Skip to content
This repository has been archived by the owner on May 14, 2024. It is now read-only.

CartPole Reinforcement Learning with Neuroevolution of augmenting topologies

License

Notifications You must be signed in to change notification settings

AnthonyDickson/CartPole-NEAT

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Build Status codebeat badge Maintainability Codacy Badge codecov

CartPole-NEAT

In this project I aim to implement NEAT in python. Then I will try solve the CartPole problem using it. I will use OpenAI Gym's environment for this. I also plan to make this program multi-threaded so that I can have multiple runs (with possibly different configurations) running in parallel.

I also want to make a supporting web application that will handle real-time plotting as I have found matplotlib to be a bit lacking in that area. This as an opportunity to practice some web development skills and learn some new tech. This web application would be a RESTful application served locally from a flask server running in docker container, backed by MongoDB or possibly PostgreSQL. For the front end of things I will probably use a combination of Twitter Bootstrap, React, and JavaScript Live.

Check out the wiki pages for relevant resources and the projects page to see the progress of brainstorming, planning, and implementation. This project is split up into three subprojects: the actual genetic algorithm implementation, the RESTful API that provides access to test data, and the dashboard that provides visualizations of previous runs and a real time visualization test runs that are in progress. Each subproject contains a README.md with further details.

Test Suite

I intend on building up a test suite as I go. This can be run with the 'test_suite.sh' script, although you probably don't need to do this yourself since Travis CI does this for you.