It's All in the Mix: Technology Choice between Driverless and Human-Driven Vehicles in Sharing Systems

Abstract

Operators of vehicle-sharing systems such as carsharing or ride-hailing can benefit from integrating driverless vehicles into their fleet. In this context, we study the impact of optimal fleet size and composition on an operator's profitability, which entails a non-trivial tradeoff between operational benefits and higher upfront investment for driverless vehicles. We analyze a strategic fleet sizing and composition problem, integrating a rebalancing problem, which we formalize as a Markov decision process. We incorporate the rebalancing problem with a time-dependent fluid approximation to devise a scalable linear programming solution approach, which we improve by state-dependent emergency rebalancing. We present a numerical study on artificial and real-world instances that reveals significant profit improvement potential of driverless and mixed fleets compared to human-driven fleets. For real-world instances, the profit improvement amounts up to 20.4% over exclusively human-driven fleets. If both vehicle types incur equal operational costs, operators optimally mix a small number of driverless vehicles with a large number of human-driven vehicles. Mixed fleets are particularly beneficial if demand varies over time, and operators consequently shift rebalancing to lower-demand periods.

Keywords: transportation, autonomous mobility-on-demand, mixed autonomy, fluid approximations, rebalancing, time-dependent demand

JEL Classification: C44