/demo-RL

Primary LanguagePython

Reinforcement Learning DDPG

Test of DDPG algorithm on a simplified satellite simulation.


Project Structure

train.py -- Main loop between Environment and Agent
train_utils.py -- Helper functions such as noise function
agent.py -- Define how the agent makes actions given observations and updates policy
networks.py -- Define actor and critic networks
prioritized_buffers -- Implementation of PER
satellite_engine -- Simplified satellite simulation
satellite_constant -- Constants for calculating satellite simulation