Disruption Recovery For Integrated Berth Allocation And Crane Assignment In Container Terminals

Berths and quay cranes are crucial resources for container terminals. The performance of a berthing and crane scheduling plan has direct impact on the service quality and the efficiency of the following operation and transport in container terminals. Uncertainties from outside environment may cause various disruptions in daily operation, which will turn the initial berthing and crane assignment plan into an infeasible one. Therefore, how to make immediate response when a disruption happens is of great significance for container terminals to maintain its competence.Based on disruption management theory and studies on berth allocation and crane assignment in container terminals, a disruption recovery method is proposed in this thesis. The recovery strategy is defined as dynamic scheduling, which refers to a post-act plan based on pragmatic situations. The recovery plan is realized by constrained berth and crane rescheduling. In the recovery plan, the new berthing position and reassigned crane number are determined based on a probability distribution. And the number of reassigned crane is changeable between two working shifts. Besides, vessels requiring early dispatch are also considered. The earnings obtained from early dispatched vessels can be regarded as compensation, namely a reduction for the recovery cost. A bi-objective model is built to present the recovery strategy. The first objective is to minimize the service time and deviation from the original departure time. The second objective is to minimize the recovery cost which includes the horizontal transportation cost, extra operation cost and the profit from vessels requiring early dispatch.A heuristic algorithm based on Squeaky Wheel Optimization(SWO) is developed to solve this model. SWO algorithm has excellent performance on resource allocation problems, which functions by figuring out the bottle neck elements of the problem and move the bottle neck elements forward during resource allocation process. In this thesis, SWO is applied to optimize the berth allocation and crane assignment process by improving the allocation sequence of vessels. In the end, the performance of the algorithm is verified by numerous computational experiments and a comparison between the SWO and First-come-first-serve heuristic is made. Scenario analysis is also conducted to show the impact of the number of delayed vessels, failed cranes and vessels requiring early dispatch. In addition, sensitivity analysis of parameters is also conducted to show their impacts.