Solving the multiple-depot vehicle scheduling problem

The Multiple-Depot Vehicle Scheduling Problem (MDVSP) is the problem where one wants to assign timetabled tasks to vehicles which are housed in multiple depots. This problem is highly relevant for the public transport domain, as competition grows fierce and companies need to stay competitive to survive. As solving the MDVSP optimally is difficult in practice, the aim of this thesis is to give a detailed description for two heuristics discussed in Pepin et al. (2009), namely truncated branch-and-cut and the Lagrangian heuristic. Our implementation of the truncated branch-and-cut heuristic has very poor behavior, which is probably caused by CPLEX’ root node heuristics. Also, in comparison with the heuristics in Pepin et al. (2009), our implementation of the Lagrangian heuristic is rather lackluster, only outperforming their tabu search heuristic. This is probably due to poor code-optimization. Nevertheless, our implementation still provides solutions within 1% optimality gap for all the test instances considered.