When Altruism Meets Autonomy: Managing Bottleneck Congestion with Strategic Autonomous Vehicles

A new arXiv paper (2604.21941) tackles the bottleneck congestion problem at highway weaving ramps, where conflicting traffic flows create complex interactions between human-driven vehicles (HDVs) and autonomous vehicles (AVs). The authors first formulate a Wardrop-based model capturing selfish HDV behavior, establishing existence, uniqueness, and validity of the resulting equilibrium. They then introduce a Stackelberg-Wardrop formulation in which AVs act as strategic leaders to optimize system performance, while HDVs respond through equilibrium adaptation. The framework also incorporates heterogeneous behavioral preferences via a Social Value Orientation (SVO) model, allowing AVs to exhibit altruistic or selfish traits.

The paper's key finding: under selfish HDV behavior, AV penetration has a non-increasing impact on system performance. Instead of smooth improvements, traffic flow remains unchanged until AVs hit critical penetration thresholds, at which point efficiency gains suddenly emerge. This structural property provides principled guidance for designing AV control and incentive mechanisms. The work demonstrates how strategically controlled autonomous agents can be deployed to induce system-level efficiency gains in mixed-autonomy transportation networks, potentially reducing congestion without requiring full autonomy. The implications for smart city planning and AV fleet management are significant, as it suggests that modest AV adoption may be ineffective—only reaching specific penetration levels unlocks meaningful benefits.