Simulation-Based Pattern Timetabling for Short-Turning Trains in a Metro Network

Effective coordination of train services across different lines of a metro network may significantly improve service quality. To this end, we introduce the Service Pattern Timetabling Problem for a Metro Network (SPTPN). In this problem a periodic timetable is established for each line so as to minimize the average passenger waiting time. The periodic timetables constructed by allowing trains to short-turning strategy (i.e., trains are allowed to reverse direction before reaching the line's terminal station). Thus, we name the resulting periodic timetables as patterns. We develop an iterated local search algorithm to solve the SPTPN, featuring four tailored neighborhood operators. Solutions are evaluated using an event-driven simulation model that accurately captures train operations and passenger behaviors under capacity constraints. Finally, the effectiveness of the proposed model and solution algorithm is demonstrated through numerical experiments.