reinforcement-learning-an-introduction

Why we included this project

Anyone working through Sutton and Barto's Reinforcement Learning: An Introduction will find this repo a practical companion. It reimplements nearly every figure and worked example from the second edition in clean, self-contained Python, from the 10-armed bandit testbed and gridworld value iteration through Monte Carlo, TD learning, Sarsa, Q-learning, Dyna-Q, and function approximation. Each script reproduces the exact plot shown in the book, so you can run a chapter's code, see the result, and check your own understanding against a reference implementation. Students who want to move from reading equations to actually running the algorithms will get the most out of it, and instructors can pull scripts straight into lab exercises. The code is organized by chapter and kept deliberately simple, a good starting point before moving on to heavier frameworks.