You are on page 1of 7

IJAMS

THE PROCESS OF DECISION MAKING REGARDING PARTNERSHIPS INCLUDING A MULTICRITERIA METHOD


Patrcia Guarnieri1,University of Brasilia, Dept. of Administration, Brasilia, Distrito Federal, Brazil, patguarnieri@gmail.com Kazuo Hatakeyama2, University of the Vale do Rio dos Sinos, Postgraduate Program in Production Engineering and Systems, So Leopoldo, Rio Grande do Sul, Brazil, khatakeyama@unisinos.br
Abstract: Nowadays, the relationships between buyers and suppliers have not more adversarial nature. The concept of partnerships or strategic alliances in supply chain management have becoming a prevailing factor in the process of supplier selection and evaluation, because collaborate with suppliers is an efficient strategy to get competitive advantage. However the task of select and evaluate suppliers is a complex matter because involves several conflicting criteria, which can have qualitative and quantitative nature, then this task has a multicriteria focus. In order to choose the most appropriate partner, the companies should define an adequate set of criteria and a set of potential suppliers, besides that a multicriteria method needs to be chosen to establish the preferences structure of the decision makers involved in this process. These decision makers are is most time the people in charge of Purchasing activity, nevertheless in many times, professionals from Logistics, Quality and the director of SCM can be also involved. The purpose of this paper is to demonstrate the importance of classify supplier to different levels of partnerships according his performance in a set of metrics/criteria, through a multicriteria method. A numerical application in presented to illustrate the method applied, which is the ELECTRE III. Moreover applying multicriteria methods the decision makers can have a structured way to choose an appropriate partner and share the benefits from partnerships with the entire supply chain. Keywords: Supplier selection, partnerships in SCM, Multicriteria methods, ELECTRE III

1. Introduction
The need to maintain relationships with suppliers has become a critical issue to business [1] which is a result of competitive pressures, the need to achieve cost efficiency and urgency of establishing relationships with key suppliers. These close relationships can provide buyers with the expertise needed to develop new products, new technologies and new processes. Besides that [2] the suppliers, nowadays, can be considered as an extension of the manufacturing process of the companies in the current competitive environment, processes are dependent on reliable supply flows that respond rapidly to demand variability. In this context the activities of identify the best suppliers for a new product or service or to evaluate the performance of a former supplier, is very complex due to the metrics used include not only quantitative measures (as cost, delivery fees, defect rate, etc.) but also qualitative factors, such as stability of management, reliability, ability to design and process, management capability, financial condition, among other factors [3]. Considering that the process of decision making regarding partnerships needs to consider multiple, and,

in several times, conflicting criteria (metrics), it is configured as a MCDM Multiple Criteria Decision Making problem by nature. Thus, considering that the measurement of supplier performance possesses multidimensional and multi-scale, the aggregation of preferences and the evaluation process have its complexity increased [4]. In the last decades, many articles have approached these issues in several situations as: Selection and Evaluation of the suppliers performance for the purchase of components/materials [5], [6], [7]; select the most appropriate alternative for outsourcing [7], [8]; selection of partners for co-development alliances [9]; supplier selection and optimal allocation order, satisfying constraints [10], [11], [12]; select partners in supply chain at different stages of product life cycle [4], [13]; evaluation of suppliers environmental performance [14]. Therefore, after literature review it is possible to state that few articles are about partnerships or strategic alliances in Supply Chain Management, using MCDM methods. The purpose of this article is to demonstrate how the utilization of MCDM methods, more specifically, ELECTRE III method can aid suppliers in the process of decision making regarding partnerships.

International Conference on Agile Manufacturing Systems 2011, Agra, India.

Guarnieri, Patricia and Hatakeyama, Kazuo

The article is structured as follows: Firstly are presented details about the method used and the steps of application. Secondly, through a numerical application are detailed the results obtained and also, the results are discussed based on numerical application. Thirdly, the conclusions shows that using MCDM methods in this context, the managers can reduce the risk in the suppliers selection processes and analyze the situation in a formal and structured way.

2. Materials and methods


It is important to emphasize that this article is neither empirical nor deductive hypothetical, actually it proposes a model of multicriteria decision support and described the modeling procedure, similar to that used in the field of Operations Research. It was proposed by [15] five phases to structure problems in a formal way, this article highlight the stages of structuring the problem and model building. In problems structuring are specified goals, alternative actions and restrictions, so this phase is essential in decision making, because if wrong, can compromise all the process. It was stated by [16] that, a model can be considered as a representation of essential aspects of a process or a system, which can provide useful information about it. The process of model construction follows patterns that depend on the complexity of the system and the availability of information, according to [15]. In this work the modeling was performed according to the structure of MCDA approach.

with a single metric, such as a currency. In this case, the MCDA methods allow the combination of goals through the subjective evaluation of a single or group of decision makers. It was stated by [17] that experts in multicriteria decision aid have a habit of splitting MCDA methods in three families: i) multi-attribute utility theory, ii) outranking methods and; iii) interactive methods. [20] calls them, respectively: i) Single-criterion synthesis approach, which eliminates any incomparability; ii) Outranking Approach, which accepts incomparability and; iii) Approach of Interactive Local Trial, which uses trial-error interactions. Besides that, the main differences among these three approaches can be emphasized: In the first approach, derived from the American thought stream, local preferences (at each attributes level) are aggregated into a unique (utility, value), which is carried out in an additive (compensatory) way. Some multicriteria methods belonging to this approach can be highlighted: MAUT, SMART, TOPSIS, AHP and Goal programming. The second approach, derived from the French thought stream, is first aimed at building binary relations (called outranking relations) in order to represent the decision makers preferences (based on the available information). The aggregation of preferences in this approach does not accept the compensation (tradeoffs) between criteria [21]. Key methods of this approach are: ELECTRE, PROMETHE [22]. Methods that belong to the third group are mainly developed within the MOLP framework. In these methods, computation steps (providing successive trade-offs) and dialogue steps (additional information on the decision-makers preferences). Also, it is important to point out, specifically to the situation presented in this article, some peculiarities: i) It is desirable that suppliers present a balanced performance in all criteria (a poor performance should not be compensated by a very good performance on another criterion, it means that tradeoffs are not desirable), ii) It is expected that all types of preference relations are accepted, in order to make the problem closer to reality. Thus, the ELECTRE (Elimination et Choix Traduisant la ralit) III method is most appropriate for this situation, considering that it is from outranking approach, which does not allow compensation between criteria and accepts all preference relations, including incomparability. The Incomparability should not be confused with indifference, it occurs between two alternatives a and

2.1 Modeling through Multicriteria Decision Aid (MCDA) approach


The concept of Multicriteria Decision Aid (MCDA) aims to provide decision makers with some tools to allow them to progress in solving decision problems, where several - often contradictory - points of view must be taken into consideration. Moreover, it is not possible to say that, in general, any decision (solution, action) is better, both originating from all points of view, therefore, the concept of optimization is not appropriate in this context [17]. The purpose is to indicate more appropriate solutions to the preference structure of a given decision maker [18]. There are basically three key stages of the process of MCDA [19]: 1. Identification and structuring of the problem; 2. Construction and use of the model; 3. Development of action plans. In addition, MCDA methods are required when there is not a way to represent all of the objectives of a problem

The process of decision making regarding partnerships including a multicriteria method

b if there is not a clear evidence in favor of some kind of preference or indifference [23]. The ELECTRE III method was chosen to rank suppliers, according to a set of criteria, an important aspect to note is that this method uses a ratio of overvalued, which means that values are assigned to the alternatives in the decision problem [18]. According to [24], in the version III of ELECTRE method [20] incorporated Fuzzy methods in the construction of ranking alternatives calculation, allowing the creation of pseudo-criteria. The Fuzzy version of this approach allows the incorporation of the inaccuracies and uncertainties in the decision making process, setting the preference and indifference thresholds. The ranking of alternatives from the best to worst is achieved by introducing weights of the criteria in a scale from 0 to 1 or to 100, for example According to [17] they are like votes given to each criterion. At the first stage of the method are used Concordance and Discordance Indexes for the construction of the outranking relation, which is based on the Outranking Degree S (a, b). This index indicates how much the credibility of 'a' outranking on 'b', with values between 0 and 1, and consists of a function increasing gi(a) and decreasing gi(b), for any criterion 'i' [18]. This index is given by: Equation (1)

At the second stage, occurs the exploitation of outranking relation, through which are obtained complete two pre-orders. In order to obtain the first preorder, we apply the procedure of descending distillation, which consists of: 1) From the set A, we select the best alternative, called the first distillation D1; 2) The alternative in D1 is removed from the set, which is now called A'; 3) The procedures 1 and 2 are repeated on the set A', resulting in a second distillation D2 and so on, until reaching the alternative with the worst performance [18]. In addition, [18] states that the second pre-order is obtained through an ascending distillation procedure, in which are selected the alternatives with lower performance.

3. Results from Numerical Application


In this section it is presented a numerical application for the purpose of illustrating the use of the multicriteria method, simulating a situation with real characteristics from Brazilian automotive industry. The steps of the numerical application are: Description of the problem context and identification of the managers objectives; Definition of a set of criteria; assigning weights to the criteria, carrying out the evaluation of alternatives vs criteria and; applying the method ELECTRE III.

3.1 Description of the problem context and identification of the managers objectives
Consider an automaker needs to evaluate and select existing suppliers to a partnership involving the supply of automotive systems/modules (automotive systems/modules are sets of parts with a full function, eg: cockpit - module direction includes steering wheel, arrow panel, horn, etc.) so, instead of buying individual pieces, the company purchased modules, which are directly assembled on the vehicle, reducing the time and manufacturing costs. However, for this purpose it is necessary that suppliers have some capacity for innovation, besides that production of modules demands the co-design of products, commitment to enable the partnership, technical and technological capacity to produce the modules; and capacity to meet the demand in small quantities on time. In order to know if suppliers meet these requirements the automaker needs to undertake a process of formal evaluation of suppliers performance based in a set performance indicators, also known as criteria, which are able to measure whether the objectives are achieved when the manufacturer chooses a particular supplier. So, through a ranking of suppliers it is possible to know which one has better performance and if he is able to assume such responsibility.

Equation (2)

The Concordance Index indicates the extent of the harmony with the statement a is, at least, as good as b. However it is possible to have situations where there is not this concordance, so there is another index which includes the possibility to establish a veto for this statement, which is called Discordance Index [18], [24]. So, this index uses a veto threshold, which allows that the credibility of statement aSb is refused, if: Equation (3) Thus, the Discordance Index is defined by:

Equation (4) Then, the Outranking Level is equal to Concordance Index, when there is not Discordance for anyone criterion. Otherwise, the Concordance Index is reduced due to Discordance levels.

Guarnieri, Patricia and Hatakeyama, Kazuo

3.2 Definition of a set of criteria


For this project consider that the Purchasing Manager and the Director of Supplier Relationship Management met to discuss the situation, they indicate the objectives and a set of appropriated indicators for this evaluation process. The set of indicators chosen for this purpose, were: 1) Innovative capacity; 2) Quality; 3) Commitment; 4) Technical and technological capacity; 5) JIT ability. Analyzing its supplier roll, the managers realized that there are few key suppliers that have such potential. Just four suppliers can meet the performance indicators chosen. Thus, the managers start to find methods most suitable for the evaluation and selection of suppliers and concluded that MCDA methods would be appropriate for this situation. In addition, the managers decided that is not desirable that were trade-offs between criteria, because a low performance in one indicator could not be compensated by a high performance on another indicator. In this specific situation it is required that suppliers have a satisfactory/balanced performance in all indicators. Then it is possible to move for the next step, which includes: to assign weights to the criteria, evaluate alternatives vs criteria and, apply the method ELECTRE III.

the lowest level. The payoffs matrix generated can be visualized in Table 1.
Table 1 Suppliers Payoff Matrix

Source: This research (2011).

Then, with data presented in Table 1 it was possible to start the use of the software ELECTRE III-IV, demo version, developed and available at website of Lamsade - Universit Paris Dauphine, to proceed the ranking of suppliers according its performance in each criterion. First of all, in order to carry out the processes of descending and ascending distillation, it is necessary establish two thresholds: =0.15 (indifference) and =0.30 (preference). Besides that, in this article the values of veto threshold for all criteria is equal to 0, considering the structure of proposed criteria and the judgments scale used, besides the suppliers already have gone through a prior selection process, remaining in the base of company, only those considered keysuppliers. After the calculation process accomplished by the software, we obtain the median and final pre-orders, which allow to rank the suppliers. These results are presented in Figure 1 and Table 2, respectively.

3.3 Assigning weights to the criteria, evaluating alternatives vs criteria and; applying the method ELECTRE III
First of all, to start the process of application of the method, the managers need to assign weights to the criteria, due to some of them be considered more relevant than the other, given the impact in the achievement of objectives. Also in this step, it is necessary to evaluate the alternatives (four suppliers) based on a set of five criteria, generating the payoffs matrix (Table 1) presented below. To simplify the presentation of data, the suppliers were codified as: S1, S2, S3, S4 and the criteria as: C1, C2, C3, C4, C5.

Source: This research (2011)

Based in the historical performance of suppliers the managers gave some weights to the suppliers according each indicator. In this case, it is considered that C1=0,25; C2=0,20; C3=0,25; C4=0,15 and; C5=0,15. After this, in order to proceed the evaluation of the alternatives vs criteria, it was constructed Likert Scales, with verbal and ordinal characteristics, with five levels for each criterion, considering 5 the highest level and 1

Figure 1 Distillation Processes

The process of decision making regarding partnerships including a multicriteria method

Table 2. Rankings

this specific situation. In Table 4 is presented the preference relations matrix.


Table 4 Preference relations matrix

According the results presented in Figure 1, both distillations process are convergent, this situation indicates that the system designed to evaluate and rank the suppliers was able to compare the alternatives. Another possibility would be the divergence between these two classifications, which indicates that there is a failure of the system in compare these alternatives, in this case the model parameters may be revised, but it is not this case. Also, it is possible to perceive according Table 2, that both orders: median and final generated the same ranking of the alternatives, which prove that the models parameters do not need to be revised. The Concordance and Credibility indexes were also generated and can be visualized in the Table 3.
Table 3 Concordance and Credibility indexes matrix

Source: This research (2011)

As shown in Table 4, regarding the preference relations generated after the pairwise comparison of the suppliers, it is possible to observe that the Supplier 4 is preferable (P) over all other suppliers, for this reason it was ranked as the best alternative; Supplier 2 is preferable (P) to Suppliers 1 and 3 and is not preferable (P-) to Supplier 4; Supplier 1 is not preferable (P-) to Suppliers 2 and 4, but it is preferable (P) to Supplier 3 and; Supplier 3 is not preferable (P-) to all suppliers, for this reason is was ranked as the worst alternative. Therefore, with the application of a MCDA method, more specifically the ELECTRE III, the managers had the possibility to structure the problem of evaluation and selection of suppliers to partnerships, regarding codesign and co-production, in a formal way, which significantly reduces the subjectivity of the decision making process and the risk of a wrong choice. Also it was possible to consider several performance indicators in the model, not only price, which is commonly used as a major factor in negotiations. Furthermore, it is important to emphasize that in relationships with suppliers involving partnerships, the price cannot be considered as the more important and only indicator, considering that there are several other capacities and skills, which should be taken into account.

Source: This research (2011).

According the application of Equation (1) by ELECTRE III method the alternatives (suppliers) were compared pairwise, analyzing the statement aSb. Thus it was possible to obtain the Concordance Matrix shown in Table 3. Through this matrix we realize that: 1)Supplier 1 is better 43% than Supplier 2; 86% than supplier 3 and 29% than Supplier 4; 2)Supplier 2 is 71% better than Supplier 1; 100% better than Supplier 3 and 64% than Supplier 4; 3) Supplier 3 is 29% better than Supplier 1; 43% better than Supplier 2 and 43% better than Supplier 4; 4) Supplier 4 is 71% better than Supplier 1; 86% better than Supplier 2 and 100% better than Supplier 3. The alternatives with better performance are respectively: S4 S2 S1 S3, as shown in the median and final pre-order (eg. Table 2). The Credibility Matrix presented the same indexes because it was decided do not use the veto threshold in

4 Conclusions
In this paper it was proposed the utilization of ELECTRE III method to evaluate and rank suppliers according its performance in a set of indicators. The main features of the problem were presented and the steps of modeling and implementation of the proposed model were described. Initially, it was carried out a literature review in order to identify the main approaches to the problem involving multiple criteria. In strategic relationships, incorporate multiple criteria is essential, given that in many cases the suppliers perform part of the processes of the buyer company, thereby selecting partners and constantly evaluate the

Guarnieri, Patricia and Hatakeyama, Kazuo

existing ones, became pressing in supply chain management due to the inclusion of partnerships in business processes. Therefore, according the context presented in numerical application, the method ELECTRE III was more appropriate, due it is a method from outranking approach, which does not allow tradeoffs between criteria, also it provides more balanced solutions. In this particular case, it desirable suppliers present a satisfactory performance in the set of criteria proposed to implement close relationships with suppliers. This systematic proposal constitutes a useful tool for decision support in a multicriteria perspective. Moreover, using structured methods to select suppliers to partnerships, the company can reduce the risk of choose an inappropriate supplier and eliminates problems as: quality failures, product losses, delayed delivery and, consequently, reduces total costs. In addition, evaluating the performance of partners can provide feedback to them and propose improvements in deficient processes, aiming a greater adherence to companys policies.

6. Boran, F. E.; Gen, S.; Kurt, M.; Akay, D. A multi-criteria intuitionistic fuzzy group decision making for supplier selection with TOPSIS method. Expert Systems with Applications Volume 36, Issue 8, (2009), pp. 11363 - 11368. 7. Ordoobadi, Sh. M. Development of a supplier selection model using fuzzy logic. Supply Chain Management: An International Journal 14/4 (2009), pp. 314 327. 8. Almeida, A. T. de. Multicriteria decision model for outsourcing contracts selection based on utility function and ELECTRE method. Computers & Operations Research 34 (2007), pp. 3569 3574. 9. Feng, B.; Fan, Z. P.; Ma, J. A method for partner selection of codevelopment alliances using individual and collaborative utilities. Int. J. Production Economics 124 (2010), pp. 159 170. 10.Razmi, J.; Songhori, M. J.; Khakbaz, M. H. An integrated fuzzy group decision making/fuzzy linear programming (FGDMLP) framework for supplier evaluation and order allocation. Int. J. Adv. Manuf. Technol. (2009) 43:590 607. 11.Lin, R.-H. An integrated FANPMOLP for supplier evaluation and order allocation. Applied Mathematical Modelling Volume 33, Issue 6, June 2009, pp. 2730 2736. 12.Ustun, O.; Demirtas, E. A. An integrated multiobjective decision-making process for multi-period lot-sizing with supplier selection. The International Journal of Management Science. Vol 36 (2008), pp. 509 521. 13.Chang, Sh. L.; Wang, R. Ch.; Wang, Sh.Y. Applying fuzzy linguistic quantifier to select supply chain partners at different phases of product life cycle. Int. J. Production Economics 100 (2006), pp. 348 359 14.Awasthi, A., et al. A fuzzy multicriteria approach for evaluating environmental performance of suppliers. International Journal of Production Economics (2010), pp. 370 - 378. 15.Ackoff, R. L.; Sasieni, M. W. Operational Research. Editora Livros Tcnicos e Cientficos: Rio de Janeiro, 1974.(in Portuguese). 16.Souza, F. M. C. de. Rational decisions in uncertainty situations Recife: Livro Rpido, 2005.(in Portuguese). 17.Vincke, P. Multicriteria decision-aid. John Wiley & Sons, 1992. 18.Almeida, A. T. de. The Knowledge and the use of multicriterion of decision support Editora Universitria UFPE: Recife, 2011. (in Portuguese). 19.Belton, V.; Stewart, T. J. (2002). Multiple Criteria Decision Analysis. Kluver Academic Publishers.

5. Acknowledgements
The authors acknowledge CAPES to support part of this research work.

6. References
1 Lambert, D. M. Supply chain management: processes, partnerships, performance. Sarasota: SCM Institute, 2004. 2. SUPPLY CHAIN COUNCIL. Supply-chain operations reference - SCOR model version 8.0. Disponvel em: http://wenku.baidu.com/view/40025537f111f18583d 05abe.html. 3. Bozarth, C. C.; Handfield, R. B. Introduction to operations and supply chain management. New Jersey: Pearson Education Inc. (2008). 4. Wang, Sh-Y.; Chang, Sh.-L.; Wang, R.-Ch. Assessment of supplier performance based on product-development strategy by applying multigranularity linguistic term sets. Omega The International Journal of Management Science 37 (2009), pp. 215 226. 5. Ghodsypour, A., S.H.; OBrien, C. Fuzzy multiobjective linear model for supplier selection in a supply chain. Int. J. Production Economics 104 (2006), pp. 394407.

The process of decision making regarding partnerships including a multicriteria method

20.Roy, B. ELECTRE III: Un alghoritme de methode de classements fonde sur une representation floue des prfrences en presence de critres multiples. Cahieres de CERO, v. 20, n. 1, (1978), pp. 3 - 24. 21.Martel, J-M. Multicriterion Decision Aid: Methods and Applications. In: CORS - SCRO 1999 Annual Conference, June 7-9, 1999 Windsor, Ontario. 22.Bana e Costa, C.A. Structuration, Construction et Exploitation d'un Modele Multicritre d'Aide la Dcision. Lisboa Portugal: Instituto Superior Tcnico, Universidade Tcnica de Lisboa, 1992. (Thesis). 23.Roy, B.; Bouyssou, D. Aide multicritre la dcision: Mthodes et cas. Paris, Economica, mai. 1993. 24.Infante, C. E. D. C.; Mendona, F. M.; Valee, R. A. B. Application of the Electre III Method in the case of the region of the Field of Vertentes in Minas Gerais" In: SPOLM Symposium of the Operational Research & Logistics of the Marine. Annals .. School of the Navy War, Urca, Rio de Janeiro, 15 to 16 September, 2011.(in Portuguese).

Production Engineering and Management (University of Strathclyde/Scotland). Presently, is a Professor at UNISINOS, Brazil. In his career, he acted in several fields of Engineering, as design of engineering courses, manufacturing processes, prototype apparatus for research and production processes, consultant in quality and cost management, SCM and logistics. He has published several papers on international and national journals, congress, and conferences. E-mail: khatakeyama@unisinos.br

Biography:
Patricia Guarnieri PhD in Production Engineering by Federal University of Pernambuco, Brazil, MSc in Production Engineering from UTFPR, Brazil. Postgraduate in Enterprise Management and Teaching of Higher Education by, respectively, Univel, Brazil and Unipan, Brazil. BS in Accountancy from the UNIOESTE, Brazil. Presently she is a researcher and lecturer in SCM and Logistics and Decision Analysis areas at University of Brasilia, Brazil. In her career, she acted in several fields of Production Engineering and Enterprise Management, such as Supply and Reverse Logistics, SCM, Financial Analysis and Environmental Accountancy. She has published several papers in international and national journals, conferences, and congress. E-mail: patguarnieri@gmail.com

Kazuo Hatakeyama
PhD in Mechanical Engineering (Univ. Wales/Wales), Post doctorate in Higher Education Management (The Victoria Univ. of Manchester/England), Information, Production and Systems (Univ. of Waseda/Japan), Technology Innovation Policies (Univ. of Brighton/England), MSc Planning of Technical Education (OSU/USA), PG Diploma

You might also like