Offline Data-Driven Reinforcement Learning for Efficient Traffic Signal Control

This study introduces an innovative offline data-driven reinforcement learning approach, called DataLight, to optimize traffic signal control and enhance urban mobility.