Railway rolling-stock-assignment-scheduling algorithm for minimizing inspection cost

Train maintenance regulation requires regular inspections of the rolling stock of railway companies. A planner who makes a schedule of rolling-stock operation should consider both assignments of rolling stock to train services and schedules of periodical maintenance. In Japan, the schedule is manually created for the middle-term, which is for half a month or full month. To reduce spent time to make the schedule, we developed an algorithm for rolling-stock-assignment schedule. To efficiently consider inspection cycles, the proposed algorithm creates assignment patterns of rolling stock which satisfy inspection cycles in advance. And then, the algorithm creates templates of schedule using the patterns, and assigns inspections to the templates. We verified the feasibility and scalability of the proposed algorithm using actual train operation data. The results indicate that the algorithm successfully generates a schedule which satisfies required regular inspection cycles within feasible calculation time.