier Min (1994) Rosenthal et al. (1995)
selection De Boer et al. (1998) Ghodsypour and O’Brien (2001)

In contrast to the equal weighting utilized in LW models, AHP is an effective method

for providing a structured determination of the weights of criteria by using pairwise
comparison to select the best suppliers. Several researchers have used AHP to deal
with the supplier selection issue. These include Nydick and Hill (1992), Barabarosoglu
and Yazgac (1997), Tam and Tummala (2001), Bhutta and Huq (2002), and Handfield
et al. (2002).
De Boer et al. (1998) proposed an outranking method (ELECTRE I) for supplier
selection. Application of such an outranking method enables the selection of a number
of attractive suppliers based on relatively limited information. To evaluate and select
the suppliers, De Boer et al. (1998) used the quantitative criteria of turnover, distance,
cost, and the qualitative criterion of quality image.

2.2 Total cost of ownership models

TCO models attempt to include the quantifiable costs that are incurred throughout the
purchased item life cycle into the supplier selection model. Monczka and Trecha (1988),
Smytka and Clemens (1993), Roodhooft and Konings (1996), Bhutta and Huq (2002),
and Chen and Yang (2003) all attempted to integrate the total cost into their evaluation
models. Bhutta and Huq (2002) made a comparison between TCO models and AHP for
the supplier selection problem. If cost data are included (as they are in TCO), AHP can
provide a more robust tool for managers to select and evaluate suppliers – thus
enabling them to make sound selections.

2.3 Mathematical-programming models

MP models can be used to formulate the supplier selection problem in terms of an
objective function to be maximized (for example, profit) or minimized (for example,
costs) by varying the values of the variables in an objective function.
Several papers have used single objective techniques. These include linear
programming (Pan, 1989; Ghodsypour and O’Brien, 1998), goal programming (Buffa
and Jackson, 1983; Karpark et al., 1999), or mixed integer programming (Chaudhry
et al., 1993; Rosenthal et al., 1995; Ghodsypour and O’Brien, 2001) to solve the supplier
selection issue. Most of these MP models took cost as the objective function, with other
criteria (such as quality, capacity, delivery, and so on) being taken into account as
constraints. However, the complexity of the MP models are not suitable for companies
that wish to solve the supplier selection issue effectively without utilizing advanced
computer programmers. There are two problems with MP models. The first is that the Supplier
criteria that are considered to be constraints are weighted equally; this rarely happens selection
in practice. The second is that they usually consider only quantitative criteria
(Ghodsypour and O’Brien, 1998; De Boer et al., 2001).
In a review of earlier research, Dickson (1966) identified 23 different criteria as
having been used in the supplier selection process – including quality, delivery,
performance history, price, technical capability, and financial position. Choi and 929
Hartley (1996) analysed the principal components of supplier selection practices in the
US’s auto industry and identified eight factors – finances, consistency, relationship,
flexibility, technological capability, customer service, reliability, and price. The
empirical results showed that the consistency factor (of quality and delivery) was the
most important factor and that finances was the least important factor. To evaluate
and select suppliers, De Boer et al. (1998) used the quantitative criteria of turnover,
distance, and cost, and the qualitative criterion of quality image. A survey by Chao et al.
(1993) and Mummalaneni et al. (1996) concluded that quality and in-time delivery are
the most important criteria. A review of 74 supplier selection articles by Weber et al.
(1991) found that cost, product quality, delivery performance, and supply capacity
were most frequently used as selection criteria. Additionally, the research had shown
that small turnover suppliers have to take on more risky business propositions, if they
are to compete against large turnover suppliers (Nicholas et al., 2000).
To evaluate suppliers effectively, it is necessary to identify quantitative and
qualitative criteria simultaneously (Weber et al., 1991). It is to be noted that quality
is considered as both a quantitative criterion and as a qualitative criterion. This is
because it depends on the perspectives of decision-makers – using “quality image” or
“quality level” (for example, process capability, defects).
From the above literature review, it is apparent that LW models do not include
quantitative factors (such as cost and delivery), whereas the MP models have the
disadvantage of not including qualitative factors (such as finance and customer
service). Many papers used AHP method to deal with supplier selection problem, most
of them using the qualitative criteria in evaluation. However, in this case study, not
only the qualitative criteria, but quantitative criteria (cost, delivery, distance and
turnover) must be considered simultaneously in supplier performance evaluation
and selection. Additionally, the quantitative criteria of cost, distance, and delivery,
therefore, had the characteristic of “smaller-is-better” and the “larger-is-better” for the
criteria of turnover. This integrated model includes:
.
using AHP method to systematically integrate various judgments from different
evaluators and obtain the weights of qualitative criteria; and
.
application of GRA not only integrates the qualitative and quantitative data, but
also considers its characteristic large-is-better or smaller-is-better.

3. Grey relational analysis

Grey system theory originated with Deng (1982). It can be used to solve uncertainty
problems in cases with discrete data and incomplete information (Deng, 1989). It is,
therefore, a theory and methodology that deals with poor, incomplete, or uncertain
systematic problems. One of the major advantages of grey system theory is that it can
generate satisfactory outcomes using a relatively small amount of data or with great
variability in factors. It achieves this by increasing the regularity of data with proper
JMTM treatment (Li et al., 1997). Like fuzzy set theory, grey theory is thus an effective
17,7 mathematical model for resolving uncertain and indeterminate problems (Hsu and
Wen, 2000).
The concept of grey relational space was proposed by Deng (1982) on the basis of
the combined concepts of system theory, space theory, and control theory. It can be
used to capture the correlations between the reference factors and other compared
930 factors of a system (Deng, 1988). GRA analyses uncertain relations between one main
factor and all the other factors in a given system (Liang, 1999). Fields covered by grey
theory include forecasting, system control, data-processing, modelling, and decision
making (Chang et al., 2000; Hsu and Wen, 2000). GRA has been successfully applied to
various decision problems – including performance evaluation of airlines (Feng and
Wang, 2000), generation scheduling of hydroelectricity (Liang, 1999), image
compression (Hsu et al., 2000), and multiple-attribute decision-making problems
(Wu, 2003; Chiou and Tzeng, 2001). The details of the GRA are summarized in
Appendix.

4. Case study
The present study proposes an integrated model for supplier selection by
combining AHP and GRA into a single evaluation model for establishing a decision
support system. To demonstrate the model, a case study is presented to evaluate
the potential suppliers of printed circuit boards (PCBs) for a notebook computer
firm.
According to the existing supplier selection process, each potential supplier had to
be audited by the buyer purchasing department, quality-assurance department, and
engineering department. The PCB suppliers had to satisfy certain requirements. It was
necessary for the supplier to be:
.
located close to the buyer (to satisfy the JIT production system);
.
a major supplier in its markets (with a high annual turnover);
.
able to offer high quality;
.
able to offer low unit cost; and
.
involved in a long-term partnership with the buyer.

Based on market research, there are three candidate suppliers (A, B, and C) which
satisfy the above requirements. The application of the integrated model includes four
steps, as shown in Figure 1.

Figure 1.
Steps in the integrated
model for supplier
selection
4.1 Define qualitative and quantitative criteria for supplier selection Supplier
A literature review of evaluative factors affecting supplier selection (Dickson, 1966; selection
Weber et al., 1991; Chao et al., 1993; Min, 1994; Choi and Hartley, 1996; Mummalaneni
et al., 1996; De Boer et al., 1998), and brainstorming with three department managers,
we identified six qualitative criteria including quality, finance, customer service,
production capacity, design and technical capability, and information technology (IT)
system. The quantitative criteria included turnover, cost, delivery and geographical 931
location. The framework of the supplier selection is shown in Figure 2.

4.2 Use of AHP to obtain relative weights of quantitative criteria

Various firms give different weights to different criteria. The present study used AHP to
determine the relative weight of each criterion. A five-point rating scale was designed
(essential, extremely important, very important, moderately important, and somewhat
important), and priority weights of the criteria were determined using pairwise
comparisons. The AHP evaluation process was conducted after interviewing three
department managers of purchasing, quality assurance, and engineering. A program
written in EXCEL has been developed to aggregate evaluators’ judgments and
calculates the weights. The consistency ratio of the pairwise comparison judgment
matrix was 0.09. This indicated that the judgment matrix had satisfactory consistency.
The weights of the evaluative criteria are presented in Table II. It can be seen that quality
was considered to be the most important criterion in the supplier selection process.

4.3 Evaluation of potential suppliers

The qualitative and quantitative data are shown in Table III. To demonstrate the GRA
model in this supplier selection problem, the suppliers were considered as alternatives i
(i ¼ 1, 2, 3), whereas evaluative factors were viewed as criteria j for each alternative

Figure 2.
Qualitative and
quantitative criteria used
in the supplier selection
process
 17,7

932

Table II.
JMTM

qualitative criteria
Relative weights of
Quality Finance Customer service Production capacity Design and technical capability IT system Turnover Distance Delivery Cost

Weight 0.228 0.019 0.038 0.091 0.095 0.081 0.021 0.030 0.190 0.208
Rank 1 10 7 5 4 6 9 8 3 2
 Customer Production Design and technical IT Turnover Distance Delivery Cost
Quality Finance service capacity capability system (million) (km) (day) ($)

Supplier a 3.87 5.00 4.40 4.00 3.67 5.00 350.03 25.00 11.00 19.80
Supplier b 4.18 3.00 4.00 4.00 3.67 5.00 430.41 15.00 11.50 20.40
Supplier c 3.79 5.00 3.50 4.67 4.50 4.00 88.51 12.00 11.00 20.10
selection

Qualitative and
Supplier

candidate suppliers
quantitative evaluation of
Table III.
933
JMTM ( j ¼ 1, 2, . . . , 10). The higher the weighted value of each qualitative criterion, the
17,7 greater its capability to support the service. In terms of quantitative criteria, the
supplier with the highest turnover was considered pre-eminent. The best geographical
location was deemed to be the one closest to the plant. Short delivery times and low
unit costs are certainly advantages to notebook computer manufacturers. The
qualitative data, therefore, had the characteristic of “larger-is-better”. In the case of
934 quantitative data, turnover was also “larger-is-better”, but distance, delivery, and cost
were “smaller-is-better”.

4.4 Use of GRA to determine best supplier

This procedure used GRA to integrate the qualitative evaluation from three evaluators
using AHP and quantitative data:
(1) Generation of referential series of x0. The qualitative and quantitative evaluation of
three potential suppliers can be represented as a matrix as follows:
2 3
3:87 5:00 4:40 4:00 3:67 5:00 350:03 25:00 11:00 19:80
6 7
Xi ¼ 6
4 4:18 3:00 4:00 4:00 3:67 5:00 430:41 15:00 11:50 20:40 7
5
3:79 5:00 3:50 4:67 4:50 4:00 88:51 12:00 11:00 20:10

Then the referential series of x0 is (4.18, 5, 4.4, 4.67, 4.5, 5, 430.41, 12, 11, 19.8), and the
compared series of x1 ¼ (3.87, 5, 4.4, 4, 3.67, 5, 350.03, 25, 11, 19.80), x2 ¼ (4.18, 3, 4, 4,
3.67, 5, 430.41, 15, 11.5, 20.40) and x3 ¼ (3.79, 5, 3.5, 4.67, 4.5, 4, 88.51, 12, 11, 20.10).
(2) Normalization of data set. The series data in this case can be treated using two
approaches: larger-is-better and smaller-is-better. Therefore, the normalized referential
series of x0 becomes x*0 : The “larger-is-better” data transformation of equation (1) in
Appendix is applied to qualitative criteria and to the “turnover” factor, whereas the
“smaller-is-better” criteria set of “distance”, “delivery”, and “cost” are transformed
using equation (2). The calculations are shown below. According to the calculation
results, the referential series of x*0 becomes (1, 1, 1, 1, 1, 1, 1, 1, 1, 1):

x*1 ð1Þ ¼ 3:8723:79

4:1823:79 ¼ 0:2; x*2 ð1Þ ¼ 4:1823:79
4:1823:79 ¼ 1:0; x*3 ð1Þ ¼ 3:7923:79
4:1823:79 ¼ 0:0

x*1 ð2Þ ¼ 5:0023:00

5:0023:00 ¼ 1:0; x*2 ð2Þ ¼ 3:0023:00
5:0023:00 ¼ 0:0; x*3 ð2Þ ¼ 5:0023:00
5:0023:00 ¼ 1:0

..
.

x*1 ð8Þ ¼ 25:0225:0

25:0212:0 ¼ 0:0; x*2 ð8Þ ¼ 25:0215:0
25:0212:0 ¼ 0:77; x*3 ð8Þ ¼ 25:0212:0
25:0212:0 ¼ 1:0

x*1 ð9Þ ¼ 11:5211:0

11:5211:0 ¼ 1:0; x*2 ð9Þ ¼ 11:5211:5
11:5211:0 ¼ 0:0; x*3 ð9Þ ¼ 11:5211:0
11:5211:0 ¼ 1:0

x*1 ð10Þ ¼ 20:4219:8 * 20:4220:4 * 20:4220:1

20:4219:8 ¼ 1:0; x2 ð10Þ ¼ 20:4219:8 ¼ 0:0; x3 ð10Þ ¼ 20:4219:8 ¼ 0:5

(3) Calculation of the grey relational coefficient g0i ð jÞ: This step is to use the grey relational
equation to calculate the grey relational coefficient by using equation (3). In this
paper, the distinguished coefficient was set at 0.5. For example, r 01 ð1Þ ¼ 0 þ 0:5ð1Þ= Supplier
0:8 þ 0:5ð1Þ ¼ 0:38; r 02 ð1Þ ¼ 0 þ 0:5ð1Þ=0 þ 0:5ð1Þ ¼ 1:0 and r 03 ð1Þ ¼ 0 þ 0:5ð1Þ=1 þ selection
0:5ð1Þ ¼ 0:33: The calculation results are shown in Table IV.
(4) Calculation of degree of grey equation coefficient G0i. The final step is to calculate the
grade of the grey relational coefficient G0i by using equation (4). According to GRA, the
alternative with the highest grey relational grade is the most important (or optimal)
alternative (Wu and Chen, 1999). Therefore, in this study, the priorities of potential 935
suppliers can be ranked in accordance with the grey relational grade values – because
the relative weights (Wi) of evaluative criteria are determined using AHP. The results
of the grey relational grade calculation for each supplier is shown below:

G01 ðaÞ ¼ ½0:087 þ 0:019 þ 0:038 þ 0:030 þ 0:031 þ 0:081 þ 0:014 þ 0:010 þ 0:190

þ 0:208

¼ 0:708

G02 ðbÞ ¼ ½0:228 þ 0:006 þ 0:020 þ 0:030 þ 0:031 þ 0:081 þ 0:021 þ 0:020 þ 0:063

þ 0:069

¼ 0:570

G03 ðcÞ ¼ ½0:075 þ 0:019 þ 0:013 þ 0:091 þ 0:095 þ 0:027 þ 0:007 þ 0:030 þ 0:190

þ 0:104

¼ 0:650

The priority of the three potential suppliers (in accordance with their grey relational
grades) is supplier a . supplier c . supplier b. However, supplier a is the optimal PCB
supplier in the entire supply chain if both qualitative and quantitative criteria are taken
into account.

5. Sensitivity analysis and feasibility of application

The GRA method is a flexible approach to the supplier selection problem when both
qualitative and quantitative criteria have to be taken into account. Evaluators can
revise the referential sequence taking into consideration the nature of criteria – such as
“larger-is-better”, “smaller-is-better” and “nominal-is-best”. For example, if supplier C’s
unit cost is less than $19.88 or delivery time is within 10.5 days (when j ¼ 0.5), the
supplier priority is changed to c . a . b. In addition, the adoption of the AHP method
is an effective method for determining the relative weights of criteria in this case.
In particular, the main advantage of AHP compared with other methods is that it
enables evaluators to represent the relative importance and interaction of multiple
criteria in the supplier selection process. Evaluators can revise the weights of criteria
 17,7

936
JMTM

Table IV.
The values of grey
relational coefficient
Customer Production Design and technical IT Turnover Distance Delivery Cost
Quality Finance service capacity capability system (million) (km) (day) ($)

Weight 0.228 0.019 0.038 0.091 0.095 0.081 0.021 0.030 0.190 0.208
Supplier a 0.38 1.00 1.00 0.33 0.33 1.00 0.68 0.33 1.00 1.00
Supplier b 1.00 0.33 0.53 0.33 0.33 1.00 1.00 0.68 0.33 0.33
Supplier c 0.33 1.00 0.33 1.00 1.00 0.33 0.33 1.00 1.00 0.50
via a pairwise comparison procedure if there are changes in the business environment Supplier
and/or changes in customer requirements. selection
It should be noted that the same conclusion regarding supplier priority (a . c . b)
based on grey relational grade is reached with varying distinguished coefficients
from 0.1 to 1.0, as shown in Figure 3. It can, therefore, be concluded that supplier
selection using the GRA model is stable and reliable. In addition, the proposed model
can effectively combine qualitative and quantitative criteria in the supplier selection 937
process. From the results of the case study, it can be asserted that the AHP and GRA
method is an effective and practicable approach to solving the supplier selection
problem of a notebook computer manufacturer.
In order to simplify the evaluation process of AHP and GRA, a program written in
EXCEL (called AHP-GRA model) has been developed to automatically calculate
the mathematical analysis, the usage of this methodology is very convenient for the
industries. The authors interviewed and discussed with the evaluators, they all
agree that the integrated model is better than the current approaches and feasible by
using the AHP-GRA decision-aiding software. Through the proposed model, it is
possible to effectively integrate the specialized knowledge and experience of dispersed
each evaluator, and the quantitative data to select a best supplier to cooperation.

6. Conclusion
The ultimate goal of supplier selection is to select appropriate suppliers that can
provide faster delivery, reduced cost and improved quality in order to increase
corporate competitiveness additionally, searching for appropriate suppliers, and
continuously assessing their performance, are critical supply chain activities in the
successful running of an outsourcing manufacturing organizations. However, supplier
selection is a complex multi-criteria decision problem that includes both qualitative
and quantitative factors. In this study, an integrated model of AHP and GRA has been
formulated and applied to a real case to examine its feasibility in selecting a best
supplier. This integrated model is more flexible than existing procedures. For example,
if new criteria emerge (in satisfying changing business requirements), they can be

Figure 3.
Sensitivity of grey
relational grade with
different distinguished
coefficient
JMTM included in the proposed model to select a best supplier. Similarly, any new potential
17,7 supplier can be included in the evaluation process.
The proposed model has several advantages:
.
Using the AHP method to reflect the weights of qualitative criteria and integrate
the various expectations from different evaluators into evaluating the suppliers.
.
This integrated model is well suited to deal with multi-criteria decisions that involve
938 both qualitative and quantitative factors, and application of the GRA method
enables the selection of a best supplier based on limited information and data.
.
The decision-aiding software (AHP-GRA model) has been developed to conduct
the supplier selection process very effectively.

It is, therefore, that the integrated model can be widely applied for the industries.
In comparison with other models, this integrated model is applicable and effective.
Furthermore, establishment of the knowledge management system of supplier
selection is suggested for the outsourcing manufacturing company. Since the historic
data/criteria can be searched by the evaluators from the database to disperse the
lessons learned, thereby accelerating the implementation and success of supplier
selection.
There are several methods for evaluating multiple attributes – including data
envelopment analysis, technique ordered preference by similarity to the ideal solution,
and MAUT. These can all be applied to the supplier selection issue. Further research is
required into the application of these methods, and a comparison of the relative
effectiveness of the results would be of value.

References
Barabarosoglu, G. and Yazgac, T. (1997), “An application of the analytical hierarchy process to
the supplier selection problem”, Production & Inventory Management Journal, 1st quarter,
pp. 14-21.
Bhutta, K.S. and Huq, F. (2002), “Supplier selection problem: a comparison of the total cost of
ownership and analytical hierarchy process approaches”, Supply Chain Management: An
International Journal, Vol. 7 No. 3, pp. 126-35.
Braglia, M. and Petroni, A. (2000), “A quality-assurance oriented methodology for handling
trade-offs in supplier selection”, International Journal of Physical Distribution & Logistics,
Vol. 30 No. 2, pp. 96-111.
Buffa, F.P. and Jackson, W.M. (1983), “A goal programming model for purchase planning”,
Journal of Purchasing and Materials Management, Fall, pp. 27-34.
Chang, C.L., Wei, C.C. and Chen, C.B. (2000), “Concurrent maximization of process tolerances
using grey theory”, Robotics and Computer Integrated Manufacturing, Vol. 16, pp. 103-7.
Chao, C., Scheuing, E.E. and Ruch, W.A. (1993), “Purchasing performance evaluation: an
investigation of different perspectives”, International Journal of Purchasing & Materials
Management, Vol. 29 No. 3, pp. 33-9.
Chaudhry, S.S., Forst, F.G. and Zydiak, J.L. (1993), “Vendor selection with price breaks”,
European Journal of Operational Research, Vol. 70, pp. 52-66.
Chen, C.C. and Yang, C.C. (2003), “Total-costs based evaluation system of supplier quality
performance”, Total Quality Management, Vol. 14 No. 3, pp. 325-39.
Chiou, H.K. and Tzeng, G.H. (2001), “Fuzzy hierarchical evaluation with grey relation model of Supplier
green engineering for industry”, International Journal of Fuzzy System, Vol. 3 No. 3,
pp. 466-75. selection
Choi, T.Y. and Hartley, J.L. (1996), “An exploration of supplier selection practices across the
supply chain”, Journal of Operations Management, Vol. 14, pp. 333-43.
Choy, K.L., Lee, W.B. and Lo, V. (2002), “An intelligent supplier management tool for
benchmarking suppliers in outsource manufacturing”, Expert Systems with Applications, 939
Vol. 22, pp. 213-24.
Choy, K.L., Lee, W.B. and Lo, V. (2003), “Design of a case based intelligent supplier relationship
management system-the integration of supplier rating system and product coding
system”, Expert Systems with Applications, Vol. 25, pp. 87-100.
Dahel, N.E. (2003), “Vendor selection and order quantity allocation in volume discount
environments”, Supply Chain Management: An International Journal, Vol. 8 No. 4,
pp. 334-42.
Davis, T. (1993), “Effective supply chain management”, Sloan Management Review, Vol. 34,
pp. 35-46.
De Boer, L., Labro, E. and Morlacchi, P. (2001), “A review of methods supporting supplier
selection”, European Journal of Purchasing & Supply Management, Vol. 7, pp. 75-89.
De Boer, L., Wegen, L. and Telgen, J. (1998), “Outranking methods in support of supplier
selection”, European Journal of Purchasing & Supply Management, Vol. 4, pp. 109-18.
Deng, J.L. (1982), “Control problems of grey system”, Systems and Control Letters, Vol. 1,
pp. 288-94.
Deng, J.L. (1988), Properties of Relational Space for Grey System, China Ocean, Beijing.
Deng, J.L. (1989), “Introduction to grey system theory”, The Journal of Grey System, Vol. 1 No. 1,
pp. 1-24.
Dickson, G.W. (1966), “An analysis of vendor selection systems and decisions”, Journal of
Purchasing, Vol. 2 No. 1, pp. 28-41.
Feng, C.M. and Wang, R.T. (2000), “Performance evaluation for airlines including the
consideration of financial ratios”, Journal of Air Transport Management, Vol. 6, pp. 133-42.
Ghodsypour, S.H. and O’Brien, C. (1998), “A decision support system for supplier selection using
an integrated analytic hierarchy process and linear programming”, International Journal
of Production Economics, Vol. 56/57, pp. 199-212.
Ghodsypour, S.H. and O’Brien, C. (2001), “The total cost of logistics in supplier selection, under
conditions of multiple sourcing, multiple criteria and capacity constraint”, International
Journal of Production Economics, Vol. 73, pp. 15-27.
Handfield, R., Walton, S., Sroufe, R. and Melnyk, S. (2002), “Appling environmental criteria to
supplier assessment: a study in the application of the AHP”, European Journal of
Operational Research, Vol. 141, pp. 70-87.
Hines, P. (1994), “Can you create your own world class supply chain?”, Purchasing & Supply
Management, September, pp. 30-2.
Hsu, C.I. and Wen, Y.H. (2000), “Application of grey theory and multiobjective programming
towards airline network design”, European Journal of Operational Research, Vol. 127,
pp. 44-68.
Hsu, Y.T., Yeh, J. and Chang, H. (2000), “Grey relational analysis for image compression”,
The Journal of Grey System, Vol. 12 No. 2, pp. 131-8.
JMTM Karpark, B., Kumcu, E. and Kasuganti, R. (1999), “An application of visual interactive goal
programming: a case in vendor selection decisions”, Journal of Multi-Criteria Decision
17,7 Analysis, Vol. 8, pp. 93-105.
Li, P., Tan, T.C. and Lee, J.Y. (1997), “Grey relational analysis of amine inhibition of mild steel
corrosion in acids”, Corrosion, Vol. 53, pp. 186-94.
Liang, R.H. (1999), “Application of grey relation analysis to hydroelectric generation scheduling”,
940 Electrical Power and Energy Systems, Vol. 21, pp. 357-64.
Min, H. (1994), “International supplier selection: a multi-attribute utility approach”, International
Journal of Physical Distribution & Logistics Management, Vol. 24 No. 5.
Monczka, R.M. and Trecha, S.J. (1988), “Cost-based supplier performance evaluation”, Journal of
Purchasing and Materials Management, Vol. 24 No. 2, pp. 2-7.
Mummalaneni, V., Dubas, K.M. and Chao, C. (1996), “Chinese purchasing managers preferences
in supplier selection”, Industrial Marketing Management, Vol. 25 No. 2, pp. 115-24.
Nicholas, J., Holt, G.D. and Love, P.E.D. (2000), “Impacts of credit control and debt collection
procedures upon suppliers’ turnover”, European Journal of Purchasing & Supply
Management, Vol. 6, pp. 237-43.
Nydick, R.L. and Hill, R.P. (1992), “Using the analytical hierarchy process to structure the
supplier selection procedure”, International Journal of Purchasing and Materials
Management, Vol. 28 No. 2, pp. 31-6.
Pan, A.C. (1989), “Allocation of order quantity among suppliers”, Journal of Purchasing and
Materials Management, Fall, pp. 36-9.
Roodhooft, F. and Konings, J. (1996), “Vendor selection and evaluation: an activity based costing
approach”, European Journal of Operational Research, Vol. 96, pp. 97-102.
Rosenthal, E.C., Zydiak, J.L. and Chaudhry, S.S. (1995), “Vendor selection with bundling”,
Decision Sciences, Vol. 26 No. 1, pp. 35-48.
Smytka, D.L. and Clemens, M.W. (1993), “Total cost supplier selection model: a case study”,
Journal of Purchasing and Materials Management, Vol. 29 No. 1, pp. 42-9.
Tam, M.C.Y. and Tummala, V.M.R. (2001), “An application of the AHP in vendor selection of a
telecommunications system”, Omega, Vol. 29, pp. 171-82.
Timmerman, E. (1986), “An approach to vendor performance evaluation”, Journal of Purchasing
and Materials Management, Vol. 26 No. 4, pp. 2-8.
Weber, C.A. and Current, J.R. (1993), “A multiobjective approach to vendor selection”, European
Journal of Operational Research, Vol. 68, pp. 173-84.
Weber, C.A., Current, J.R. and Benton, W.C. (1991), “Vendor selection criteria and methods”,
European Journal of Operational Research, Vol. 50 No. 1, pp. 2-18.
Wu, H.H. (2003), “A comparative study of using grey relational analysis in multiple attribute
decision making problems”, Quality Engineering, Vol. 15 No. 2, pp. 209-17.
Wu, J.H. and Chen, C.B. (1999), “An alternative form for grey relational grades”, The Journal of
Grey System, Vol. 11 No. 1, pp. 7-12.

Further reading
Narasimhan, R. (1983), “An analytical approach to supplier selection”, Journal of Purchasing and
Materials Management, Vol. 19 No. 4, pp. 27-32.
Verma, R. and Pullman, M.E. (1998), “An analysis of the supplier selection process”, Omega,
Vol. 26 No. 6, pp. 739-50.
Appendix. Grey relational analysis Supplier
The algorithm of GRA is stated as follows (Wu and Chen, 1999):
.
Step 1. Generate the referential series of x0 ¼ ðx0 ð1Þ; x0 ð2Þ; . . . ; x0 ð jÞ; . . . ; x0 ðnÞÞ; in which
selection
xi is the compared series of ðxi ð1Þ; xi ð2Þ; . . . ; xi ð jÞ; . . . ; xi ðnÞÞ; where i ¼ 1; 2; . . . ; m: The
compared series xi can be represented in a matrix form.
.
Step 2. Normalize the data set. The series data can be treated using one of the
following three types: “larger-is-better”, “smaller-is-better”, and “nominal-is-best”. For 941
“larger-is-better” data transformation, xi ð jÞ can be transformed into x*i ð jÞ: The formula is
defined as in equation (A1):
xi ð jÞ 2 min xi ð jÞ
j
x*i ð jÞ ¼ ðA1Þ
max xi ð jÞ 2 min xi ð jÞ
j j

where max xi ð jÞ is the maximum value of entity j and min xi ð jÞ is the minimum value of
j j
entity j. For “smaller-is-better”, xi ð jÞ can be transformed into x*i ð jÞ by equation (A2):
max xi ð jÞ 2 xi ð jÞ
j
x*i ð jÞ ¼ ðA2Þ
max xi ð jÞ 2 min xi ð jÞ
j j

Therefore, the normalized referential series of x0 becomes x*0 : The original data set needs
to be normalized with one of the three types of data transformations.
.
Step 3. Calculate the distance of D0i ð jÞ – that is, the absolute value of the difference
between x*0 and x*i at the point of j.
.
Step 4. Calculate the grey relational coefficient g0i ð jÞ using the equation (A3):
D min þ zD max
g0i ð jÞ ¼ ðA3Þ
D0i ð jÞ þ zD max
where D min ¼ min min D0i ð jÞ; D max ¼ max max D0i ð jÞ; and j is the distinguished
i j i j
coefficient (j [ [0,1]).
.
Step 5. Calculate the degree of the grey equation coefficient G0i. If the weights (Wi) of
criteria are determined, the grey relational grade is defined as follows:
X
n
G0i ¼ ½W i ð jÞ £ g0i ð jÞ
ðA4Þ
j¼1

According to GRA, if any of the alternatives has a higher grey relational grade than others, it is
the most important (or optimal) alternative. The present study uses this technique to rank
candidate suppliers in accordance with GRA values.

Corresponding author
Ching-Chow Yang can be contacted at: chinchow@cycu.edu.tw

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

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