Integrating crew scheduling and crew rostering for rail freight with train delays
Preprint
Abstract
Crew scheduling and crew rostering are two key stages in the design of yearly plannings for train drivers. These stages are usually handled sequentially, which may lead to suboptimal solutions. In this work, we address the two stages in an integrated manner for an application to rail freight at the SNCF, the main French railway operator, and propose a method based on column generation that solves both stages simultaneously, while taking into account all the eighteen working rules. We demonstrate the practical relevance of our approach on real freight instances from the SNCF, by displaying substantial cost reductions.Integrating the two stages also makes it possible to better account for uncertainty. We introduce a simple metric to measure the ``fragility'' of rosters with respect to train delays, and we incorporate a stochastic constraint based on this metric into the integrated formulation. Our approach successfully handles this constraint, and experiments on real instances show that the resilience to delays can be improved while maintaining costs at least as good as those obtained with the standard sequential approach in the deterministic setting.Keywords
All Related Versions
This publication has 22 references indexed in Scilit:
- 50 years of Operations Research in public transportEURO Journal on Transportation and Logistics, 2025
- Railway crew scheduling: Models, methods and applicationsEuropean Journal of Operational Research, 2020
- A review on railway delay managementPublic Transport, 2020
- Algorithms for non-linear and stochastic resource constrained shortest pathMathematical Methods of Operations Research, 2018
- Fuzzy-Logic Controlled Genetic Algorithm for the Rail-Freight Crew-Scheduling ProblemKI - Künstliche Intelligenz, 2017
- Integration of duty scheduling and rostering to increase driver satisfactionPublic Transport, 2017
- Cyclic and non-cyclic crew rostering problems in public bus transitOR Spectrum, 2014
- A decomposition approach for the integrated vehicle-crew-roster problem with days-off patternEuropean Journal of Operational Research, 2013
- Genetic algorithm based approach for the integrated airline crew-pairing and rostering problemEuropean Journal of Operational Research, 2009
- A bus crew scheduling system using a set covering formulationTransportation Research Part A: General, 1988