Skip to content

katjawittfoth/Frozen_Lake

Repository files navigation

Frozen Lake

Training a Reinforcement Learning agent to solve Frozen Lake game from OpenAI gym.

This is my project for the Reinforcement Learning class taken as an elective for the Master's in Data Science program at the University of San Francisco.

The goal is to train an agent to get to the goal from the start without falling into a hole on a frozen lake. The lake is extremely slippery, so you don't always go in the direction you indented to, which adds to the complexity of the problem.

I used a Q-learning as well as DQN reinforcement learning algorithms to train my agent.

Project Presentation

Q-learning agent

DQN agent

Frozen_Lake GIF

About

Reinforcement Learning agent to solve Frozen Lake game from OpenAI gym

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages