NodeRL Adventure

This is a collection of Reinforcement Learning Algorithms written entirely in TypeScript for educational purposes.

WORK IN PROGRESS!!

Plan

1. Multi-Arm Bandits Algorithms: Exploration/Exploitation
   • Epsilon-Greedy
   • Upper Confidence Bound
   • Thomson Sampling
2. Dynamic Programming
   • Iterative Policy Evaluation
   • Policy Improvement
   • Policy Iteration
   • Truncated Policy Iteration
   • Value Iteration
3. Monte Carlo Methods
   • MC Prediction
     • State Values
     • Action Values
   • MC Control
4. Temporal-Difference Methods
   • TD Prediction: TD(0)
   • TD Prediction
     • State Values
     • Action Values
   • TD Control: Sarsa
   • TD Control: Q-Learning
   • TD Control: Expected Sarsa
5. Value-Based Methods
   • Deep Q-Networks (DQN)
     • Vanilla DQN
     • N Step DQN
     • Double DQN
     • Dueling DQN
     • DQN with Prioritized Experience Replay (PER)
     • DQN with Noisy Networks
     • Categorical DQN (C51)
     • Quantile Regression DQN
     • Rainbow
     • Normalized Advantage Functions (NAF)
6. Policy-Based Methods
   • REINFORCE
   • Off-Policy
7. Actor-Critic Methods
   • Vanilla Actor-Critic
   • Advantage Actor Critic (A2C)
     • A2C with Generalized Advantage Estimation (GAE)
   • Trust Region Policy Optimization (TRPO)
   • Proximal Policy Optimization (PPO)
   • Actor-Critic with Experience Replay (ACER)
   • Actor-Critic using Kronecker-Factored Trust Region (ACKTR)
   • Deep Deterministic Policy Gradient (DDPG)
     • DDPG with Hindsight Experience Replay (HER)
   • Twin Delayed Deep Deterministic (TD3)
   • Soft Actor Critic (SAC)
     • SAC Discrete
8. Multi-Agent Algorithms
   • Multi-Agent DDPG (MADDPG)
   • Multi-Agent TD3
   • Multi-Agent SAC
9. Model-Based Algorithms
   • TODO