A Multi-agent Traffic Signal Control System Using Reinforcement Learning

This paper presents a control method based on multi-agent for traffic signals. A reinforcement learning algorithm is used to optimize traffic flow in the intersection. The genetic algorithm intends to introduce a global optimization criterion to each of the local learning processes that optimize the cycle of traffic signals and green-ratio. Area-wide coordination is achieved by game theory. We combine local optimization with global optimization to optimize traffic signal in multi-intersection. Simulation results indicate that our presented method is superior than traditional control one.