الگوریتم فازی ART: روش طبقه بندی برای انتخاب و ارزیابی تامین کنندگان
|کد مقاله||سال انتشار||تعداد صفحات مقاله انگلیسی||ترجمه فارسی|
|21276||2010||6 صفحه PDF||سفارش دهید|
نسخه انگلیسی مقاله همین الان قابل دانلود است.
هزینه ترجمه مقاله بر اساس تعداد کلمات مقاله انگلیسی محاسبه می شود.
این مقاله تقریباً شامل 4680 کلمه می باشد.
هزینه ترجمه مقاله توسط مترجمان با تجربه، طبق جدول زیر محاسبه می شود:
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Expert Systems with Applications, Volume 37, Issue 2, March 2010, Pages 1235–1240
For most of managers purchasing is a strategic issue. Thus, to select the suitable suppliers has strategic importance for every company. The objective of supplier selection is to reduce purchasing risk, maximize overall value to the purchaser and build a long term, reliable relationship between buyers and suppliers. Many methods have been proposed and used for supplier evaluation and selection; most of them try to rank the suppliers from the best to the worst and to choose the appropriate supplier(s). Supplier evaluation and selection is a complex and typical multi criteria decision-making problem. Because of human judgment needs in many area of supplier selection such as preferences on alternatives or on the attributes of suppliers or the class number and borders supplier selection becomes more difficult and risky. In this study, a new tool for supplier selection is proposed. In this paper, we applied Fuzzy Adaptive Resonance Theory (ART)’s classification ability to the supplier evaluation and selection area. The proposed selection method, using Fuzzy ART not only selects the most appropriate supplier(s) and also clusters all of the vendors according to chosen criteria. To explain the Fuzzy ART method a real-life supplier selection problem is solved and suppliers are categorized according to their similarities. The obtained results show that the proposed method is well suited as a decision-making tool for supplier evaluation and selection problem.
Generally in the process of supplier evaluation and selection, firms are ranked by grading with respect to various criteria, classified and best suited one/s is chosen. As a result of this classification, for example with a high classified firm a long term, less controlled, trust based commercial relationship can be established or vice versa. Consequently, managing supplier categorization has become momentous in terms of profitability, productivity and success in achieving time targets. Although many methods have been proposed and used for supplier evaluation and selection, most of them try to rank the suppliers from the best to the worst or to choose the best supplier among others. This study focuses on supplier evaluation and selection from the point of a new perspective based on Fuzzy ART neural networks. In following sections, we introduce the supplier evaluation and selection problem, and formulate the problem as a decision-making model and solve with Fuzzy ART algorithm to categorize the suppliers using many criteria. The rest of the paper is organized as follows: Section 2 mentions supplier selection and its shortcomings. Section 3 describes Adaptive Resonance Theory (ART) neural networks. Section 4 explains Fuzzy Adaptive Resonance Theory which is one of the ART networks. The proposed Fuzzy ART method for supplier selection problem is presented in Section 5. A real case sample problem – solved by this method – and its results take part in Section 6. The final section is for discussion and conclusions.
نتیجه گیری انگلیسی
Supplier selection problems have been solved by several methods. The main ones are linear weighting methods, total cost approaches, mathematical programming techniques and statistical methods. Mathematical programming techniques cause a significant problem in considering qualitative factors. AHP cannot effectively take into account risk and uncertainty in estimating the supplier’s performance. The artificial intelligence (AI) methods can cope better with complexity and uncertainty than ‘traditional methods’, because they are designed to be more like human judgment functioning. In AI systems, users only have to provide the information on performance of a supplier on the criteria. The AI methods subsequently make the actual trade off of the users, based on what they have ‘learned’ from the experts or cases in the past. The mostly used method in practice is linear weighting method and it depends heavily on human judgment. The selection criteria is weighted and rated by human. In most of cases the weights are chosen equally. In supplier selection and evaluation process, as a multiple-attribute decision-making problem, the decision-makers always express their preferences on alternatives or on the attributes of suppliers, which can be used to help rank and categorize the suppliers or select the most desirable one(s). To classify the vendors the DMs should also judge the number of groups and the necessary boundaries for the groups. Every additional DMs judgment necessity makes the supplier selection and evaluation problem more difficult and risky. While the supplier selection method covers an important requirement, it is well known that it has several weaknesses. Therefore in this study, Fuzzy ART neural network algorithm is proposed as an alternative solution of conventional methods of supplier selection and evaluation problem and successful results are acquired. The proposed method makes some significant and remarkable contributions to weak points of the traditional methods. One of the contributions of the proposed method is its clustering ability. The suppliers are categorized by similarity degrees between them. The conventional methods are applied to rank the suppliers and select the most appropriate supplier. To classify the suppliers by TWP one needs to assign a threshold value or values. This threshold is a clean cut distinct for supplier classes and rated by human DM. But, Fuzzy ART method not only selects the most appropriate supplier and also clusters all of the vendors through the matching function. In proposed algorithm, the matching function determines the supplier categories and their memberships. As a result, failure modes are classified according to their similarity without any threshold assigned by DM. Besides this, in the supplier clustering studies, the cluster number and so the cluster boundaries are determined by DM at the beginning of the study (Chen et al., 2006). However in Fuzzy ART method cluster number and cluster boundaries are not defined manually, both are formed mathematically according to matching process. In conventional methods; although, all evaluation criterion (producing critical/safety part, producing similar part, etc.) are rated separately, the categorization of the suppliers are done based only on an aggregated value, for example TWP. The calculated TWP value is used to measure supply performance of the vendor. But two suppliers that have exactly the same TWP value can have totally different grades according to evaluation criteria. It means different suppliers that have the same TWP value can be dissimilar from the supply performance point of view. Oppositely, two vendors with different TWP can be in the same category of supply behavior. The aggregated TWP value can cause loss of fifteen criteria’s separated effects and failures in evaluation process. In proposed Fuzzy ART supplier evaluation algorithm all evaluation criteria ratings are taken into the account separately for each vendor to place the suppliers in the most appropriate group. So suppliers with relatively lower TWP can be classified into relatively higher supplier category or vice versus. This contribution of the algorithm is very important and useful. The proposed method is very effective to deal with the supplier evaluation and selection problem. Fuzzy ART method not only selects the most appropriate supplier but also clusters all of the vendors according to their similarity. The algorithm is adaptive and can easily, rapidly applied to all sectors, all firms in the supply chain. Proposed method is very flexible; can be executed whatever how much the data size. Application of the method does not require any expertise field; by the aid of a small computer program can easily be used in practice.