Divide-and-price: A decomposition algorithm for solving large railway crew scheduling problems

The railway crew scheduling problem consists of generating crew duties to operate trains at minimal cost, while meeting all work regulations and operational requirements. Typically, a railway operation uses tens of thousands of train movements (trips) and requires thousands of crew members to be assigned to these trips. Despite the large size of the problem, crew schedules need to be generated in short time, because large parts of the train schedule are not finalized until few days before operation.We present a column generation based decomposition algorithm which achieves high-quality solutions at reasonable runtimes. Our divide-and-price algorithm decomposes the problem into overlapping regions which are optimized in parallel. A trip belonging to several regions is initially assigned to one region (“divide”). The corresponding dual information from optimization is then used as a bonus to offer the trip to other regions (“price”). Pricing and assignment of trips are dynamically updated in the course of the optimization. Tests of our algorithm on large-scale problem instances of a major European freight railway carrier yielded promising results.

► Decomposition of crew scheduling instances substantially accelerates column generation algorithms.
► Potential loss of quality is met by allowing overlapping of regions.
► Trips are assigned to regions based on dual information.
► Simultaneous optimization of regions enables dynamic updating of the decomposition.