This project is a school project for the 5th year Deep Reinforcement Learning subject at @ESGI. By doing this project we learned a lot about the Rust language, the tch-rs library and the functioning of the most famous deep reinforcement learning models.
Algorithms implemented are:
- DQN (Deep Q Learning)
- DDQN (Double Deep Q Learning)
- DDQN with Experience Replay
- DDQN with Prioritized Experience Replay
- ReINFORCE
- ReINFORCE with learned baseline
- MCRR (Monte Carlo Random Roll-out)
- MCTS (Monte Carlo Tree Search) 🚧
- PPO A2C (Proximal Policy Optimization with Actor-to-Critic) 🚧
Environments implemented are:
- Line World
- Grid World
- Pac Man (game is done, but didn't had time to execute algorithms in it) 🚧
Made by Cédric GARVENES and Réda MAIZATE