ハイブリッドタブーサーチによる階層型ロジスティクスの最適化

抄録

Recent globalization of market raises outstandingly the importance of logistic optimization toward just-in-time and agile manufacturing. With this point of view, in this study, we have formulated a site location and route selection problem as a p-Hub problem with capacity constraints. It refers to a non-linear integer programming deciding simultaneously location of hubs and routes from plants to customers via hub facilities. To solve the problem practically as well as efficiently, we have developed a new and novel meta-heuristic method termed hybrid tabu search and implemented it in a hierarchical manner. It relies on a tabu search as the upper level algorithm, and a revised Dijkstra method under Lagrange relaxation for capacity constraints as the lower one. To accelerate the efficiency, we give unique methods to generate an initial hub location based on the minimum spanning tree of the nodes, and to adjust the Lagrange multipliers in imitation of auction mechanism regarding transport cost. Moreover, we adopt a multi-start routine in the aid of ant method to prevent from trapping into the local optimum. Through the numerical experiments that outperformed two popular commercial software, we confirmed effectiveness of the proposed method even for real-life applications.