پديد آورندگان :
كمالي، حميدرضا نويسنده دانشجوي دكتري دانشكده‌ي مهندسي صنايع، دانشگاه يزد Kamali, H.R , صادقيه ، احمد نويسنده استاد دانشكده مهندسي صنايع، دانشگاه يزد Sadegheih, A , وحدت‌زاد ، محمد علي نويسنده استاديار دانشكده مهندسيي صنايع،دانشگاه يزد Vahdat-Zad, M.A , خادمي‌زارع ، حسن نويسنده دانشيار دانشكده‌ي مهندسي صنايع، دانشگاه يزد Khademi-Zare, H
چكيده فارسي :
در اين نوشتار زنجيره‌ي تامين حلقه‌بسته با لحاظ انواع هزينه‌ها و نيز محدوديت‌هاي ظرفيت و زمان، براي محصول‌هاي تكنولوژيك كه هزينه‌ي توليد و قيمت فروش آنها نزولي است، به‌منظور تعيين مقدار و زمان سفارش‌دهي، توليد و تحويل برنامهريزي مي‌شود. بدين‌منظور از چهار روش متاهيوريستيك الگوريتم ژنتيك، بهينه‌سازي انبوه ذرات، تكامل تفاضلي و كولوني زنبورهاي مصنوعي به‌همراه برخي تغييرات در حل مسيله استفاده شده و جواب آنها براي مسايل با ابعاد كوچك با جواب بهينه‌ي حاصل از حل مدل برنامهريزي مختلط عدد صحيح مسيله مقايسه شده است. نتايج حاصل از تحليل عددي براي مسيله‌ي با ابعاد كوچك و نيز با ابعاد بزرگ نشان مي‌دهد كه خطاي روش تكامل تفاضلي قابل قبول است و نيز كم‌ترين خطا در بين چهار روش ذكر شده است.
چكيده لاتين :
In a global economy, providing products, at the right time in the right quantity and at a low cost, can be regarded as a key to success. Efficient supply chains have an important role in guaranteeing this success. The objective of this paper is to plan a single product, multi-echelon, multi-period closed loop supply chain (CLSC) for high-tech products, and, finally, the decisions made regarding component procurement, production, distribution, recycling and disposal. The considered planning problem is like a Knapsack problem. Therefore, it can be concluded that it is NP-hard. To plan the explored CLSC problem, the time horizon is divided into some equal periods, and planning is done for them. The more the number of divisions or periods and the closer the planning to reality, the more the dimensions of the problem and the more the amount of solving time needed. This is especially true in NP-hard problems. When analytic methods such as the branch and bound method (for solving MILP model) are used, an increase of the problem dimensions leads to a drastic increase in solving time. Thus, in the case of these problems, metaheuristic algorithms should be used to make a near optimal solution. So, four proposed heuristic-based variables, including the genetic algorithm (GA), particle swarm optimization (PSO), differential evolution (DE), and the artificial bee colony (ABC), were implemented in order to solve the mixed integer linear programming model (MILP). Finally, the computational results obtained through these four methods were compared with the solutions obtained by GAMS optimization software. The solution revealed that the DE methodology performs very well in terms of both quality of solution obtained and computational time. The results of this study indicated an approximate solution for selecting active markets among potential markets. Also, for determining the time and quantity of components and products to produce and ship in a CLSC, in general, and for high-tech products, in particular, by dividing the time horizon into many periods, which increases the accuracy of planning.
