1 paper · 1 filter
Haoran Su, Hanxiao Deng
Reinforcement learning for traffic signal control is bottlenecked by simulators: training in SUMO takes hours, reproducing results often requires days of platform-specific setup, a…