User guide

This page contains what you need to know to install the package and get started with its basic functionalities, as well as information and explanations about the different components of the library and how to use these to implement and deploy MPC-based RL agents.

Getting started

• Installation
  • Using pip
  • Using source code
• A simple Reinforcement Learning task for an LTI system
  • Environment
  • Controller
  • Learning

Theory

• Reinforcement Learning
• Model Predictive Control
  • Parametric MPC
• Reinforcement Learning using Model Predictive Control
  • MPC as function approximation
  • Q-learning with MPC
  • Deterministic Policy Gradient with MPC
• Gradient-free approaches

Agents

• Inheritance hierarchy
  • Callbacks
  • Agents
• Base agent classes
  • Agent
  • LearningAgent
  • RlLearningAgent (gradient-based)
  • GlobOptLearningAgent (gradient-free)
• Concrete agent classes

Callbacks

TODO

Learnable parameters

TODO

Optimizers

TODO

Exploration

TODO

Updating

TODO

Experience

TODO

Warmstarting

TODO