Berth Allocation Based on Improved Memetic Algorithm in Container Terminal

The berth allocation problem in container terminal is a NP-hard problem. In past, the genetic algorithm was often adopted to solve the problem. The genetic algorithm initially seemed to be quite useful at finding generally good global solutions but later proved to be quite inefficient at finding the absolute optimum solution in last few mutations. Therefore, an improved memetic algorithm was introduced in the present paper. In basic genetic operations, an improved order crossover and a new mutation based on neighborhood search was employed to avoid infeasible solutions. Moreover, an improved simulated annealing algorithm was utilized for the local search. Finally, various numerical examples were analyzed, which showed that compared to the traditional genetic algorithm the proposed memetic algorithm can speed up the convergence and is more suitable for the berth allocation problem in container terminal.