Battery Swapping Strategy for Electric Transfer-Vehicles in Seaport: A Deep Q-Network Approach

To achieve green seaports, electric transfer-vehicles (ETVs) are introduced as an alternative to traditional diesel-powered vehicles for transferring cargo. This paper proposes a two-stage energy-transport scheduling method to meet the transferring jobs while minimizing the charging cost consisted of vehicle-job matching, vehicle dispatch and energy management. The optimal matching between vehicles and jobs is made in the first stage. Based on the job lists obtained in the first stage, ETVs are dynamically dispatched to minimize the job delay and charging cost in the second stage. The two-stage energy-transport scheduling model is formulated and solved using a deep Qnetwork method. Simulation experiments with realistic data derived from one real-world seaport in northern China verify the effectiveness of the proposed method. Comparative results demonstrate that cost reductions of 25.8% and 50.37% can be obtained using the proposed method, comparing to mathematical programming and rule-based matching case, respectively.