A bi-layer model for berth allocation problem based on proactive-reactive strategy

Abstract

To reduce the impact of equipment failure, vessel arrival time delay and other unexpected factors, container terminal managers need to develop a more robust berth scheduling plan while considering the recovery costs. To deal with the berth allocation problem (BAP) with uncertain arrival time, a proactive-reactive hybrid strategy to deal with the impact of uncertainty is proposed. According to the characteristics of BAP and proactive-reactive strategy, a bi-layer programming model is constructed. A robust baseline plan with buffer time is generated in the first layer, and a recovered scheduling plan is generated in the second layer according to the actual arrival time. The decisions between the two layers of models affect each other. In addition, an Adaptive cross-entropy algorithm (ACEA) is designed to solve the first layer model. Different from the traditional cross-entropy algorithm (CEA), an adaptive search strategy is added to further search and optimize the neighborhood of the elite solutions. Furthermore, heuristic adjustment strategies (HAS) are designed to deal with the second layer. The computational results demonstrate that the bi-layer model proposed in this paper can better balance the baseline schedule cost and the recovered schedule cost. Numerical experiments are carried out to assess the effectiveness of the proposed model and the efficiency of the proposed algorithms.