Optimal Distributed Control of Multiagent Systems in Contested Environments via Reinforcement Learning
This topic contains 0 replies, has 1 voice, and was last updated by arXiv 1 year, 2 months ago.

Optimal Distributed Control of Multiagent Systems in Contested Environments via Reinforcement Learning
This paper presents a modelfree reinforcement learning (RL) based distributed control protocol for leaderfollower multiagent systems. Although RL has been successfully used to learn optimal control protocols for multiagent systems, the effects of adversarial inputs are ignored. It is shown in this paper, however, that their adverse effects can propagate across the network and impact the learning outcome of other intact agents. To alleviate this problem, a unified RLbased distributed control frameworks is developed for both homogeneous and heterogeneous multiagent systems to prevent corrupted sensory data from propagating across the network. To this end, only the leader communicates its actual sensory information and other agents estimate the leader state using a distributed observer and communicate this estimation to their neighbors to achieve consensus on the leader state. The observer cannot be physically affected by any adversarial input. To further improve resiliency, distributed Hinfinity control protocols are designed to attenuate the effect of the adversarial inputs on the compromised agent itself. An offpolicy RL algorithm is developed to learn the solutions of the game algebraic Riccati equations arising from solving the Hinfinity control problem. No knowledge of the agent dynamics is required and it is shown that the proposed RLbased Hinfinity control protocol is resilient against adversarial inputs.
Optimal Distributed Control of Multiagent Systems in Contested Environments via Reinforcement Learning
by Rohollah Moghadam, Hamidreza Modares
https://arxiv.org/pdf/1708.09630v2.pdf
You must be logged in to reply to this topic.