Optimal Matching Problem for Ride-Sharing Considering Users’ Schedule Rearrangement

Abstract

This paper proposed an optimal matching problem between drivers and riders for ride-sharing. The model assumes that all Space-time Extended Network (STEN) links have measurable utility based on the activity desired by each user. The ride-sharing service is designed to maximise the social utility of all users while considering the capacity constraints of vehicles. As user activity over the entire study period is taken into account, the model accommodates changes, and even cancellations, in the timing of user activity without addition of any further parameters. Furthermore, the model allows evaluation of the diffusion effects of autonomous vehicles used for ride-sharing. A case study using a hypothetical network confirmed that autonomous vehicle diffusion not only reduced detours during pick-up, but also encouraged riders to travel. Vehicle diffusion increased total user utility and helped to equalise opportunities for user activities. Although autonomous vehicle diffusion may increase total travel time, this can be mitigated by vehicle-sharing.