A New Detour Strategy in Ride-Hailing Markets with Mixed Autonomy Considering Congestion Externalities

With the rapid development of urban economies, the soaring number of ride-hailing vehicles has exacerbated network congestion during peak hours. In recent years, several companies have begun exploring the use of autonomous vehicles in on-demand ride services. By implementing appropriate rerouting strategies to manage the routes of shared autonomous vehicles (SAVs), it may be possible to alleviate road congestion during high-demand conditions. In a mixed ride-hailing market that includes both human-driven vehicles (HVs) and SAVs, we propose a novel SAV management strategy for peak periods and formulate a bi-level model to describe the market equilibrium. Based on differences in routing strategies, we distinguish between HVs and SAVs and develop a traffic assignment model at the lower level. At the upper level, we characterize the equilibrium state of the mixed market and formulate an idle SAV repositioning model under the proposed strategy. The model incorporates elastic passenger demand and choice preferences and derives the optimal management decision of the ride-hailing platform aimed at maximizing profit. Our study evaluates the impacts of the proposed strategy on the road network, the platform, passengers, and HV drivers, and demonstrates that it can support government goals by reducing total system time costs.