A Multi-Agent System for Dynamic Urban Traffic Regulation

Abstract

Traffic congestion requires adaptive signal timing. We outline a decentralized multi-agent reinforcement learning algorithm that coordinates green light sequences dynamically.