Reinforcement Learning
Implementation of various RL agents
This project provides a collection of reinforcement learning agents and offers a unified framework to train, evaluate and visualize the performance of the agents on various Gym environments.
The following agents have been implemented:
- Tabular learning
- Q-Learning
QLearningAgent - SARSA Learning
SARSAAgent
- Q-Learning
- Deep-Learning based function approximation
- Dueling Deep Q-Network
DDQNAgent - REINFORCE algorithm
ReinforceMCwithBaselineAgent, ReinforceMCwithoutBaselineAgent
- Dueling Deep Q-Network
- Hierarchical Learning
- Semi-Markov Decision Process Q-Learning
SMDPQLearningAgent - Intra-Option Q-Learning
IntraOptionQLearningAgent
- Semi-Markov Decision Process Q-Learning
More information about this project can be found at this repository