Figure 7
From: A scalable approach to optimize traffic signal control with federated reinforcement learning
The performance of the proposed method in new road network: (a) fine-tuning reward (b) halting number of vehicles; (c) average cumulative waiting time of all intersections.
