港口航道与近海工程2025，Vol.62Issue(6)：5-11,7.DOI:10.16403/j.cnki.ggjs20250602
基于强化学习的自动化集装箱码头混行交通调度算法研究
Research on Scheduling Algorithm for Mixed Traffic in Automated Container Terminals Based on Reinforcement Learning
摘要
Abstract
With the deepening of automation transformation of container terminals,avoiding traffic conflicts between trucks with or without operators is the key to improving terminal operation efficiency and ensuring production safety.In view of the difficulty of traditional scheduling methods in coordinating the spatiotemporal conflicts of heterogeneous trucks and the randomness of driver behavior,this study proposed an intelligent scheduling method based on deep reinforcement learning and developed a masked proximal policy optimization algorithm(PPO-MASK).The algorithm uses the action mask mechanism to mask invalid actions,effectively solving the problem of traffic flow conflicts between trucks with or without operators in the port area.Experimental results show that the algorithm shows good performance in the operating environment of large container ports,with at least 5%performance improvement compared to other scheduling algorithms,providing a highly adaptable decision support algorithm for the scheduling of horizontal transportation systems in smart ports.关键词
自动化码头/混行交通/无人集卡/调度优化/AnyLogic仿真/强化学习Key words
automated terminal/mixed traffic/unmanned container truck/scheduling optimization/AnyLogic simulation/reinforcement learning分类
交通工程引用本文复制引用
Sun Xubo,Xu Xinglu,Li Jia,Wang Wenyuan..基于强化学习的自动化集装箱码头混行交通调度算法研究[J].港口航道与近海工程,2025,62(6):5-11,7.基金项目
国家自然科学基金项目(52301314 （）
52272318) （）
