Distributed Traffic Signal Control with Fairness Using Deep Reinforcement Learning

Currently, traffic congestion is a frequent phenomenon, owing to the increased traffic volume in road-traffic networks. To alleviate traffic congestion, it is important to control traffic signals appropriately. In this study, we consider each traffic signal as an agent and adopt a distributed control method in which neighboring agents cooperate with each other locally. In traffic signal control, it is important to control traffic signals with fairness, so that the variation in the average waiting times along different roads is minimized. In this study, we propose a method that applies deep reinforcement learning to achieve control that satisfies this fairness.