A Dynamic Berth Allocation Model Based on Stochastic Consideration

Abstract

When allocating berth in terminal, inputs such as quay-crane handling time are stochastic parameters. A dynamic stochastic berth allocation model was proposed to simulate the real decision-making. To decrease total ship waiting time, the model ignores the "first comes first served" rule, and copes with the corresponding evil consequence of exceeding acceptable waiting time of ships with probability. Since consuming CPU time for solving the model is beyond polynomial time, a heuristics algorithm was developed based on genetic algorithm. And a reduced basic search space for the algorithm was discussed by considering approximate optimal solution properties. The experiment results reveal that the model is capable of allocating berths efficiently and dynamically to calling containerships in real stochastic environments and the algorithm can lead to satisfactory results in allowable CPU time.