Titolo della tesi: System Optimal Ride-sharing and Traffic Equilibrium
This thesis explores socially optimal ridesharing schemes and their potential to reduce congestion and emissions. The work presents a framework where a social planner manages ride-matching between passengers and drivers to minimize travel costs (fuel and travel time) and environmental costs. The study focuses on both the theoretical development and practical implementation of these schemes, offering solutions for static networks as well as dynamic and congested urban networks. The core contributions lie in optimizing matching systems, ensuring stable and fair pricing mechanisms, and also incorporating financial sustainability.
The first chapter of the thesis proposes a static ridesharing model where a social planner optimally matches passengers with drivers. The objective is to minimize total travel and environmental costs within a network. A linear programming problem is formulated to compute the optimal matches, and the existence and uniqueness of solutions are examined. Importantly, no participant incurs a loss compared to traveling alone, as traveling alone is always an option. The chapter also proves that without environmental costs, the optimal matching is stable, meaning no rider-driver pair has an incentive to switch partners. Using the Sioux Falls network, the results show that even with modest participation, significant reductions in CO2 emissions can be achieved.
The second chapter introduces a peer-to-peer ride-sharing scheme designed to minimize total car-kilometers traveled while ensuring stable and fair matchings between passengers and drivers. The optimization problem is solved using a mixed-integer linear programming approach, and a pricing scheme based on fair surplus division is proposed. The surplus each traveler gains is midway between the minimum and maximum they could achieve in stable solutions, ensuring financial sustainability without subsidies. Empirical analysis using toy networks and the Sioux-Falls network demonstrates how surplus-based pricing can achieve efficient and fair cost-sharing.
The final chapter examines the implementation of ridesharing systems in densely populated urban areas, using a real-world case study from the Île-de-France Network. The study presents a framework that integrates mode choice, route assignment, and role flexibility in driver-passenger matching to minimize social costs and reduce emissions. By employing time-dependent travel predictions and optimizing driver-passenger matchings within short time windows, the model helps reduce congestion, and total kilometers traveled. Additionally, it uses an innovative link-specific, speed-based emission model to calculate each traveler's CO2 emissions based on their chosen route. The results indicate that this approach can lead to significant improvements in both environmental and traffic conditions, making ridesharing a more sustainable alternative in large urban networks.