Heuristic algorithms for the handicapped persons transportation problem

The authors examine the problem of determining an optimal schedule for a fleet of vehicles used to transport handicapped persons in an urban area. The problem is a generalization of the well-known advance-request Pickup and Delivery Problem with Time Windows. Due to the high level of service required by this kind of transport, several additional operational constraints must be considered. The problem is NP-hard in the strong sense, and exact approaches for the solution of real-life problems (typically with hundreds of users to be transported) are not practicable. The authors describe a fast and effective parallel insertion heuristic algorithm which is able to determine good solutions for real-world instances of the problem in a few seconds on a personal computer. They also present a Tabu Thresholding procedure which can be used to improve the starting solution obtained by the insertion algorithm. The application of the proposed procedures to a set of real-life instances for the city of Bologna, involving about 300 trips each, is also discussed. The heuristic algorithms obtain very good results compared with the hand-made schedules, both in terms of service quality (all the service requirements are met) and overall cost.