What is it about?
We study regular and chaotic behaviors in large population games capturing congestion effects, such as road traffic. In realistic settings with diverse self‑interested agents that dynamically adapt their choices of shared common resources, we prove that certain conditions, such as heavier overall demand or faster adaptation rates, can make day‑to‑day congestion patterns destabilize and become unpredictable, leading to inefficient utilization of resources. Interestingly, upon averaging out the system’s behavior over time, despite day-to-day unpredictability, the resulting traffic flows still match the efficient Nash equilibrium predicted by standard static game theory. This emergence of macroscopic order mirrors thermodynamics, where molecular chaos averages out to steady pressure, except now the passive particles are active decision‑making programs chasing their own goals in their own ways.
Featured Image
Photo by Kyndall Ramirez on Unsplash
Read the Original
This page is a summary of: Heterogeneity, reinforcement learning, and chaos in population games, Proceedings of the National Academy of Sciences, June 2025, Proceedings of the National Academy of Sciences,
DOI: 10.1073/pnas.2319929121.
You can read the full text:
Contributors
The following have contributed to this page
