Abstract

The timetable is vulnerable to the impact of emergencies such as severe weather and equipment faults, potentially disrupting train services. Swift and efficient timetable adjustments are crucial for maintaining the operational efficiency of high-speed railways (HSRs). This study introduces a railway traffic management advisory system (RTMAS) tailored for high-speed trains, aiding dispatchers in the rapid rescheduling of timetables during emergencies. The RTMAS's structure and its emergency handling protocols are designed with an emphasis on human-in-the-loop and user-friendly principles. The core modules, including risk event scenarios, delay time prediction, expert systems, rescheduling scheme generation, and strategy evaluation systems, and their functions are given from the whole process of the dispatchers' emergency responses. The system employs metaheuristic algorithms and deep reinforcement learning (DRL) for predicting delays and generating rescheduling plans. A practical examination of the Beijing-Shanghai HSR Line validates the effectiveness of the RTMAS. The results indicate that the system not only furnishes dispatchers with near-optimal rescheduling solutions but also markedly diminishes total delays and the number of trains impacted by emergencies.