218 現実的条件下でのロジスティックス最適設計問題のハイブリッド型メタ解法

抄録

Logistics optimization has been noted increasingly as a key issue to improve the business efficiency under global competition. So far studies have concerned with rather simple problems and major emphasizes have been paid to developing a new algorithm, and to evaluating its validity through benchmarking. However, easy applications following these outcomes often cause a dramatic increase in problem size, and hence in such difficulty that makes impossible to solve the resulting problem by any currently available software. We need to make more earnest efforts to cope with the complex and complicated real-world situations. With this point of view, in this study, we concern with the hierarchical logistic network from two aspects that have been scarcely considered previously though they are very popular in real-world. We consider there a mufti-commodity product problem and a problem with stair-wised discount cost of transportation. To cope with these conditions, we have developed extended methods of the hybrid meta-heuristic approach for each problem. Novel iterative procedures are invented respectively while keep applying the graph algorithm towards the high speed solution of the lower level problem of the decomposed original problem. Validity of the former method is verified through comparison with the commercial software while we implement the algorithm as software amenable for supporting a daily logistic planning, and confirm its effectiveness through a real-world application for the latex.