This repository is an introduction to reinforcement learning (RL) and provides code for the learning algorithms and agents outlined in:
- Reinforcement Learning: An Introduction
- David Silver's Reinforcement Learning Course
- CS294 - Deep Reinforcement Learning (Berkeley)
- CS8803 - Reinforcement Learning (Georgia Tech)
The repository is subsetted into different folders, each one corresponding to an area of RL and will contain an overview and agents used to solve particular problems in that area. The code is in a mix of Python 2 and 3, depending on the environment. The environments derive from OpenAI Gym, Berkeley and some I have coded up myself. The repository is an ongoing project and new folders will be continually added.
Agents and learning algorithms have been coded up from first principles, usually based in Numpy, but more advanced techniques require Tensorflow and Keras.