Adaptive large neighborhood search for metro train timetable rescheduling in power-limited areas

Traction power supply faults in a metro system may lead to a power supply shortage in a local area, which limits the total traction power of trains and reduces the line capacity utilization. To mitigate these impacts, this paper proposes a train timetable rescheduling approach for the power-limited area. Firstly, considering dynamic train power and regenerative power utilization, a 0–1 integer linear programming model to reschedule arrival and departure times is formulated to maximize the line capacity utilization under power supply capacity constraints. Then, an adaptive large neighborhood search algorithm is developed, incorporating three sets of destroy/repair operators that coordinate train operations through staggered traction and traction-braking matching. According to the comparative simulations based on the real-world data from Beijing Metro Line 7, the proposed approach improves the line capacity utilization by 46.58%–70.79% with dynamic power consideration, and by 0.58%–3.22% when further incorporating the regenerative power.