Control vs. Congestion: Learning to Untangle Mixed-Autonomy Flow
Session Number
1
Advisor(s)
Ermin Wei, Northwestern University
Location
B115
Discipline
Computer Science
Start Date
15-4-2026 10:15 AM
End Date
15-4-2026 11:00 AM
Abstract
Recent advancements in vehicle autonomy have spurred interest in understanding the impact of autonomous vehicles on traffic systems. In this paper, we study a traffic assignment problem in a mixed-autonomy setting where both human-driven and autonomous vehicles coexist. We model the interaction between the two types of vehicles as a simultaneous routing game, where human drivers act selfishly to minimize their own travel times while autonomous agents are altruistic and aim to minimize the overall social cost. We characterize the equilibrium of this mixed-autonomy game via a variational inequality formulation, extending classical traffic equilibrium formulations to capture the two-player (selfish vs. altruistic) interaction. We analyze the existence, uniqueness, and computation of the equilibrium leveraging VI theory. Further, we investigate the impact of autonomous agents on social cost, identifying conditions under which their introduction leads to improvements, deteriorations, or no effect at all. We find that network structure plays a critical role in determining these outcomes, drawing parallels to Braess' paradox in classical traffic networks
Control vs. Congestion: Learning to Untangle Mixed-Autonomy Flow
B115
Recent advancements in vehicle autonomy have spurred interest in understanding the impact of autonomous vehicles on traffic systems. In this paper, we study a traffic assignment problem in a mixed-autonomy setting where both human-driven and autonomous vehicles coexist. We model the interaction between the two types of vehicles as a simultaneous routing game, where human drivers act selfishly to minimize their own travel times while autonomous agents are altruistic and aim to minimize the overall social cost. We characterize the equilibrium of this mixed-autonomy game via a variational inequality formulation, extending classical traffic equilibrium formulations to capture the two-player (selfish vs. altruistic) interaction. We analyze the existence, uniqueness, and computation of the equilibrium leveraging VI theory. Further, we investigate the impact of autonomous agents on social cost, identifying conditions under which their introduction leads to improvements, deteriorations, or no effect at all. We find that network structure plays a critical role in determining these outcomes, drawing parallels to Braess' paradox in classical traffic networks