We’ve just released Open Control - a python package that implements basic algorithms for analysis and design of optimal feedback controllers.

Features:

• Classical control methods

• Linear quadratic regulator (LQR) computation

• Adaptive Dynamic Programming (ADP) for optimal linear/nonlinear control systems

• Off-policy, On-policy learning algorithms for linear/nonlinear systems

• Experience replay algorithms for linear/nonlinear systems (TODO)

• Learn Reinforcement Learning by the tutorial notebook and examples code.

OpenControl provides algorithms for wide range of system dynamics, including classical control module and reinforcement learning control systems with ADP module. For learning purpose, we recommend starting with our example code to define a system and run the simulation.

Development

You can check out the latest version of our source code by cloning the repository:

git clone https://github.com/VNOpenAI/OpenControl`

Installation

Before all, create a new virtual conda environment and install requirements

pip install requirements.txt

You can do minimal installation of OpenControl with

pip install OpenControl

Check version

python
>>> import OpenControl 
>>> OpenControl.__version__

Documentation

Please read the lastest version of our documentation here

How to contact OpenControl team?

For any question, drop an email at phi9b2@gmail.com.