PowerGridworld: A Framework for Multi-Agent Reinforcement Learning in Power Systems
Authors: David Biagioni, Xiangyu Zhang, Dylan Wald, Deepthi Vaidhynathan, Rhoit Chintala, Jennifer King, Ahmed S. Zamzam
Corresponding author: David Biagioni
All authors are with the National Renewable Energy Laboratory (NREL).
Description
PowerGridworld provides users with a lightweight, modular, and customizable framework for creating power-systems-focused, multi-agent Gym environments that readily integrate with existing training frameworks for reinforcement learning (RL). Although many frameworks exist for training multi-agent RL (MARL) policies, none can rapidly prototype and develop the environments themselves, especially in the context of heterogeneous (composite, multidevice) power systems where power flow solutions are required to define grid-level variables and costs. PowerGridworld is an opensource software package that helps to fill this gap. To highlight PowerGridworld’s key features, we include two case studies and demonstrate learning MARL policies using both OpenAI’s multi-agent deep deterministic policy gradient (MADDPG) and RLLib’s proximal policy optimization (PPO) algorithms. In both cases, at least some subset of agents incorporates elements of the power flow solution at each time step as part of their reward (negative cost) structures.
Please refer to our preprint on arXiv for more details. Data and run scripts used to generate figures in the preprint are available in the paper directory.
Basic installation instructions
Env setup:
conda create -n gridworld python=3.8 -y
conda activate gridworld
git clone [email protected]:NREL/PowerGridworld.git
cd PowerGridWorld
pip install -e .
pip install -r requirements.txt
Run the pytests to sanity check:
pytest tests/
pytests --nbmake examples/envs
Examples
Examples of running various environments and MARL training algorithms can be found in examples.
Funding Acknowledgement
This work was authored by the National Renewable Energy Laboratory (NREL), operated by Alliance for Sustainable Energy, LLC, for the U.S. Department of Energy (DOE) under Contract No. DE-AC36-08GO28308. This work was supported by the Laboratory Directed Research and Development (LDRD) Program at NREL.
Citation
If citing this work, please use the following:
@article{biagioni2021powergridworld,
title={PowerGridworld: A Framework for Multi-Agent Reinforcement Learning in Power Systems},
author={Biagioni, David and Zhang, Xiangyu and Wald, Dylan and Vaidhynathan, Deepthi and Chintala, Rohit and King, Jennifer and Zamzam, Ahmed S},
journal={arXiv preprint arXiv:2111.05969},
year={2021}
}
1 Dec 17, 2021
56 Nov 15, 2022
8 Sep 30, 2022
1.1k Jan 2, 2023
669 Dec 29, 2022
1 May 15, 2022
0 Jul 26, 2022
183 Dec 27, 2022
8 Nov 7, 2022
7 Mar 28, 2022
51 Dec 14, 2022
17 Jul 20, 2022
139 Jan 01, 2023
191 Dec 16, 2022
9 Oct 31, 2022
70 Nov 13, 2022
1 Aug 10, 2022
29 Oct 01, 2022
65 Nov 28, 2022
108 Dec 09, 2022
1.5k Jan 08, 2023
6.7k Dec 31, 2022
17 Jun 17, 2022
9.3k Jan 07, 2023
238 Jan 04, 2023
6 Nov 04, 2022
104 Nov 12, 2022
51 Nov 23, 2022