Abstract
The goal of this research project is to improve the operational efficiency of shared-ride mobility-on-demand services (SRMoDS). SRMoDS ranging from UberPool to micro-transit have the potential to provide travelers mobility benefits that are comparable to existing ride-hailing services without shared rides such as UberX, but at a lower cost and with fewer harmful externalities. To meet the project’s goal, this study proposes a bi-criteria network pathfinding approach that considers proximity to potential future traveler requests in addition to travel time. This pathfinding approach was built on top of a state-of-the-art dynamic vehicle routing and matching modules. The study tests the proposed pathfinding approach using the network of the City of Anaheim. The results indicate that the proposed bi-criteria pathfinding can potentially reduce both traveler waiting and in-vehicle travel time; however, the effectiveness depends on several factors. Important factors include the relative supply-demand imbalance as well as several hyperparameters in the optimization-based control policy. Moreover, the results indicate that the bi-criteria policy is only advisable when the SRMoDS vehicle has one or fewer in-vehicle passengers. Although the operational benefits found in this study are relatively small, future research efforts related to tuning hyperparameters should allow bi-criteria pathfinding to significantly improve SRMoDS.