Reinforcement learning for FrozenLake

Reinforcement Learning to solve the FrozenLake environment from Gym.
The following modules have been implemented:

1. Value iteration
2. Policy iteration
3. Tabular Q learning

Installation

git clone https://github.com/rnitin/frozen_lake.git
pip install -r requirements.txt

Description

VI.ipynb

Jupyter notebook to perform value iteration on 4x4 Frozen lake environment.

PI.ipynb

Jupyter notebook to perform policy iteration on 4x4 Frozen lake environment.

QL.ipynb

Jupyter notebook to perform tabular Q learning on 4x4 Frozen lake environment.

Acknowledgements

This program was developed as a part of the Reinforcement Learning course offered by Prof. Srinivas Shakkottai at Texas A&M University.
The notebooks use the fancy_visual method which was provided by the instructors.