This repository contains the code to train the baseline agent provided in the 2022 edition of Learning to Run a Power Network and to recreate the experiments (as well as the figures) of the paper Reinforcement learning for Energies of the future and carbon neutrality: a Challenge Design.
This agent uses the Proximal Policy Optimization algorithm.
You need python>=3.8.6
To train the baseline:
pip[3] install -r requirements.txt
python[3] train.py [args]Run python[3] train.py -h to see all available arguments.
To reproduce the figures (6), (7) and (8) of
Reinforcement learning for Energies of the future and carbon neutrality: a Challenge Design,
check the corresponding make_figure_n.ipynb notebook.
You need to train several instances of the baseline agent before running these notebooks.
We used this code to make the experiments of Reinforcement learning for Energies of the future and carbon neutrality: a Challenge Design but we are still working on cleaning it up.