Skip to content

A TensorFlow based DQN agent who moves in a small grid world

License

Notifications You must be signed in to change notification settings

Lostefra/ReinforcementLearningToy

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ReinforcementLearningToy

A TensorFlow based agent who moves in a small grid world

This repository was created in preparation of my MSc thesis, in order to put in practice the study of Deep Reinforcement Learning.

The task of the agent is to go within a 4x4 grid world from the top left corner to the bottom right, in the minimum number of steps possible. The agent gets a reward when it gets to the bottom right corner and it gets penalised for each step it does and in case it falls from the 4x4 grid.

The agents currently implemented are: DQN, C51 DQN, REINFORCE.

Both the agent and the environment are implemented using the tf_agents library.

Releases

No releases published

Packages

No packages published

Languages