Scheduling method for demand-responsive transportation system

This paper offers an addition to a family of previously proposed vehicle-scheduling models for a demand-responsive transportation system. The assumed operating environment consists of many-to-many travel demand, multi-vehicle, advanced reservation, and time-window constraint for pickup and drop-off times. The model builds vehicle schedules one vehicle at a time, maximizing the number of trip requests to be handled by each vehicle. The method follows a two-step approach: building of an initial route, and insertion of additional trips into the initial route. In inserting trips to the initial route, the model checks feasibility of insertion for each trip first, and then identifies the maximum number of trips that can be inserted using a tree-structured search technique. The proposed method is evaluated for a set of specific examples by examining the sum of time differences between the original and actual pickup and drop-off times, and the number of vehicles required. The model is relatively straightforward in its computational procedure and offers possibilities of other applications.