BETSE Gymnasium Environment and Koopman Operator Learning

Project Overview

This project integrates the BETSE simulation engine with Gymnasium to enable reinforcement learning and Koopman operator learning for controlling bioelectric gradients in a 7-cell system.

File Structure

• gym_betse/: Contains the Gymnasium environment, agents, and utilities.
• koopman/: Contains the Koopman operator learning framework using PyTorch.
• train.py: Script for training RL agents.
• train_koopman.py: Script for training the Koopman operator model.

Getting Started

1. Install the required packages:

   pip install -e gym_betse/
   

2. Run the training script for RL agents:

   python gym_betse/train.py
   

3. Run the training script for the Koopman operator model:

   python koopman/train_koopman.py
   

Requirements

• Python 3.7+
• PyTorch
• Gymnasium
• BETSE
• h5py
• numpy

Authors (in alphabetical order)

• Akriti Adhikari
• Matthew Bailey
• Brian Brown
• Jonah Checkets
• Charlie Clark
• Caleb Crandal
• Andrew Criddle
• Jacob DeGraw

License

This project is licensed under the MIT License.