Graph Neural Networks for Decentralized Controllers
This addresses scalability and implementation issues in multi-agent systems like robotics and smart grids, offering a novel approach but likely incremental in combining GNNs with decentralized control.
The paper tackles the problem of controlling dynamical systems with autonomous agents by proposing a framework using graph neural networks (GNNs) to learn decentralized controllers from data, demonstrating its potential through the flocking problem with properties like equivariance and stability.
Dynamical systems comprised of autonomous agents arise in many relevant problems such as multi-agent robotics, smart grids, or smart cities. Controlling these systems is of paramount importance to guarantee a successful deployment. Optimal centralized controllers are readily available but face limitations in terms of scalability and practical implementation. Optimal decentralized controllers, on the other hand, are difficult to find. In this paper, we propose a framework using graph neural networks (GNNs) to learn decentralized controllers from data. While GNNs are naturally distributed architectures, making them perfectly suited for the task, we adapt them to handle delayed communications as well. Furthermore, they are equivariant and stable, leading to good scalability and transferability properties. The problem of flocking is explored to illustrate the potential of GNNs in learning decentralized controllers.