期待効用値を用いた最適コンテナ船荷役の近似解法

Abstract

It is rather difficult to obtain an optimal sequence of loading containers into a containerized ship by computerized algorithm because it has two different objective functions; optimization of vessel condition like GM, and minimization of the number of containers to shift for picking up target containers on a container terminal. This paper describes a heuristic algorithm for the optimal loading sequence problem. The problem formulated as an integer programming problem with one objective function including the expected number of containers to shift and the contribution rate for GM is solved by the algorithm which consists of two solution methods, i.e., the classical assignment problem by Hungarian method and the integer programming by branch and bound method. In our algorithm a basic approximate solution is obtained by the assignment problem first. If the GM value associated with this solution is satisfied by the ship's condition, then this solution becomes optimal. Otherwise, a revised approximate solution, which offers a GM value more than or equal to the minimum GM, is calculated by the integer programming method.