Reinforcement Learning environments and agents/policies used for the FNAL accelerator application

Software Requirement

• Python 3.7
• The environemnt framework is built of OpenAI Gym
• Additional python packages are defined in the setup.py
• For now, we assumes you are running at the top directory

Installing

• Pull code from repo

git clone https://github.com/fermilab-accelerator-ai/control-for-accelerators-in-hep.git

• Install control-for-accelerators-in-hep (via pip):

cd control-for-accelerators-in-hep
pip install -e . --user

Directory Organization

├── setup.py
├── scripts                           : a folder contains RL steering scripts  
├── dataprep                          : a folder with code to read and prep data
├── surrogates                        : a folder contains surrogate model code
├── agents                            : a folder contains agent codes
├── gym_accelerator                   : a folder containing the accelerator environments
├── cfg                               : a folder contains the agent and environment configuration
├── utils                             : a folder contains utilities