A reactive GRASP metaheuristic for the container retrieval problem to reduce crane's working time

摘要

The container retrieval problem (CRP) is a very important issue for container terminals. The CRP seeks to find an optimal sequence of operations for the crane to retrieve all the containers from the bay according to a predefined order. An optimal sequence of operations is obtained by either reducing the number of container relocations or reducing any kind of working cost performed by the crane, i.e., energy, time, etc. Although the former is the main objective function discussed in the literature, minimizing the number of relocations does not ensure the solution with the minimal working cost, as evidenced in this paper. Therefore, in this study, a crane's trajectory is defined to better measure the crane's working cost, and the optimization goal is to minimize the crane's working time considering the crane's trajectory. Moreover, it proposes exact methods and a reactive GRASP algorithm for the CRP. The experimental results show that the proposed algorithm is able to provide better solutions for both the number of container relocations and the crane's working time, when compared to heuristic approaches in the recent literature.