Can driving automation make demand responsive transit viable? A strategic modeling approach

This research investigates the viability and potential advantages of deploying a Demand-Responsive Transit (DRT) service in low-demand areas, with a particular focus on the role of driverless technology. Central to the analysis is a strategic modeling framework that captures the interaction between a transit regulator and a service operator. We propose a two-stage game-theoretic model in which the regulator sets the fare per kilometer and the operator responds by optimizing operations through an integer linear programming (ILP) model. This framework integrates demand estimation based on fare and service attributes with operational decision-making, enabling a systematic and computationally efficient evaluation of policy scenarios without relying on complex agent-based simulations. Findings reveal that although human-operated DRT services can improve consumer surplus, they are economically inefficient due to high operational costs and required subsidies, which often lead to their failure. In contrast, driverless DRT services emerge as a viable alternative, with substantially lower costs and improved social welfare compared to human-operated DRT, especially when using low-capacity vehicles. However, despite their advantages, driverless DRT services still require significantly higher subsidies than fixed-route public transportation. Furthermore, optimizing fare levels remains essential to balancing financial sustainability with maximizing social welfare. Our results suggest that a moderate fare increase of approximately 50 % is necessary to enhance social welfare, though the service remains dependent on external funding. Therefore, policymakers and transit agencies should explore the potential of driverless technology while carefully considering pricing strategies and subsidy structures to ensure long-term feasibility in rural and low-density areas.