This is a code repository of the paper 'Learning Scalable Multi-Agent Coordination by Spatial Differentiation for Traffic Signal Control'
Maybe you need to create a specialised conda environment and then run following codes:
git clone https://github.com/Skylark0924/Gamma-Reward-Perfect.git
cd Gamma-Reward-Perfect
conda install environment.yaml
pip install -r requirements.txt
Note: The package of Cityflow may need to be installed manually. Please see their website for details.
python ray_gamma_reward.py
Arguments are listed in ray_gamma_reward.py, you can change them in the script or by using the terminal.