A NEW MODEL AND HEURISTIC ALGORITHMS FOR THE MULTIPLE-DEPOT VEHICLE SCHEDULING PROBLEM

Abstract

The multiple-depot vehicle scheduling problem (MDVSP) addresses the work of assigning vehicles to serve a given set of time trips with the consideration of certain requirements representing the market rules. Extensive studies in the literature address the MDVSP, but because of the complexity of the problem, the findings of those researchers are still not enough to represent real world situations in Taiwan. Formulation for the MDVSP typically contains the following assumptions: ( 1) the size of fleet or maximum number of available vehicles at each depot is known already, (2) all trip serving costs are usually simplified as a single term in the objective function, which fails to reflect public transit operator concerns, (3) the applied deadheading strategy is static less flexibility, (4) there is no discussion of differences of route change frequency in the problem. This paper presents a new MDVSP model to address the above issues. A greedy heuristic algorithm based on the divide-and-conquer technique is also proposed to solve the MDVSP effectively. Computational tests are performed on the Kinmen Bus Administration (KBA) and results demonstrate that the proposed new model and the greedy heuristic algorithm for the MDVSP are effective in solving real world problems.