Effectiveness of Dynamic Insertion Scheduling Strategy for Demand-Responsive Paratransit Vehicles Using Agent-Based Simulation

This paper deals with the scheduling of paratransit vehicles. The current scheduling method utilized by paratransit providers is to provide a door-to-door ride for those customers who have made reservations. Thus, the paratransit providers know in advance the pickup and drop-off locations of each customer. Using this information, they are able to determine a route for each vehicle to minimize the total operating costs. In the current scheduling method, vehicles are not allowed to pick up unscheduled customers. This practice often leads to low seat utilization. To address this shortcoming, this paper explores the idea of allowing vehicles to pick up unscheduled customers who are in close proximity to the prescheduled stops (referred to as the dynamic response area or DRA). To this end, this paper develops an agent-based simulation model to evaluate the effectiveness of this strategy. The model was tested using the Chicago network. The results of the simulation experiments indicate that (1) the proposed strategy is able to serve more customers using the same fleet size, and (2) the proposed strategy will not significantly affect the scheduled customers' in-vehicle travel time.