Iterated local search for the multiple depot vehicle scheduling problem

The multiple depot vehicle scheduling problem (MDVSP) is a well-known and important problem arising in public transport. Although many solution approaches have been published in the literature, algorithms using metaheuristics appeared only very recently (Large Neighborhood Search and Tabu Search). In this paper, we introduce an Iterated Local Search algorithm for the MDVSP, incorporating a neighborhood schema called ”block moves”, based on the notion of ejection chains. Using a set of benchmark instances, we show empirically that the proposed algorithm performs better than the best metaheuristics implemented so far and obtains high quality results within short computational times.