نوع مقاله : پژوهشی
دانشکده مهندسی صنایع و سیستمها، دانشگاه تربیت مدرس
عنوان مقاله [English]
In this paper, we deal with a procurement problem in a decentralized two-echelon supply chain, in which a buyer (manufacturer) aims to procure a bundle of needed items from a number of suppliers. The problem is modeled via a bi-level programming model, in which the buyer acts as a leader and the suppliers separately act as followers on lower level. To solve this bi-level mathematical model, a hybrid algorithm based on particle swarm optimization (PSO-A*) is proposed. The proposed mechanism, by satisfying the partners' constraints, is able to reach a near-optimal solution which persuades the partners to contract. In this paper, a comprehensive pattern is proposed for embedding the negotiation process in mathematical models and their solution procedure. This study aims at developing a bi-level programming to deal with a negotiation-based procurement problem, according to the realistic assumptions, in which the buyer is considered as a leader and makes optimal decisions according to suppliers' proposals in lower level as followers. Such a mechanism provides an alignment among suppliers' production planning and order allocation to avoid instantaneous orders, inability of suppliers to supply orders, and impose high inventory cost. In addition, it supports the partnership with valued suppliers through suitable order allocation by taking suppliers' capacities into consideration. This research has been done based on the assumptions derived from the interviews with the experts in supplying automotive parts company called SAPCO and a number of its partners. To evaluate the performance of the proposed algorithm, the results of the PSO-A* algorithm are compared with those of PSO-Exact and PSO-Greedy algorithms. Based on computational analysis, it can be observed that the PSO-A* algorithm is more efficient compared to the PSO algorithm in which its lower level sub problems are solved through an exact solver; it is also more effective compared to the PSO-Greedy algorithm.