Maintenance scheduling in rolling stock circulations in rapid transit networks

The railway routing problem determines specific paths for each individual train, given its type and composition and considering possible maintenance locations and durations. The objective is to minimize operating costs and penalties related to waiting times and maintenance all while considering train scheduling and maintenance constraints. The model is solved using Branch and Bound and Column Generation approaches. In the paper the different approaches are compared for different planning horizons and model parameter settings. The computational tests have been run in a real RENFE network