# Self-organising traffic lights

## Back to Basics: Deep Reinforcement Learning in Traffic Signal Control

We revisit basic assumptions in Deep RL for traffic signal control and define an MDP that outperforms current SotA.