Modeling and Managing Mixed On-Demand Ride Services of Human-Driven Vehicles and Autonomous Vehicles

We model a monopoly ride-sourcing market where the platform adopts the service types of human-driven vehicles (HVs) and autonomous vehicles (AVs). Both congestion externality under mixed traffic flow and heterogeneity on riders' perceived utility of the ride-sourcing service are considered when formulating the mode choice behavior of riders. We analyze the impact of the platform's fleet size and its price for riders on demand rates and riders' waiting time in the market equilibrium state. The analytical results show that the demand rates of mixed on-demand ride service types are not necessarily monotonous to the price for riders or the fleet size, due to the congestion externality and existence of a wild goose chase regime. Under either profit maximization or welfare maximization strategies, numerical results demonstrate that a higher pure AV traffic flow capacity benefits human ride-sourcing drivers and both types of riders. The platform should arrange more AVs than HVs even under the high AV depreciation cost. In a surging demand scenario, the platform should encourage riders to switch from the HV service to the AV service through price regulation. Moreover, an extended scenario considering the integrated service is discussed. The economic analysis and gained managerial insights benefit the on-demand ride services platform's decision making on operational strategies of HVs and investment in AVs with mixed traffic congestion effects.