Optimization Approach to Rolling Stock Scheduling for Timetable Temporarily Changed due to Disaster

Abstract

When railway operators devise a rolling stock schedule plan, it is desirable that the schedule plan for the new timetable is similar to the previous schedule. In this study, we focus on rolling stock scheduling following a large-scale natural disaster. Generally, partial suspensions occur some sections in the line from the damage caused by the disaster, so a temporary timetable is generated. Subsequently, a rolling stock schedule that is close to the previous one must be created. We propose a two-phase rolling stock scheduling algorithm for the temporarily changed timetable. The proposed scheduling algorithm creates a set of duties based on the multi-commodity network flow problem in the first phase. Following that, the algorithm provides a set of rosters by solving the travelling salesman problem. Finally, we confirm that the proposed algorithm can produce a good solution in terms of evaluation criteria and computational time.