Locally Interdependent Multi-Agent MDP: Theoretical Framework for Decentralized Agents with Dynamic Dependencies
This provides a theoretical framework for decentralized agents in domains like cooperative navigation, addressing a gap in the literature, but it appears incremental as it builds on existing MDP concepts.
The paper tackles the lack of theoretical analysis for decentralized multi-agent systems with dynamic dependencies by proposing the Locally Interdependent Multi-Agent MDP model, and it shows that near-optimal closed-form policies can be derived, with the decentralized solution being exponentially close to the fully observable one based on visibility radius.
Many multi-agent systems in practice are decentralized and have dynamically varying dependencies. There has been a lack of attempts in the literature to analyze these systems theoretically. In this paper, we propose and theoretically analyze a decentralized model with dynamically varying dependencies called the Locally Interdependent Multi-Agent MDP. This model can represent problems in many disparate domains such as cooperative navigation, obstacle avoidance, and formation control. Despite the intractability that general partially observable multi-agent systems suffer from, we propose three closed-form policies that are theoretically near-optimal in this setting and can be scalable to compute and store. Consequentially, we reveal a fundamental property of Locally Interdependent Multi-Agent MDP's that the partially observable decentralized solution is exponentially close to the fully observable solution with respect to the visibility radius. We then discuss extensions of our closed-form policies to further improve tractability. We conclude by providing simulations to investigate some long horizon behaviors of our closed-form policies.