Reinforcement learning for train timetable rescheduling under perturbation: A general value-based approach

Key Points

• The value-based approach effectively reschedules train timetables under perturbation, enhancing operational flow.
• A notable reduction in delay times by 30% was observed during simulations with varying perturbations.
• Assessment using a reinforcement learning framework unveils novel strategies for rescheduling amidst disruptions.
• This model may enable faster recovery times for train services, yet validation in real-world scenarios is needed.