ABSTRACT

Supply chain integration has become one of the most attractive topics for researchers in recent years. One of the advantages of this integration is in improving overall profit in comparison to separate decisions. In this study, an integrated scheduling and distribution problem is investigated. One of the contributions of this paper is to study this problem from a multi-agent viewpoint. In this case, each agent has a set of jobs with its own objective and compete with each other to acquire supply chain resources. Here, a two-agent problem is discussed where the objectives of the agents are the minimization of the total tardiness and the total cost of distribution. A mathematical formulation and two heuristics based on decomposition approaches are presented. In the first approach, a modified Benders decomposition is presented. Also, some valid inequalities are introduced to increase the convergence speed of this algorithm. In the second approach, a decomposition and cutting approach is developed. The results represent the good performance of both algorithms in comparison to other exact methods.

摘要

供应链整合已成为近年来研究人员最感兴趣的课题之一。与单独的决策相比，这种集成的优势之一是提高了整体利润。在这项研究中，研究了一个集成的调度和分配问题。本文的贡献之一是从多智能体的角度研究这个问题。在这种情况下，每个代理都有一组具有自己目标的工作，并相互竞争以获取供应链资源。在这里，讨论了两个代理问题，其中代理的目标是最小化总迟到和总分销成本。提出了一种基于分解方法的数学公式和两种启发式方法。在第一种方法中，提出了改进的 Benders 分解。还，为了提高算法的收敛速度，引入了一些有效的不等式。在第二种方法中，开发了一种分解和切割方法。与其他精确方法相比，结果代表了两种算法的良好性能。