Reinforcement Learning Algorithms Implementation

---

Reinforcement Learning Examples

In each file, there is a jupyter notebook and its corresponding python file. Where in the note book records all the intermediate process and plots, in the python file is the compiled process you can run directly.

Grid World

• Determnistic Grid World with Value Iteration

• Non-Deterministic Grid World with Q-Learning

Tic Tac Toe

• Tic-Tac-Toe Python Implementation

Muilti-Arm Bandit

• Muilti-Arm Bandit Implementation

Blackjack

• Monte Carlo Simulation on Blackjack

• Blackjack Solution

Cliff Walking

• Cliff Walking Implementation

Random Walk(n-step TD method)

• Random Walk Implementation

DynaMaze

• DynaMaze Implementation
• DynaMaze Extensions Implementation

Random Walk(General)

• 1000-State Random Walk

Tile Coding

• Tile Coding Implementation

Mountain Car

• On-policy Function Approximation

Server Access

• Generalisation on Continuing Tasks

Baird Counter Example

• Generalisation of Off-Policy Learning

TD-Lambda

• Random Walk offline-Lambda
• Random Walk TD(\lambda)

Mountain Car-Lambda

• Sarsa-Lambda

Policy Approximation

• Short-Corridor