An integrated scheduling method for AGVs in an automated container terminal considering battery swapping and battery degradation

Automated guided vehicles (AGVs) are powered by batteries that degrade over time. AGV batteries are procured in batches in automated container terminals, resulting in variations in battery degradation across different batches. It is crucial to consider the differences in the degradation states of AGV batteries while developing task scheduling and battery swap time. This article introduces an integrated scheduling model that incorporates task assignment and battery swap time for AGVs, specifically considering batteries in varying degradation states with the aim of minimizing task completion time. A hybrid genetic algorithm (HGA) with neighborhood search (NS) combined with a look-ahead greedy approach was used to solve the model, and numerical experiments confirmed the applicability of the proposed model. This study presents an effective scheduling method for handling task dispatching and battery swap plans in automated terminals. The performance of the HGA is compared with several methods, including: (i) solving the underlying mixed integer linear programming (MILP) exactly with CPLEX, (ii) the current scheduling method (YP) employed in Shanghai Yangshan Port, (iii) an HGA without NS procedure, (iv) a genetic algorithm (GA) without the NS procedure and look-ahead greedy approach, and (v) a matheuristic approach (MA) from the literature. Experiment results indicate that the proposed HGA offers significantly shorter computational times compared to solving the underlying MILP exactly with CPLEX. Additionally, in terms of solution quality, the HGA outperforms the YP, HGA without NS, GA, and MA by 46.19%, 2.96%, 20.62%, and 21.32%, respectively.