Assessing the operational performance of ridesharing-enabled shared autonomous vehicles fleets

Abstract

Shared autonomous vehicles (SAVs) are quickly spreading in major cities and may become a preferred mobility solution in the near future. Dynamic ridesharing (DRS) is envisioned to enhance the performance of SAV systems by increasing vehicle occupancy and lowering empty vehicle kilometres travelled (VKT). Nonetheless, existing literature assesses the impact of DRS-SAV utilising unrealistic traffic models or heuristic matching and pricing methods. To address this gap, this paper presents a simulation-based service design assessment framework to test real-time SAV operation strategies. Travellers’ mode choices are explicitly modelled, and advanced DRS operation strategies, involving optimal matching and pricing, are tested in a mixed-traffic urban network. The results indicate that advanced DRS methods and accounting for travellers’ mode choices greatly increase the number of served travellers with even smaller VKT, while incurring only a slight increase in waiting and travel time. If properly managed, DRS can significantly reduce the congestion caused by private trips and empty VKT.