Multi-Modal Morning Commute With Endogenous Shared Autonomous Vehicle Penetration Considering Parking Space Constraint

Autonomous vehicles are poised to generate revolutionary impact on urban mobility and parking availability. This study aims to model the morning commute dynamics in a multi-modal transportation system and examine how to regulate the market in the presence of parking space constraint and shared autonomous vehicles (SAVs). Other than the SAVs, commuters can also choose to take regular vehicles (RVs) or public transit (PT) to travel from home to workplace. SAVs will compete the bottleneck capacity with RVs. The dynamic departure patterns and endogenous penetration rate of SAVs are determined depending on parking supply and other related factors. The conditions for various portfolios of travel modes are derived in details. Analytical propositions, numerical examples, as well as policy implications are given with respect to different market factors such as the additional SAV cost, and parking supply. The results show that there are two possible system optimal solutions depending on the value of transit fare. When the transit fare is low, the system optimum occurs in Scenario CS II&III, wherein all three travel modes are used, the first SAV commuter arrives at the bottleneck just when the last RV commuter leaves, and the parking supply and the additional SAV cost should be controlled to a suitable value; when the transit fare is relatively high, the system optimum appears in Scenario IV, an extreme scenario wherein the parking supply is zero, and only the travel modes of PT and SAVs are used. This paper is helpful in quantifying the impacts of new vehicle technologies and shedding lights on future mobility and parking management.