یکپارچه سازی TOPSIS فازی و برنامه نویسی آرمانی چند دوره برای خرید محصولات مختلف از تامین کنندگان چندگانه
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|19248||2011||12 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Purchasing and Supply Management, Volume 17, Issue 1, March 2011, Pages 42–53
In this paper, a two-phase approach for supplier selection and order allocation problem under a fuzzy environment is proposed. We investigate a problem in which a single buyer orders multiple products from multiple suppliers in multiple periods. To account for inherent ambiguity and vagueness in most real-world data and information, in the first phase of the approach, a fuzzy multiple criteria decision making (FMCDM) method is used to obtain the overall ratings of alternative suppliers, and to select the most qualified ones for further evaluations. In the second phase, using the goal programming (GP) technique, we construct a multi-objective mixed integer linear programming (MOMILP) model to determine the order quantities of each selected supplier for each product in each period. In the MOMILP model, two goals are considered: 1) total value of purchasing (TVP) and 2) meeting the total available periodic budget. Constructing the approach in this way effectively reduces the risk of purchasing. This is because besides evaluating suppliers with regard to a set of quantitative and qualitative criteria, there is also a systematic way to purchase from more than one supplier. Finally, a numerical example is conducted to clarify the proposed approach and to show its usefulness.
The supplier selection problem deals with defining potential suppliers, selecting the best set of suppliers among them, and determining the shipment quantity of each (Weber et al., 1991). According to Patton (1997), and Michaels et al. (1995), supplier evaluation and selection is a key element in the industrial buying process, and appears to be one of the major activities of the professional industries. Many industrial managers and practitioners select suppliers based on their experience and perceptions. These approaches are obviously subjective and their weaknesses have been investigated in several studies (e.g., Hwang and Yoon, 1981 and Kontio, 1996). In fact, supplier selection has been shown to be a multiple criteria decision making process, and if other real-world considerations are taken into account, may turn out to be an even more complicated problem (Kumar et al., 2006). As a pioneer in the supplier selection problem, Dickson (1966) identified 23 different criteria for selecting suppliers. These criteria include quality, delivery, performance history, warranties, price, technical capability, and financial position. After Dickson many researchers proposed various MCDM methods for supplier selection. For example, Yahya and Kingsman (1999) used the analytic hierarchy process (AHP) (Saaty, 1980) to determine priorities in selecting suppliers. On the other hand, some other researchers have focused on applying mathematical programming (MP) methods for supplier selection. For example, Weber and Current (1993) analyzed the supplier selection problem as a multiple objective decision. In traditional practices, first suppliers are selected, and then the buyer, by taking some other considerations and side constraints into account, makes the final decision on how much to order from each. But in the recent decade, researchers have concentrated on integrated approaches in which the issues of supplier selection and order allocation are simultaneously investigated. For example, Ghodsypour and O’Brien (1998) integrated the AHP and linear programming, and proposed a two-stage approach for this problem. Since then, much research in the field of integrated supplier selection and order allocation has been done. However, a large body of these works considers data as crisp and exact. This treatment of data is far from the real-world situations where uncertainty and vagueness is the prominent characteristic. In this paper, we handle these uncertainty of data and impreciseness of human’s judgments through fuzzy sets theory (Zadeh, 1965), and propose a two-phase approach to deal with the multi-product, multi-period, supplier selection and order allocation problem. More specifically, in the first phase, an MCDM approach under fuzzy environment, based on a modified Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) is used to obtain the overall ratings of alternative suppliers. As a result of the first phase’s evaluations, the most qualified suppliers are selected for further evaluations. In the next phase, using the goal programming (GP) method we construct a multi-objective mixed integer linear programming (MOMILP) model to determine the quantity of each product in each period that should be allocated to each selected supplier. Here, we consider a single buyer who wants to purchase multiple products from multiple suppliers in multiple periods, and whose demand for each product in each period is known in advance. Furthermore, the buyer has set some other considerations, such as the total available budget, and total acceptable defective parts purchased, in making the final decision. The rest of the paper is organized as follows. In the next section we present a literature review on applications of MCDM and multi-objective decision making (MODM) techniques used for solving the supplier selection and order allocation problem. In Section 3, we propose our two-phase approach for the problem under consideration. In Section 4, a numerical example is presented to show the applicability and usefulness of our proposed approach. Managerial implications and limitations of the proposed approach are discussed in Section 5. Finally, conclusions are given in the last section.
نتیجه گیری انگلیسی
One of the most important problems that every organization has to deal with is the supplier evaluation and selection. In the real life, due to some issues such as decreasing the risk of purchasing or providing a competitive environment, it is very common to purchase from multiple suppliers. Therefore, selecting the most satisfactory suppliers and assigning appropriate order quantities to each is a very important business process. To account for natural ambiguity and vagueness in most of the real-world data and information, in this paper we propose an integrated two-phase approach for supplier selection and order allocation problem under fuzzy environments. The problem under consideration consists of a single buyer who purchases multiple products from multiple suppliers in multiple periods. In the first phase, using a fuzzy MCDM approach, we calculate the overall score of alternative suppliers. The importance weights of the criteria are calculated using fuzzy AHP method, and then using a modified fuzzy TOPSIS we calculate the final rating of each supplier. In the second phase, using the goal programming (GP) technique, we construct a multi-objective mixed integer linear programming (MOMILP) model in which two goals of total value of purchasing (TVP) and meeting the total budget of purchasing in each period are taken into consideration. Suppliers’ ratings calculated in the first phase of the approach are used as the parameters of the first goal of the GP model. Additionally, to take into account the criticality of the quality issue, we have considered the defect rate as a constraint rather than as a goal. To illustrate the applicability of our approach, a numerical example is presented at the end of the paper which is followed by sensitivity analysis of the GP model for different levels of periodic budget and TVP. From the obtained results, good consistency of the model is observed. As we stated in the assumptions of the GP model, we did not consider variations of the purchasing price due to the ordered quantities. Therefore, a very practical extension of the mathematical model of the second phase would be considering some form of quantity discount for making purchasing decisions. Utilizing some fuzzy multi-objective methods in the second phase, instead of a crisp GP model, for performing a fully fuzzy decision process can be considered as another possible direction for future research. Furthermore, other MCDM and MODM methods could also be applied, and comparisons with the proposed framework could be carried out.