Reinforcement-learning (In-progress)

I'm trying to implement some reinforcement-learning algorithms. Most of my implementation based on three lectures below:

• Richard Sutton's & Andrew Barto's Book Reinforcement Learning: An Introduction (2nd Ed)
• David Silver's Lecture
• Udacity's Reinforcement-learning lecture (Georgia Tech).

My codes are like a rewrite from Denny Britz's Repo, But because I can't write such a beautiful code like he does yet :( So I try to implement many of it by myself ;)

Table of contents

0. Environment

• Grid World (Environment, DP-Policy Evaluation, DP-Policy Iteration, DP-Value Iteration)
• Black Jack (Environment)
• Windy Grid World (Environment)

1. Dynamic Programming

• Grid World (Environment, DP-Policy Evaluation, DP-Policy Iteration, DP-Value Iteration)

2. Simple Model-Free

• Black Jack Monte-Carlo (Prediction, Control, Simulation)
• Black Jack Temporal-Difference (TD[0]) (Prediction, Control a.k.a SARSA, Simulation)
• Windy Grid World Temporal-Difference (TD[0]) (Control a.k.a SARSA, Simulation)

3. Eligibility Traces

• Black Jack TD-lambda (Prediction