Fleet sizing and pricing for hybrid ownership of shared autonomous vehicles in a multimodal transportation system

Shared privately-owned autonomous vehicles (SPAVs) are expected to be a primary alternative to existing car-sharing and on-demand mobility services when fully automated vehicles are soon to be privatized. In this paradigm, PAV owners make decisions on whether/when/where/for how long to share their PAVs, which would make a dynamic fleet of SPAVs. To improve the effectiveness of such a mobility system, a certain fleet of shared business-owned autonomous vehicles (SBAVs) may also be required. This study proposes a bi-level programming model considering the hybrid ownership of shared autonomous vehicles (SAVs). The upper level optimizes the fleet size of SBAVs and rental price of SAVs involving an endogenous fleet of SPAVs, and the lower level involves an SAV self-relocation strategy and attends to travelers’ daily activity-travel scheduling behaviors. A memetic algorithm involving iterative adjustments of the supply and demand sides is developed for the bi-level model. Results from numerical examples indicate that SBAVs provided by the operator complement SPAVs to serve the mobility demand to improve system performance, while PAVs owners adapt sharing behaviors and daily schedules to reduce activity-travel disutilities.