城市轨道交通列车时刻表和车底运用计划研究进展与综述

Review on Train Timetabling and Rolling Stock Circulation Planning Problems for Urban Rail Transit Systems

列车时刻表与车底运用计划是城市轨道交通运营规划的重要环节,也是运营企业实现降本、增效、提质的基础。围绕该领域的最新研究进展,从研究问题、建模方法和求解算法三个方面,对列车时刻表和车底运用计划问题进行了系统梳理和综述。在研究问题方面,既有文献侧重于列车时刻表优化、时刻表与车底运用协同优化,以及可变编组条件下的列车时刻表优化。在建模方法方面,文献主要采用一般整数规划和离散时空网络两类方法。在求解算法方面,启发式方法因其操作简便、灵活高效而被广泛应用;精确算法因能提供全局搜索和求解质量的评估机制,成为当前常用的方法;此外,强化学习方法在求解上述问题中也得到初步应用。未来研究需要关注通信、控制等技术变革带来的列车运营模式创新及其衍生的新问题,探讨城市轨道交通与其它交通方式的一体化融合,设计高性能求解算法以应对大规模城市轨道交通运营管理问题的计算挑战。

The train timetable and rolling stock circulation plan are crucial aspects of urban rail transit operation planning which serve as the foundation for operating companies to reduce costs,improve efficiency,and enhance quality.This paper systematically reviews and summarizes the latest research advancements in the field of train timetabling and rolling stock planning problems.The review is organized from three perspectives:studied problems,modeling methods,and solution algorithms.In terms of studied problems,existing literature focused on train timetabling optimization,the joint optimization of the train timetable and rolling stock circulation plan,and train timetabling optimization with variable train compositions.Regarding modeling methods,conventional integer programming and discrete space-time network approaches have emerged as the two mainstream approaches for modeling train timetabling and rolling stock circulation planning problems.In terms of solution algorithms,heuristic methods are widely applied due to their simplicity,flexibility,and efficiency in solving train timetabling and rolling stock circulation planning problems.Exact solution algorithms,which offer global search and solution quality assessment mechanisms,are currently mainstream.Additionally,reinforcement learning methods have also been preliminary applied to solving the aforementioned problems.Future research needs to focus on the innovation of train operation modes brought about by technological changes in communication,control,and other areas,as well as the new problems derived from them.It is necessary to explore the integration of urban rail transit with other transportation modes,and design high-performance solving algorithms to address the computational challenges of large-scale urban rail transit operation management issues.