cor3bit/pydeco

PyDeco - Experiments in Decentralized and Distributed Control Algorithms

Experiments in Decentralized and Distributed Control Algorithms

---

Usage

• Install the library with pip:

$ pip install git+https://github.com/cor3bit/pydeco.git

TODO

Supported Environments

• Vanilla LQ
• Centralized LQ (each agent has access to the full state)
• Distributed LQ (local information)
• Non-linear Vehicle Platoon Model [in progress]

Supported Controllers

• Single-agent Analytical LQR
• Single-agent GPI
  • RLS Policy Evaluation
  • Q-learning Policy Evaluation
  • Q-learning GN Policy Evaluation
• Multi-agent GPI
  • RLS Policy Evaluation
  • Q-learning Policy Evaluation
  • Q-learning GN Policy Evaluation

Acknowledgements

TODO

References

1. Steven J. Bradtke, B. Erik Ydstie, Andrew G. Barto, "Adaptive linear quadratic control using policy iteration" (1994)
2. Siavash Alemzadeh, Mehran Mesbahi, "Distributed Q-Learning for Dynamically Decoupled Systems" (2018)
3. Daniel Goerges, "Distributed Adaptive Linear Quadratic Control using Distributed Reinforcement Learning" (2019)
4. Hang Wang, Sen Lin, H. Jafarkhani, Junshan Zhang, "Distributed Q-Learning with State Tracking for Multi-agent Networked Control" (2020)