Applied Mathematical Modelling 31 (2007) 24752486

www.elsevier.com/locate/apm

Analytic network process in supplier selection:

A case study in an electronic rm
Cevriye Gencer *, Didem Gurpinar
Gazi University, Faculty of Engineering and Architecture, Department of Industrial Engineering, 06570 Maltepe/Ankara, Turkey
Received 1 November 2005; received in revised form 1 August 2006; accepted 9 October 2006
Available online 8 December 2006

Abstract
Supplier selection, which is the rst step of the activities in the product realization process starting from the purchasing
of material till to the end of delivering the products, is evaluated as a critical factor for the companies desiring to be successful in nowadays competition conditions. With the scope of this paper, supplier selection was considered as a multi criteria decision problem. A model aiming the usage of analytic network process (ANP) in supplier selection is developed
owning to the evaluation of the relations between supplier selection criterias in a feedback systematic. The proposed model
is implemented in a company of electronic.
 2006 Elsevier Inc. All rights reserved.
Keywords: Supplier selection; Analytic network process; Multi criteria decision making; Purchasing

1. Introduction
Purchasing decisions have a major impact on companies, because of this fact systematic methods must be
followed up. There are two main reasons for this [1]:
First in many companies, the cost of the purchased goods and services accounts for more then 60% of the
cost of goods sold.
Second, over 50% of all quality defects can be traced back to purchase material.
Since 1960s, supplier selection criterias and suppliers performance have been a focal point of many
researchers.
While the traditional vendor evaluation methods primarily considered nancial measures in the decision
making process, more recent emphasis on the incorporation of multiple vendor criteria into evaluation process
[2].

Corresponding author. Tel.: +90 0312 231 74 00/2852; fax: +90 0312 230 84 34.
E-mail address: ctemel@gazi.edu.tr (C. Gencer).

0307-904X/$ - see front matter  2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.apm.2006.10.002

2476

C. Gencer, D. Gurpinar / Applied Mathematical Modelling 31 (2007) 24752486

The studies about supplier selection are based on the years of 1960s. Dickson identied 23 criteria for supplier selection based on a survey of 273 purchasing manager. He showed that quality was perceived to be most
important criteria followed by delivery and performance history [2].
It is never expected from a supplier being perfect, according to all supplier selection criterias. For example,
a supplier product may have a high quality, but cost of the products may not be the lowest. On the other hand,
another suppliers products cost may be the lowest, this is very good for a company, but on the same time
delivery performance may be the worst. As it seen from the example, for making good decisions, supplier
selection process must be handled systematically.
There are many methods used in supplier selection such as cluster analysis [3], case based reasoning systems
[4], statistical models [3], decision support systems [4,3], data envelopment analysis [3,1,5,6], analytic hierarchy
process [79,3] total cost of ownership models [3,10], activity based costing [11], articial intelligence [4,3],
mathematical programming [12,6,13,2,14].
The methods used in supplier selection are intending the eectively of the purchasing decisions and implementing decision-making mechanism systematically.
Gaballa [14] is the rst author who applied mathematical programming to supplier selection in a real case in
1974. He used a mixed integer-programming model to formulate this decision-making problem for the Australian Post Oce.
There are many papers in literature about supplier selection. Some of them are: Ghodsypour and OBrien
[12,14], Zhu [13], Talluri [6], Talluri and Narasimhan [2], Boer et al. [3], Weber et al. [5], Barbarosoglu and
Yazgac [7], Tam and Tummala [8], Dagdeviren and Eren [9], Degraeve et al. [10].
2. Analytic network process
Many decision problems cannot be building as hierarchical because of dependencies (inner/outer), inuences between and within clusters (criterias, alternatives). ANP is very useful to solve this kind of problems.
ANP provides a general framework to deal with decisions without making assumptions about the independence of higher-level elements from lower level elements and about the independence of the elements within a
level. In fact ANP uses a network without the need to specify levels as in a hierarchy [15].
ANP is rstly introduced by Saaty, which is based on 1-9 scale, in his book named The Analytic Hierarchy
Process (1980). After that, Thomas L. Saaty developed this issue in his published book named The Analytic
Network Process (1996).
Saaty suggested the usage of AHP to solve the problem of independence on alternatives or criteria and the
usage of AHP to solve the problem of dependence among alternatives or criteria. The structural dierence
between AHP (hierarchy) and ANP (network) is also shown in Fig. 1. As it seen from the gure a hierarchy
is the simple and special case of a network.

(a)

(b)

Goal
X

Fig. 1. (a) Structure of AHP, (b) Structure of ANP.

C. Gencer, D. Gurpinar / Applied Mathematical Modelling 31 (2007) 24752486

2477

Nodes of the network represent components of the system, arcs denote interactions between them. In order
to build the decision problem, all interactions among the elements should be considered. As it seen from
Fig. 1b; X ! Y means that the elements of a component Y depends on component X.
The control hierarchy, critical for ANP analysis, provides overriding criteria for comparing
each type of interaction that is intended by the network representation. There are two types of control criteria
[16].
A control criterion may be directly connected to the structure as the goal of a hierarchy if the structure is in
fact a hierarchy. In this case the control criterion is called a comparison-linking criterion.
Otherwise a control criterion does not connect directly to the structure but induces comparisons in a network. In this case the control criterion is called a comparison- inducing criterion.
With the scope of this paper, the second control criteria (a comparison-inducing criterion) is chosen
according to the relationship between criterias and alternatives.
In the application of ANP, software like, Ecnet, Super Decision or mathematical programs like Excel,
Maple, Mathematica can be used.
Since its introduction, ANP has been getting more widely used in decision making and numerous applications have been published in literature. For example; ANP model is used by Momoh and Zhu [27], for the
selection of generating power units for appropriate price allocation in a competitive power industry, Sarkis
and Meade [18], for selection of logistics strategy; Buyukyazc [19], for nding the most important factor
for education based on memorization; Erkiletlioglu [20] for determination of the military logistics policy;
Meade and Presley [28], for project selection based on research and development; Saaty and Niemira [21],
for nancial-crisis forecasting; Yurdakul [22], for machine selection; Chung et al. [17], for product mix planning in semiconductor fabricator; Ulutas [23], for selection of the appropriate energy policy for Turkey; Partovi [24], for locating facilities strategically.
Lee et al. [25], used an integrated approach of ANP and goal programming for interdependent information
system Project selection; Karsak, Sozer and Alptekin [26], are also used the same approach for selection of the
best pencil design.

3. The proposed model in supplier selection

Numerous models have been developed for supplier selection. According the below mentioned factors a
model is developed by using ANP in supplier selection.
Several criterias and alternatives can be evaluated with the scope of the decision problem.
Both objective and subjective factors can take into consideration in the decision problem.
There exists an interaction between and within supplier selection criterias and alternatives.
The proposed ANP model in supplier selection is given in Fig. 2.
Step 1: The rst step of the algorithm is the analysis of the supplier selection problem. The main goal of the
supplier selection problem is selecting the best supplier that meets the requirements or criterias of the
company.
Step 2: For the proposed supplier selection model, overall 45 criteria is determined under the below mentioned three main criteria clusters:
Business structure of the supplier.
Manufacturing capability of the supplier.
Quality system of the supplier.
All of the criterias and sub-criterias and detail criterias are given a code letter. These codes, which are given in
Appendix A will be used in the supermatrix.

2478

C. Gencer, D. Gurpinar / Applied Mathematical Modelling 31 (2007) 24752486

Analysis of supplier selection problem

Determining the goal and supplier

selection criterias

Determining the alternative suppliers

ANP

Building the supplier selection model

Identification of the network structure
and relationships

Making the paired comparisions

Solution algorithm
Building the supermatrix

Decision
Finding the limiting priorities

Fig. 2. The proposed ANP model in supplier selection.

Step 3: In this step alternatives are determined. Selecting the alternatives from the successful ones in their eld
of activity by using the preliminary elimination will increase the quality of the decision.
Step 4: In this step the interactions between and within clusters and elements are determined. The supplier
selection network models control hierarchy according to the determined criterias is given in Fig. 3.
Step 5: In this step supermatrix is constructed according to the network built in step 4. The supermatrix structure is shown in Equality 1 and the detailed version of the supermatrix is given in Appendix B.
1

W=

W 11

W 14

W 21

W 22

W 24

W 31

W 33

W 34

W 41

W 42

W 43

In the above equality 1, number 1 shows the criteria cluster belonging to business structure of the supplier; number 2 shows the criteria cluster belonging to manufacturing capability of the supplier; number 3 shows the criteria cluster belonging to quality system of the supplier; number 4 shows the criteria
cluster belonging to alternative suppliers.
In the supermatrix given equality 1, W11, W14, W21, W22, W24, W31, W33, W34, W41, W42, W43, W44
shows the sub-matrixes. Also the clusters, which have no interaction, is shown in the supermatrix

C. Gencer, D. Gurpinar / Applied Mathematical Modelling 31 (2007) 24752486

1. Business Structure of the Supplier

General
information of
the supplier

Organization
Profile of the
supplier

Financial status
of the supplier

3. Quality System of The Supplier

Quality System
Certificate of
the Supplier

Quality System
Documentation
of the Supplier

Quality System
Application of
the Supplier

2479

2. Manufacturing Capability of the Suppler

Equipment
Status of the
Supplier

Manufacturing
Capability of
the Supplier

Material Delivering
Capability of the
Supplier

4. Alternatives

First alternative
supplier

Second
alternative
supplier

Nth alternative
supplier

Fig. 3. The supplier selection network models control hierarchy.

with zero (0). In the supermatrix W21 means that cluster 2 (manufacturing capability of the supplier)
depends on cluster 1 (business structure of the supplier).
Similarly with AHP, the 19 scale developed by Saaty is used and paired comparisons are made to
build up the supermatrix. Also in this step the consistency of each comparison is checked. Each column of a supermatrix is either a normalized eigenvector or all of its block entries are zero.
Step 6: Generally in this step the supermatrix will be an unweighted one. Because in each column it consists of
several eigenvectors which of them sums to one (in a column of a stochastic) and hence the entire column of the matrix may sum to an integer greater than one. The supermatrix needs to be stochastic to
derive meaningful limiting priorities. So for this reason to get the weighted supermatrix, rstly the
inuence of the clusters on each cluster with respect to the control criterion is determined. This yields
an eigenvector of inuence of the clusters on each cluster. Then the unweighted supermatrix is multiplied by the priority weights from the clusters, which yields the weighted supermatrix.
Step 7: Finally the supermatrix will be steady state by multiplying the weighted supermatrix by itself until the
supermatrixs row values converge to the same value for each column of the matrix. According to this
point the supplier, which have the highest priority will be chosen.

4. Application of supplier selection

The proposed model of supplier selection was implemented in an electronic company. There are 615
employee in the company, 286 of them are engineer and 329 of them are technician. The electronic company
is realising the design, manufacturing, purchasing and after sales services of its electronic products. The

2480

C. Gencer, D. Gurpinar / Applied Mathematical Modelling 31 (2007) 24752486

activities like supplier selection, evaluating the suppliers performance and maintaining the list of the approved
suppliers are performed by a team consisting of 16 people under the coordination of purchasing department.
This team includes many disciplines like purchasing, manufacturing, quality and engineering departments.
This team reviews the candidate suppliers according to the evaluation criterias and after this evaluation,
the suppliers which exceed the pre-determined score are added to the approved supplier list of the company.
All the purchase orders are given from the approved suppliers.
The proposed decision model of supplier selection was implemented for 45 criteria under three main criteria
cluster and three main alternative suppliers. In the application, the control hierarchy which is shown in Fig. 3
and the supermatrix structure, which is given in Appendix B was used.
The paired comparisons were made by taking the experts opinion working in the companys purchasing
team. Also all the calculations were performed by using Ms Excel.
After this step, we had our unweighted supermatrix which is given in Table 1. Then clusters priority
weights were calculated by using experts opinion. Multiplying this priority weights by the unweighted supermatrix we had the weighted supermatix which is illustrated in Table 2.
The nal step is the calculation of the limiting priorities of the weighted supermatrix. In the 25th power of
the weighted supermatix, we reached the supermatix which have limiting priorities. According the limiting
supermatix is given in Table 3, the supplier J1 is the highest priority from the other alternatives. Because
of this reason we selected the supplier J1.
5. Conclusions
Supplier selection, which is one of the companys most important processes, must be systematically considered from the decision makers.
For this reason, supplier selection is evaluated by researchers for many years in a large framework consisting various techniques from the experimental to the analytical ones and its successful applications were performed in numerous sectors.
In this paper, supplier selection was considered as an multi criteria decision problem and a model is proposed by using ANP. The evaluation criterias were developed according to the structure of the business scenario and the model was applied to an actual case study. The selected supplier by using ANP is also approved
from the purchasing manager.
This paper shows that, ANP is a decision tool by making strategical decisions, such as selecting a supplier
to have a long term relationships or a procuring a critical material for the company. Generally, managers
might be inclined not to use a sophisticated method, but by using a user friendly software like super decision,
developed by Saaty, the decision making process by using ANP will be handled more easy.
Evaluating the supplier from both objective and subjective criterias will gain exibility to the decision process. When we think a suppliers all functional departments, all these departments have relationships with the
other ones, so we can easily say that all the criterias related with the supplier selection must have inuences
between the criteria clusters.
Another important nding is that the proposed model is more reecting the relation of how the selection
criterias aect the selected suppliers and at the same time what is more important for the suppliers among the
selection criterias.
This paper moves us one step closer to the usage of ANP in real world situations. We think that, this
paper can be extended by adding the risk, cost and opportunity control hierarchies and at the same
time it can be developed by using the limiting priorities as an input in the mathematical program
techniques.

Table 1
Unweighted supermatrix of supplier selection by using ANP

C. Gencer, D. Gurpinar / Applied Mathematical Modelling 31 (2007) 24752486

2481

C. Gencer, D. Gurpinar / Applied Mathematical Modelling 31 (2007) 24752486

Table 2
Unweighted supermatrix of supplier selection by using ANP

2482

Table 3
Limiting supermatrix of supplier selection by using ANP

C. Gencer, D. Gurpinar / Applied Mathematical Modelling 31 (2007) 24752486

2483

2484

C. Gencer, D. Gurpinar / Applied Mathematical Modelling 31 (2007) 24752486

Appendix A. The criterias determined supplier selection by using ANP

Criteria code

Criteria name

1.
A.
A1
A2.
A3.
A4.
A5.
B.
B1.
B2.
B3.
C.
C1.
C2
C3.

Business structure of the supplier

General information of the supplier
Facility location
The number of working years in this sector
The references
The service capability
Communication capability
Organization prole of the supplier
Organization structure
The number of personnel
Education status of the personnel
Financial status of the supplier
The last term prot
Exporting status
Appropriateness of the materials price to the market price

2.
D.
D1.
D2.
E.
E1.
E2.
E3.
E4.
F.
F1.
F2.
F3.
F4.

Manufacturing capability of the supplier

Equipment status of the supplier
Machine capacity and capability
Manufacturing technology
Manufacturing capability of the supplier
Facilities manufacturing capacity
Technical capability
Manufacturing planning capability
Handling and packaging capability
Material delivering capability of the supplier
Appropriateness of the quantity
Appropriateness of the delivery date
Appropriateness of the packaging standards
The period of procuring materials

3.
G.
H.
H1.
H2.
H3.
H4.
I.
I1.
I2.
I3.
I4.
I5.
I6.
I7.
I8.

Quality system of the supplier

Quality system certicate of the supplier
Quality system documentation of the supplier
Quality manual
Documentation control
Archive of quality records
Usage of worth instructions
Quality system application of the supplier
Process control capability
Product identication
Receiving Inspection
Calibration control
Non-conforming material control system
Corrective and preventive action system
Audit mechanism
Training

Supplier code
J1
J2
JN

Supplier name
First alternative supplier
Second alternative supplier
Nth alternative supplier

Appendix B. The super matrix of supplier selecton by using ANP

C. Gencer, D. Gurpinar / Applied Mathematical Modelling 31 (2007) 24752486

2485

2486

C. Gencer, D. Gurpinar / Applied Mathematical Modelling 31 (2007) 24752486

References
[1] E.K. Lee, Development of a Supplier Selection and Management System and Safety Stock Policies in Supply Chain, Thesis for the
Degree of Doctor of Philosophy, Department of Industrial Engineering Graduate School, Korea University, 2001, pp. 4555.
[2] S. Talluri, R. Narasimhan, Vendor evaluation with performance variability: a maxmin approach, Eur. J. Operat. Res. 146 (2003)
543552.
[3] L. Boer, E. Labro, P. Morlacchi, A review of methods supporting supplier selection, Eur. J. Purchasing Supply Manage 7 (2001) 75
89.
[4] K.L. Choy, W.B. Lee, V. Lo, Design of a case based intelligent supplier relationship management system the integration of supplier
rating system and product coding system, Expert Syst. Appl. 25 (2003) 87100.
[5] C.H. Weber, J.R. Current, A. Desai, Non-cooperative negotiation strategies for vendor selection, Eur. J. Operat. Res. 108 (1998) 208
223.
[6] S. Talluri, A buyerseller game model for selection and negotiation of purchasing bids, Eur. J. Operat. Res. 143 (2002) 171180.
[7] G. Barabarosoglu, T. Yazgac, An application of the analytic hierarchy process to the supplier selection problem, Prod. Inventory
Manage. J. First Quarter (1997) 1421.
[8] M.C.Y. Tam, V.M.R. Tummala, An application of the AHP in vendor selection of a telecommunications system. OMEGA, Int. J.
Manage. Sci. 29 (2001) 171182.
[9] M. Dagdeviren, T. Eren, Tedarikci Firma Seciminde Analitik Hiyerarsi Prosesi ve 0-1 Hedef Programlama Yontemlerinin
niversitesi Muhendislik Mimarlk Fakultesi Dergisi 16 (2) (2001) 4152.
Kullanlmas, Gazi U
[10] Z. Degraeve, E. Labro, F. Roodhooft, An evaluation of vendor selection models from a total cost of ownership perspective, Eur. J.
Operat. Res. 125 (2000) 3458.
[11] F. Roodhooft, J. Konings, Vendor selection and evaluation an activity based costing approach, Eur. J. Operat. Res. 96 (1996) 97102.
[12] S.H. Ghodspour, C. OBrien, A decision support system for supplier selection using an integrated analytic hierarchy process and
linear programming, Int. J. Prod. Econ. 5657 (1998) 199212.
[13] J. Zhu, A buyerseller game model for selection and negotiation of purchasing bids: extensions and new models, Eur. J. Operat. Res.
134 (2004) 150156.
[14] S.H. Ghodspour, C. OBrien, The total cost of logistics in supplier selection. Under conditions of multiple sourcing. multiple criteria
and capacity constraint, Int. J. Prod. Econ. 73 (2001) 1527.
[15] Thomas L. Saaty, The Analytic Hierarchy Process, RWS Publications, Pittsburg, 1990, pp. 184192.
[16] L.T. Saaty. Fundamentals of the Analytic Network Process. ISAHP Japan, 1999, pp. 1214.
[17] S.H. Chung, A.H.I. Lee, W.L. Pearn, Analytic network process approach for product mix planning in semiconductor fabricator, Int.
J. Prod. Econ. 96 (2005) 1536.
[18] L. Meade, J. Sarkis, Strategic analysis of logistics and supply chain management systems using the analytical network process,
Transp. Res. E. 34 (3) (1998) 201215.
niversitesi Fen Bilimleri Enstitusu, Ankara, 2000, pp. 4147.
[19] M. Buyukyazc, Analitik S
ebeke Prosesi, Hacettepe U
_
niversitesi Fen Bilimleri
[20] A. Erkiletlioglu, Askeri Lojistik Politikalarnn Belirlenmesinde Analitik S
Teknik U
ebeke Yontemi, Istanbul
_
Enstitusu, Istanbul,
2000, pp. 921.
[21] Saaty L. Thomas, Niemira P. Michael, An analytical network process model for nancial-crisis forecasting, Int. J. Forecasting 20
(2004) 573587.
[22] M. Yurdakul, AHP as strategic decision making tool to justify machine tool selection, J. Mater. Process. Technol. 146 (2004) 365376.
[23] Ulutas H. Berna, Determination of appropriate energy policy for Turkey, Energy 30 (2005) 11461161.
[24] Y.F. Partovi, An analytic model for locating facilities strategically, OMEGA, Int. J. Manage. Sci. 34 (2006) 4155.
[25] Lee, W. Jin, Kim H. Soung, Using analytic network process and goal programming for independent information system project
selection, Comput. Operat. Res. 27 (1998) 367382.
[26] E.E. Karsak, S. Sozer, E.S. Alptekin, Product planning in quality function deployment using a combined analytic network process
and goal programming approach, Comput. Ind. Eng. 44 (2002) 171190.
[27] J.A. Momoh, J.Z. Zhu, Application of AHP/ANP to unit commitment in the deregulated power industry, IEEE International
Conference on Systems, Man, and Cybernetics 1 (1998) 817822.
[28] L.M. Meade, A. Presley, R&D project selection using the analytic network process, IEEE Transactions on Engineering Management
49 (1) (2002) 5966.