Integrated Optimization For Berth And Quay Crane Scheduling In Container Terminals Under Uncertainty

For a container terminal, berth and quay crane are interrelated scarceresources. A reasonable scheduling can improve resource utilization andcustomer satisfaction. Moreover, coping with inevitable operationaldisruptions is crucial for its productivity and service level. This paperstudies the integrated berth allocation and quay crane assignment problem(BACAP) under uncertainty.Based on the literature review and operational practices, the paperdetermines the boundaries and characteristics of the problem studied andbuilds mathematic models of initial schedule under deterministicenvironment and baseline schedule under uncertainty. The objectivefunction of initial schedule model is to minimize the total leaving time ofvessels. By reserving various sizes of time buffers to hedge againstuncertainties, the baseline schedule model is formulated to optimize thequality robustness and solution robustness of baseline schedule, with thesurrogate objective function containing the expected total leaving time andtotal buffer time of vessels.After models are built, the paper designs a proactive-reactivescheduling method. On proactive scheduling phase, a two-level iterativealgorithm by separating decision objects is developed to solve the model:on the first level of the algorithm, the Adaptive Genetic Algorithm ispresented to optimize the priority list of vessels served, the QC numbersallocated and time buffers inserted; on the second level of the algorithm,the paper develops a local search heuristic algorithm to gain the integrated scheduling plan of berth allocation and QC assignment. On reactivescheduling phase, two reactive strategies named “Right-shifting Strategy”and “Heuristic Adjusting Strategy” are designed to deal with disruptionsduring schedule execution.Numeral experiments are conducted to evaluate the validity andsuperiority of proposed models and algorithms and the robustness ofbaseline schedule obtained. More experiments are conducted to analyzehow parameters such as the degree of uncertainty, limit of buffers andreactive strategies influence the decision-making in scheduling.