You are on page 1of 6

Proceedings of the IEEE International Conference on Automation and Logistics August 18 - 21, 2007, Jinan, China

An Application of ANP with Benefits, Opportunities, Costs and Risks in Supplier Selection: A Case Study in a Diesel Engine Manufacturing Firm
Xuguang Tan
School of Mechanical Engineering Tianjin University Tianjin, Tianjin City, China make0098@sohu.com

Ke Ma
School of Management Tianjin University Tianjin, Tianjin City, China make0098@126.com

Wei Guo
School of Mechanical Engineering Tianjin University Tianjin, Tianjin City, China wguo@tju.edu.cn

Tian Huang
School of Mechanical Engineering Tianjin University Tianjin, Tianjin City, China htiantju@public.tpt.tj.cn

Abstract - Supplier selection problem, considered as a multicriteria decision-making (MCDM) problem, is one of the most important issues for firms. Lots of literatures about it have been emitted since 1960s. However, the criteria used by most of them are independent, which usually does not correspond with the real world. From the perspective of reducing this shortcoming, this paper develops a model by using analytic network process (ANP), an extension of analytic hierarchy process (AHP), which incorporates the relationship of interdependence and feedback in a multilevel decision network, the criteria included in this model are categorized in four control hierarchies, that is, benefits, opportunities, costs and risks (BOCR). The proposed model is then implemented in a diesel engine manufacturing firm. Index Terms - ANP, BOCR, MCDM, Supplier selection, Supply chain

I. INTRODUCTION In manufacturing industry, supplier selection problem is considered as one of the most important issues, in that the correctness of the supplier selection is directly or indirectly related to the cost, agility and the customer satisfaction of the whole supply chain [1]. As a survey shows, 70% of the total product cost on average is made up of the purchase of goods and services [2]. Thus, it is very meaningful to do some researches on supplier selection. Several factors may affect the suppliers performance. Researchers have done a lot of work in this field. Dickson identified 23 different criteria for supplier selection including quality, delivery, performance history, warranties, price, technical capability and financial position [3]. Weber et al. reviewed 74 supplier selection articles from1966 to 1991, and showed that more than 63% of them were in a multi-criteria environment [4]. Hence, supplier selection problem is a multiple-criteria decision-making (MCDM) problem which includes both tangible and intangible criteria [5]. For nearly five decades, many literatures have been emitted to study the supplier selection problem, the approaches they used including Linear weighting methods [6, 7], mathematical programming models [8-11] , statistical methods [12, 13] and MCDM methods[14-16]. Among them, analytic hierarchy process (AHP) is considered as one of the most useful approaches as it has the advantage of

incorporating tangible as well as intangible criteria especially where the subjective evaluations of different individuals constitute an important part of the decision process [17]. However, AHP approach has a fatal weakness, that is, it is very hard for the AHP approach to be applied appropriately in solving the problem with the criteria which is interdependent on each other in the real world [18]. From the perspective of reducing this shortcoming, this paper introduces the analytic network process (ANP), an extension of AHP [19], for obtaining a set of suitable weights of the criteria to evaluate the suppliers. The purpose of this study is to solve the multi-criteria supplier selection problem by using ANP with four control hierarchies, that is, benefits, opportunities, costs and risks (BOCR). This paper is divided into four sections. The following section presents a brief review of the ANP approach. Section 3 proposes a model and then it is implemented in a diesel engine manufacturing firm. Short conclusions are given in Section 4. II. ANALYTIC NETWORK PROCESS ANP, first presented by T.L. Saaty in 1996, is a general form of AHP, it allows for more complex interrelationships among the decision levels and attributes [20-22]. ANP provides a general framework to deal with decisions without making assumptions about the independence of the elements between and within levels. It allows both interaction and feedback within clusters (inner dependence) and between clusters (outer dependence). In fact, ANP uses a network to make decisions without the need to specify levels as in a hierarchy [23]. Fig. 1 illustrates the structural relationship in ANP method [24]. The system is divided into two parts. The first part is control hierarchy, consisting of network relationships between goal, criteria and sub-criteria. It affects the internal relationships of the system. The second part is network hierarchy, consisting of network relationships between elements and clusters. In ANP analysis, the control hierarchy provides overriding criteria for comparing each type of interaction.

1-4244-1531-4/07/$25.00 2007 IEEE.

1446

Goal Criteria Control Level

denote C k ( k = 1,2," , n ), let each component k contain mk elements, denoted by ( ek 1 , ek 2 ,", ekmk ), and let Wij ,

ggg
Sub-Criteria

( i = 1,2,", n , j = 1,2, " , n ) denotes the sub-matrix. Equation (1) shows a standard form of a supermatrix [18].
e11 C1 # e11 " e1m 1 C1 e1m1 W11 # # ek 1 Ck # Wk1 # ekmk # Wn1 Cn e
n1

" C k " Cn ek1 " ekmk en1 " enmn " W1k # # " Wkk # # " Wnk " W1n # # " Wkn # # " Wnn

(1)

Group C1

Group C2 Network Level

Group

Group

enmn
Group CN

Fig. 1 Structure relationship in ANP method [24].

Step 4: Selection of the best alternative The supermatrix will be steady state by multiplying the weighted supermatrix by itself until the row values converge to the same value for each column of the matrix. The alternative with the highest overall priority should be selected. III.
CASE IMPLEMENTATION

There are two types of control criteria [25]. 1) One type of control criteria may be directly connected to the structure as the goal of a hierarchy. In this case, the control criteria are called comparison-linking criteria. 2) The other type of control criteria do not connect directly to the structure, but induces comparisons in a network. In this case the control criteria are called comparison-inducing criteria. With the scope of this paper, the second type of control criteria (comparison-inducing criteria) is chosen according to the relationship between criteria and alternatives. ANP comprises four major steps [18, 26]. Step 1: Problem structuring and model construction In the first step, problems should be stated clearly. The goals, criteria and sub-criteria related with the problems should also be selected via brainstorming or other methods. And then, the network with the interactions between and within clusters and elements should be constructed based on all these above. Step 2: Making pair-wise comparisons According to the network proposed in the first step, pairwise comparisons should be made by a number of experts. And then experts preferences should be integrated to form comparison matrices. Consistency test should also be applied to verify the consistency of each comparison matrix. Step 3: Supermatrix formation In Step 3, a supermatrix should be formed, which lists down all the sub-matrices consisting of all the clusters and necessary elements in order on the left and upper sides of the matrix. The supermatrix, whose concept resembles the Markov chain process [18], is a partitioned matrix, it is formed by grouping the local priority vectors obtained in Step 2 and then locating them in appropriate positions based on the flow of effect from one component to another, or from a component to itself as a loop in order to determine the global priorities [27]. Let the components of a decision system

W Company is one of the largest diesel engine manufacturers in China. The products it produced include bus diesel engines, truck diesel engines, generating diesel engines, marine diesel engines, engineering diesel engines and agricultural diesel engines, most of them are very famous in the diesel engine manufacturing industry. During its more than 60 years development process, W Company gradually organized a diesel engine supply chain system, and became its core firm. With the market competition more and more fiercely, W Company needs to compress the total cost of the whole supply chain urgently, thus, the first step it took is to make a comprehensive evaluation to its suppliers in order to find out the problems in its outsourcing phase. Based on this reason, a team consisting of 15 people has been formed to help W Company with its problem under the coordination of purchasing department. Take this case for example, W Company needs to evaluate 4 suppliers named S1 to S4, all of them supply the same product X to W Company. In the following part, a model based on ANP approach will be proposed, and it will then be applied in W Company to assess its suppliers. A. Step 1: Problem structuring and model construction In Fig. 2, the proposed ANP control hierarchy on supplier selection is given, which is the first part of the model. As displayed in Fig. 2, the goal of the supplier selection problem is selecting the best supplier. Four sub-networks, that is, benefits, opportunities, costs and risks (BOCR) have been involved in the network. As the benefits, opportunities, costs and risks usually have unequal importance, four strategic criteria, that is competitive advantages, economic benefits, creativity and improvement, have been incorporated into the model in order to rate BOCR.

1447

Selecting the best supplier

Competitive advantages

Economic benefits

Creativity

Improvement

Fig. 2 ANP control hierarchy.

To evaluate the suppliers performances, 43 different criteria and their relationships are defined under the BOCR merits by the team according to [5, 28-30] combined with the experiences of the persons who come from the purchasing department of W Company. In Fig. 3, the sub-network under
Benefits 3 Production 1 Equipment utilization rate 2 Production capacity 3 Manufacturing cycle time 4 The capability of making production plan 5 Production preparation capability 6 The capability of completing production plan 7 Qualified rate 6 Alternatives S1 S2 S3 S4 5 Sales 1The capability of completing sales orders 2 The veracity of dealing with sales d Opportunities 1 Outsourcing 1The capability of making purchasing plan 2 Purchasing period 3 The capability of completing purchasing orders

the BOCR merits and the relationships between the clusters are illustrated. B. Step 2: Making pair-wise comparisons After the first step, pair-wise comparisons are performed
Costs

2 Transportation 1Transportation capacity 2 Transportation period 3 Transportation frequency

1 Costs 1 Transportation cost 2 Inventory cost 3 Outsourcing cost 4 Sales cost 5 Manufacturing cost 6 Maintenance cost 7 Material cost

2 Alternatives S1 S2 S3 S4

4 Inventory 1Days of average inventory 2 Stock capacity per unit area

Risks

1 Opportunities 1 Informatization level 2 Sales forecast veracity 3 Communication capability 4 The degree of emergent order fulfillment 5 The ability of changing lines

1 Risks 1 The quantity of suppliers per type of material 2 Purchasing plan change 3 Purchasing defect materials 4 Production plan change 5 Return 6 Out of stock 2 Alternatives S1 S2 S3 S4

2 Alternatives S1 S2 S3 S4

Fig. 3 Sub-network under BOCR merits.

1448

by the persons from the purchasing department, the comparison results are combined by geometric mean. Here, the elements in a cluster are compared by applying Saatys 19 scales with respect to an element in another cluster which they are connected to (or on elements in their own cluster). Consistency test also should be applied to verify the consistency of each comparison matrix. As the amount of the workload in the pair-wise comparisons and the consistency tests is very huge, in this case, 86 pair-wise comparison matrices and consistency tests should be built about the judgments. Therefore, the software Super Decisions is used here for analysis. The Super Decisions software is a simple easy-to-use package for constructing decision models with dependence and feedback. C. Step 3: Supermatrix formation With the helps of Super Decisions, the supermatrix could be obtained easily. Take the opportunities sub-network as an example, the unweighted supermatrix, weighted supermatrix and the limit supermatrix are shown in Table I, Table II and Table III, respectively. D. Step 4: Selection of the best alternatives After rating results, the total weights, ideal weights and normal weights of the four suppliers are calculated for the BOCR merits, which are shown in Fig.4.

According to Fig. 4, S3 and S2 are the best suppliers in benefits. S1 is best one in opportunities. The most costly one is S3. S2 is the most risky alternative among all the suppliers. Additionally, a meeting made up of five members from the purchasing department is organized to assess the BOCR based on the four strategic criteria proposed before by using five-step scales (very high, high, medium, low and very low). During the meeting, they were asked some questions such as What level should be given to Benefits according to Competitive advantage?, the consensus results have been given in Table IV. In Table IV, the weights of the strategic criteria are shown in the parentheses. Also, the rates of the BOCR are shown in the last column based on the values of measurement levels: very high: 0.42, high: 0.26, medium: 0.16, low: 0.10, very low: 0.06 [31]. The overall scores of the four suppliers could be acquired by using (2). In (2), Wi denotes the overall score of the ith supplier. Wib , Wio , Wic , Wir denotes the total weight of the ith supplier for BOCR obtained from Fig. 4; Rb , Ro , Rc , R r denotes the rating results of the BOCR obtained from Table IV. Wi = Rb * Wib + Ro * Wio + Rc * (1 Wic ) + Rr * (1 Wir ) . (2) In Table V, the overall weights of the four suppliers have been shown. According to Table V, the supplier S1 is

TABLE I UNWEIGHTED SUPERMATRIX OF THE OPPORTUNITIES SUB -NETWORK 1 Infor~ 1 Infor~ 2 Sales~ 3 Commu~ 4 The degre~ 5 The abili~ S1 S2 S3 S4 0 0 0 0 0 0.519258 0.209293 0.063053 0.208397 2 Sales~ 0.333333 0 0.666667 0 0 0.298205 0.369395 0.209875 0.122526 3 Commu~ 1 0 0 0 0 0.510337 0.093702 0.208558 0.187403 4 The degre~ 0.217638 0 0.091402 0 0.690959 0.050108 0.168105 0.291067 0.49072 5 The abili~ 0 0 0 0 0 0.320801 0.354157 0.262237 0.062806 S1 0.118454 0.312712 0.336185 0.037828 0.194821 0 0 0 0 S2 0.13998 0.192592 0.228757 0.396006 0.042664 0 0 0 0 S3 0.099206 0.303745 0.21807 0.317921 0.061057 0 0 0 0 S4 0.189586 0.231867 0.219345 0.322901 0.036302 0 0 0 0

TABLE II WEIGHTED SUPERMATRIX OF THE OPPORTUNITIES SUB -NETWORK 1 Infor~ 1 Infor~ 2 Sales~ 3 Commu~ 4 The degre~ 5 The abili~ S1 S2 S3 S4 0 0 0 0 0 0.519258 0.209293 0.063053 0.208397 2 Sales~ 0.166667 0 0.333333 0 0 0.149102 0.184697 0.104937 0.061263 3 Commu~ 0.5 0 0 0 0 0.255169 0.046851 0.104279 0.093702 4 The degre~ 0.108819 0 0.045701 0 0.34548 0.025054 0.084052 0.145533 0.24536 5 The abili~ 0 0 0 0 0 0.320801 0.354157 0.262237 0.062806 S1 0.118454 0.312712 0.336185 0.037828 0.194821 0 0 0 0 S2 0.13998 0.192592 0.228757 0.396006 0.042664 0 0 0 0 S3 0.099206 0.303745 0.21807 0.317921 0.061057 0 0 0 0 S4 0.189586 0.231867 0.219345 0.322901 0.036302 0 0 0 0

1449

TABLE III LIMIT SUPERMATRIX OF THE OPPORTUNITIES SUB -NETWORK 1 Infor~ 1 Infor~ 2 Sales~ 3 Commu~ 4 The degre~ 5 The abili~ S1 S2 S3 S4 0.159247 0.110027 0.151165 0.092694 0.075339 0.164159 0.095206 0.070597 0.081567 2 Sales~ 0.159247 0.110027 0.151165 0.092694 0.075339 0.164159 0.095206 0.070597 0.081567 3 Commu~ 0.159247 0.110027 0.151165 0.092694 0.075339 0.164159 0.095206 0.070597 0.081567 4 The degre~ 0.159247 0.110027 0.151165 0.092694 0.075339 0.164159 0.095206 0.070597 0.081567 5 The abili~ 0.159247 0.110027 0.151165 0.092694 0.075339 0.164159 0.095206 0.070597 0.081567 S1 0.159247 0.110027 0.151165 0.092694 0.075339 0.164159 0.095206 0.070597 0.081567 S2 0.159247 0.110027 0.151165 0.092694 0.075339 0.164159 0.095206 0.070597 0.081567 S3 0.159247 0.110027 0.151165 0.092694 0.075339 0.164159 0.095206 0.070597 0.081567 S4 0.159247 0.110027 0.151165 0.092694 0.075339 0.164159 0.095206 0.070597 0.081567

Benefits

Costs

Opportunities Fig. 4 Weights of the suppliers for BOCR sub-networks.

Risks

the best one, while on the contrary, S4 is bad, this is mainly because S1 rates higher in benefits and opportunities, and lower in both Costs and Risks; On contrary, S4 rates lower in benefits and opportunities, and higher in risks and costs. Between S1 and S4, the two suppliers S2 and S3 are in the middle position. The result has been submitted to the purchasing manager of W Company, and the manager satisfied with the result, further measures will be taken to management the four suppliers. IV. CONCLUSIONS As a MCDM problem, supplier selection problem is an extremely important issue for most firms, therefore, it should be considered systematically.

For many years, researchers have developed lots of methods to deal with this problem. They have been used in numerous sectors successfully. In this paper, a model by using ANP approach is proposed, it could reflect the criteria relationships, interdependence and feedback more clearly, and this is very close to the real word. In this model proposed, 43 criteria have been introduced in order to evaluate the performances of the four suppliers which are belongs to a diesel engine manufacturing firm. To depict the problem more systematically, the 43 criteria are categorized in four control hierarchies, that is, benefits, opportunities, costs and risks (BOCR). 86 pair-wise comparison matrices and consistency tests have been used in this model, which could be acquired easily with the help of the software Super Decisions. At the end of this paper, the four suppliers of the diesel engine

Benefits Opportunities Costs Risks

TABLE IV WEIGHTS OF THE STRATEGIC CRITERIA AND THE RATING RESULTS Strategic criteria Competitive advantage Economic benefits Creativity (0.45) (0.28) (0.16) Very high Very high Very high Very high High Medium Medium High Low High Low Medium

Improvement (0.11) High Very high High Low

Rates 0.365 0.300 0.171 0.164

1450

Suppliers S1 S2 S3 S4

TABLE V Overall weights of the suppliers Ideal weights 1.000 0.707 0.704 0.607

Normal weights 0.331 0.234 0.233 0.201

manufacturing firm have been evaluated successfully. The result has been approved by the purchasing manager of the firm. REFERENCES
[1] C. Gopal and G. Cahill, Logistics in Manufacturing, USA, Chicago, 1992. [2] A. Ghobadian, A. Stainer, and T. Kiss, A computerized vendor rating system, Proceedings of the First International Symposium on Logistics, the University of Nottingham, Nottingham, UK, pp. 321-328, July 1993. [3] G. W. Dickson, An analysis of vendor selection systems and decisions, Journal of Purchasing, vol. 2, no.1, pp. 5-17, 1966. [4] C.A. Weber, J. R. Current, and W. C. Benton, Vendor selection criteria and methods, European Journal of Operational Research, vol.50, no. 1, pp. 2-18, 1991. [5] E.A. Demirtas and O. Ustun, Analytic network process and multi-period goal programming integration in purchasing decisions, Computer & Industrial Engineering, in press. [6] E. Timmerman, An approach to vendor performance evaluation, Journal of Purchasing and Materials Management, vol. 22, no. 4, pp. 28, 1986. [7] R.E. Gregory, Source selection: a matrix approach. Journal of Purchasing and Materials Management, vol. 22, no. 2, pp. 24-29, 1986. [8] A. C. Pan, Allocation of order quantity among suppliers, Journal of Purchasing and Materials Management, vol.25, No. 3, pp. 36-39, 1989. [9] C. A. Weber and J. R., Current, Theory and methodology: a multiobjective approach to vendor selection, European Journal of Operational Research, vol.16, pp. 173-184, 1993. [10]R. B. Handfield. GL. Ragatz, KJ. Petersen, and RM. Monczka, Involving suppliers in new product development, California Management Review, vol. 42, no. 1, pp. 59-72, 1999. [11]R. Narasimhan, An analytic approach to supplier selection, Journal of Purchasing and Materials Management, pp. 27-32, 1983. [12] S.H. Ghodsypour and C. OBrien, 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, 2001. [13]Z. Gao and L. Tang, A multi-objective model for purchasing of bulk raw materials of a large-scale integrated steel plant, International Journal of Production Economics, vol. 83, pp. 325-334, 2003. [14]R. Narasimhan, An analytic approach to supplier selection, Journal of Purchasing and Materials Management, pp. 27-32, 1983.

[15]S. H. Ghodsypour and C. OBrien, 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, 1998. [16]C.H. Weber, J.R. Current, and A. Desai, Non-cooperative negotiation strategies for vender selection, European Journal of Operational Research, vol. 108, pp. 208-223, 1998. [17]S. Iwasaki and K. Tone, A search model with subjective judgments: Auditing of incorrect tax declarations, Omega-International Journal of Management Science, vol. 26, no. 2, pp. 249-261, 1998. [18]T.L. Saaty, Decision Making with Dependence and Feedback: the Analytic Network Process, RWS, Pittsburgh, PA, 1996. [19]T.L. Saaty, The Analytic Hierarchy Process, McGraw-Hill, New York, 1980. [20]N. Kameya, H. Miyagi, N. Taira, and K. Yamashita, Multiple-criteria decision-making using sectional supermatrix, The Proceedings of the International Technical Conference on Circuits/System, Computers and Communications, Phuket, Thailand, pp. 838-840, 2002. [21]L.M. Meade, Strategic analysis of logistics and supply chain management systems using the analytical network process, Transportation Research Part E-Logistics and Transportation Review, vol. 34, no. 3, pp. 201-215, 1998. [22]M. A Raisinghani, Balanced Analytic Approach to Strategic Electronic Commerce Decisions: A Framework of the Evaluation Method, Information Technology Evaluation Methods and Management, Hershey PA: Idea Group Publishing, 2001, pp. 185-197. [23]T.L. Saaty, The Analytic Hierarchy Process, RWS Publications, Pittsburg, 1990, pp.184-192. [24]L.F. Wang, The theory and algorithm of ANP, System Engineering Theories and Practices, no. 3, pp.44-50, 2001. [25]L.T. Saaty. Fundamentals of the analytic network process ISAHP Japan, pp.12-14, 1999. [26]L.M. Meade and J. Sarkis, Analyzing organizational project alternatives for agile manufacturing process -- an analytical network approach, International Journal of Production Research, vol.37, no.2, pp. 241-261, 1999. [27]C.W. Chang, C.R. Wu, C.T. Lin, and H.L. Lin, Evaluating digital video recorder systems using analytic hierarchy and analytic network process, Information Sciences, in press. [28]C. Gencer and D. Gurpinar, Analytic network process in supplier selection: A case study in an electronic firm, Applied Mathematical Modelling, in press. [29]M.C. Y. Tam and V.M. R. Tummala, An application of the AHP in vendor selection of a telecommunications system, Omega, vol. 29, pp.171-182, 2001. [30]S. Erdogmus, M. Kapanoglu, and E.Koc, Evaluating high-tech alternatives by using analytic network process with BOCR and multiactors, Evaluation and Program Planning, vol. 28, pp.391-399, 2005. [31]T.L. Saaty, The Analytic Network Process. Pittsburgh: RWS Publications. 2001.

1451

You might also like