Decentralized Cooperative Metaheuristic for the Dynamic Berth Allocation Problem

Abstract

The increasing demand of maritime transport and the great competition among port terminals force their managers to reduce costs by exploiting its resources accurately. In this environment, the Berth Allocation Problem, which aims to allocate and schedule incoming vessels along the quay, plays a relevant role in improving the overall terminal productivity. In order to address this problem, we propose Decentralized Cooperative Metaheuristic (DCM), which is a population-based approach that exploits the concepts of communication and grouping. In DCM, the individuals are organized into groups, where each individual shares information with its group partners. This grouping strategy allows to diversify as well as intensify the search in some regions by means of information shared among the individuals of each group. Moreover, the constrained relation for sharing information among individuals through the proposed grouping strategy allows to reduce computational resources in comparison to the 'all to all' communication strategy. The computational experiments for this problem reveal that DCM reports high-quality solutions and identifies promising regions within the search space in short computational times.