Learning Decentralized Strategies for a Perimeter Defense Game with Graph Neural Networks
This work addresses decentralized coordination in multi-agent defense systems, offering a scalable solution that generalizes to large team sizes, though it is incremental as it builds on existing expert algorithms and GNN methods.
The paper tackled the problem of finding decentralized strategies for multi-agent perimeter defense games by designing a graph neural network-based learning framework to map local perceptions and communication to actions, achieving results that stay closer to an expert policy and capture more intruders than baseline algorithms.
We consider the problem of finding decentralized strategies for multi-agent perimeter defense games. In this work, we design a graph neural network-based learning framework to learn a mapping from defenders' local perceptions and the communication graph to defenders' actions such that the learned actions are close to that generated by a centralized expert algorithm. We demonstrate that our proposed networks stay closer to the expert policy and are superior to other baseline algorithms by capturing more intruders. Our GNN-based networks are trained at a small scale and can generalize to large scales. To validate our results, we run perimeter defense games in scenarios with different team sizes and initial configurations to evaluate the performance of the learned networks.