Coordinating Fleet Sizing and Management for Mixed On-Demand Ride Services of Autonomous and Human-Driven Vehicles

With the advent of mixed on-demand ride services where autonomous vehicles (AVs) are jointly managed with human-driven vehicles (HVs), the purpose of this study is to size and operate a mixed fleet of AVs and HVs, as well as to explore the influence of AV penetration rate on mixed ride services. In light of the decision interaction between the service provider (i.e., the platform) and HV drivers, a bi-level model is formulated to coordinate strategic fleet sizing and tactical operations. The upper-level problem aims to determine the fleet sizing, vehicle assignment and centralized rebalancing of AVs, thereby increasing the total profit of the mixed ride service system. Upon the determination of upper-level decisions, the decentralized passenger-searching movements of HVs are modeled by the lower-level problem, where each HV aims to maximize its own utility. A tailored reformulation algorithm is presented to address the bi-level programming problem. The real-world case study in Chengdu illustrates the potential benefits of incorporating AVs into on-demand ride services. The findings provide case-specific insights into developing a mixed ride service system throughout the prolonged transition phase.