Meta-Heuristics for the Design of a Demand-Responsive Transit Line

Traditional public transport is evolving towards more flexible services and operation modes in order to better serve the needs of the population, capture additional demand, and increase its profitability. The transportation system that we study integrates traditional bus transportation and on-demand services. Demand-responsive lines provide a classical, scheduled, transportation service among a limited number of specific points. Such compulsory stops are locations where a demand for transportation may be naturally assumed to be present: stops of express lines, rail or subway stations, or any other points which must be serviced according to the policies of the local transit authority. The vehicle is allowed to transit by each compulsory stop during a time window. Beside the compulsory stops, a set of optional stops to be activated on demand is available to users. Between each pair of consecutive compulsory stops a set of optional stops is defined, which can be served by the vehicle on its way from a compulsory stop to the next. A user issues a service request specifying the pick up and drop off stops. In the absence of requests involving optional stops, the vehicle travels along the shortest path on the network within each pair of consecutive compulsory stops. The acceptance of a request implies the rerouting of the vehicle for that part of route involving the optional stop(s) related to the request. The detour may cause a delay of the transit time at the following stops. Such a Demand Adaptive System (das) represents a good compromise between an expensive personalized service that precisely fulfills individual requests, and the cheap alternative supplied by the traditional public transport, that usually does not provide transportation along individual itineraries. It is a system that may well address the needs of relatively low demand areas (e.g., North American city suburbs) or periods (e.g., non-rush hours, evenings, week ends, etc.).