A Study for Solving Multi-Depot Dial-a-Ride Problem Considering Soft Time Window

Dial-a-ride is the most widely available transit service for disabled persons or seniors in the United States and Europe. This paper studies a static dial-a-ride problem considering multiple depots, heterogeneous vehicles, and soft time windows. In this paper, we apply a heuristic based on clustering first-routing second(HCR) to a real-world large dial-a-ride problem from Maryland Transit Administration(MTA). MTA`s real operation is compared with the results of developed heuristic for 24 cases. The objective function of the proposed model is to minimize the total cost composed of the service provider`s cost and the customers` inconvenience cost. For the comparison, the objective function values of HCR do not include waiting cost, delay cost, and excess ride cost. The objective function values from HCR are better than those from MTA`s operation for all cases. This result shows that our heuristic method can make the real operation better and more efficient.