![](https://img.shields.io/badge/Chat%20on%20Slack-%23reinforcement--learnin-ff69b4"){kind=icon}
ReinforcementLearning.jl, as the name says, is a package for reinforcement learning research in Julia.
Our design principles are:
- Reusability and extensibility: Provide elaborately designed components and interfaces to help users implement new algorithms.
- Easy experimentation: Make it easy for new users to run benchmark experiments, compare different algorithms, evaluate and diagnose agents.
- Reproducibility: Facilitate reproducibility from traditional tabular methods to modern deep reinforcement learning algorithms.
julia> ] add ReinforcementLearning
julia> using ReinforcementLearning
julia> run(E`JuliaRL_BasicDQN_CartPole`)Check out the Get Started page for a detailed explanation. The underlying design decisions and implementation details are documented in this blog.
ReinforcementLearning.jl itself is just a wrapper around several other packages inside the JuliaReinforcementLearning org. The relationship between different packages is described below:
+-----------------------------------------------------------------------------------+
| |
| ReinforcementLearning.jl |
| |
| +------------------------------+ |
| | ReinforcementLearningBase.jl | |
| +----|-------------------------+ |
| | |
| | +--------------------------------------+ |
| +---->+ ReinforcementLearningEnvironments.jl | |
| | +--------------------------------------+ |
| | |
| | +------------------------------+ |
| +---->+ ReinforcementLearningCore.jl | |
| +----|-------------------------+ |
| | |
| | +-----------------------------+ |
| +---->+ ReinforcementLearningZoo.jl | |
| +----|------------------------+ |
| | |
| | +-------------------------------------+ |
| +---->+ DistributedReinforcementLearning.jl | |
| +-------------------------------------+ |
| |
+-----------------------------------------------------------------------------------+
- ReinforcementLearningBase.jl Two main concepts in reinforcement learning are precisely defined here: Policy and Environment.
- ReinforcementLearningEnvironments.jl Typical environment examples in pure Julia and wrappers for 3-rd party environments are provided in this package.
- ReinforcementLearningCore.jl Common utility functions and different layers of abstractions are contained in this package.
- ReinforcementLearningZoo.jl
Common reinforcement learning algorithms and their typical applications (aka
Experiments) are collected in this package. - DistributedReinforcementLearning.jl
This package is still experimental and is not included in
ReinforcementLearning.jlyet. Its goal is to extend some algorithms inReinforcementLearningZoo.jlto apply them in distributed computing systems.
ReinforcementLearning.jl is a MIT licensed open source project with its
ongoing development made possible by many contributors in their spare time.
However, modern reinforcement learning research requires huge computing
resource, which is unaffordable for individual contributors. So if you or your
organization could provide the computing resource in some degree and would like
to cooperate in some way, please contact us!
If you use ReinforcementLearning.jl in a scientific publication, we would
appreciate references to the following BibTex entry:
@misc{Tian2020Reinforcement,
author = {Jun Tian and other contributors},
title = {ReinforcementLearning.jl: A Reinforcement Learning Package for the Julia Language},
year = 2020,
url = {https://github.com/JuliaReinforcementLearning/ReinforcementLearning.jl}
}
Thanks goes to these wonderful people (emoji key):
This project follows the all-contributors specification. Contributions of any kind welcome!