22. An automatic generation method of maintenance worker schedules at stations and depots

For Shinkansen (bullet trains) and other express trains, maintenance work such as cleaning inside the cars is carried out at stations and depots.
To accomplish this efficiently, it is necessary to have a maintenance worker schedule that assigns multiple work groups to handle maintenance on various train sets. Skilled personnel manually prepare each day's schedule based on the train schedules and vehicle rosters, which differ daily. It takes approximately three hours to prepare a day's worth of schedules, and it is necessary to save labor and de-skill the preparation process.

The maintenance worker schedule must meet various conditions that vary by station and depot, including work conditions such as arrival and departure times and rest periods for each work group, as well as conditions related to interval time between each job (Figure 1).
These conditions have different priorities, and it is not possible to quickly generate high-quality plans by simply formulating and solving them using conventional methods.
Therefore, we developed an automatic generation method of maintenance worker schedules by applying a technique called tabu search, which is capable of generating high-quality solutions quickly.
The developed method generates a tentative schedule that takes into account the priorities of various conditions, and automatically refines the schedule so that the workload is reduced. (Figure 2).

The developed method generates of a day’s worth of maintenance worker schedules in less than three minutes. A comparison between the automatically generated schedule and the schedule prepared by the person in charge revealed that interval time (a key measure of workload) was reduced by an average of 25% at stations and by an average of 56% at depots (Figure 3).
This approach enables labor-saving and de-skilling of scheduling operations, while also reducing the workload for maintenance workers.