Gearing Up for FIFA 2022 Using RLlib-Powered Traffic Control

Learn about innovative traffic control solutions developed for the FIFA World Cup 2022 in Qatar in this 33-minute conference talk by Anyscale. Discover how researchers at Qatar Computing Research Institute (QCRI) are addressing the unprecedented challenge of managing mobility for thousands of attendees within a 20-mile radius. Explore three key components of their approach: a parallel, congestion-optimized traffic microsimulator for analyzing various scenarios; a Graph Convolutional Networks-based prediction model for traffic in-flow estimation; and a multi-agent reinforcement learning framework using RLlib for coordinated traffic light control. Gain insights into the integration of RLlib to accelerate multi-agent learning and witness a demonstration of its application in optimizing traffic flow for this major sporting event.