A method integrating simulation and reinforcement learning for operation scheduling in container terminals

The objective of operation scheduling in container terminals is to determine a schedule that minimizes time for loading or unloading a given set of containers. This paper presents a method integrating reinforcement learning and simulation to optimize operation scheduling in container terminals. The introduced method uses a simulation model to construct the system environment while the Q-learning algorithm (reinforcement learning algorithm) is applied to learn optimal dispatching rules for different equipment (e.g. yard cranes, yard trailers). The optimal scheduling scheme is obtained by the interaction of the Q-learning algorithm and simulation environment. To evaluate the effectiveness of the proposed method, a lower bound is calculated considering the characteristics of the scheduling problem in container terminals. Finally, numerical experiments are provided to illustrate the validity of the proposed method.