Joint Locomotive Scheduling and Driver Rostering in Rail Freight Traffic

Joint Locomotive Scheduling and Driver Rostering in Rail Freight Traffic

This project aims at developing solution methods for problems arising in the operations management of the railway companies. More specifically, locomotives and their drivers need to be assigned to trains in order to generate shift plans. These in turn need to consider working time regulations, locomotive maintenance needs as well as the compatibilities between trains, drivers and locomotives. Also, some of the company-specific planning requirements need to be taken into account.

For that end, we develop a large and complex binary model. It combines the structures of set covering with multiple-choice constraints with multicommodity-flow problem. We also devise a novel decomposition approach to allow for the solution of the problem, dividing the model into a master locomotive scheduling problem and driver rostering subproblem. We also supplement it with a presolve heuristic for the difficult subproblem.

When implemented into the planning systems of the industrial partner, our approach will enable efficiency increases. It will also allow for a significant simplification of planning process.

This project is developed in collaboration with DB Cargo Polska. It is a company from the network of DB Cargo AG. With 2800 employees, 220 locomotives and 2820 wagons, it is the second largest rail freight carrier in Poland in terms of weight carried.

Our project is a part of a ROMSOC – an European Industrial Doctorate program aiming at development of novel mathematical modeling, simulation, and optimization as well as model order reduction techniques for numerous industrial applications (including adaptive optics, lightning, microchip design, finance and others).

Friedrich-Alexander-Universität Erlangen-Nürnberg