A graph theoretic approach to ensure a scalable performance measure in Multi-Agent Systems with MIMO agents

Abstract

In this paper we consider issues of scalable performance measures in multi-agent systems, that is, performance measures that are independent of the size of the network. We include recent results for finding such performance measures and algorithms for a distributed design process, such that networks are able to reach a desired performance measure given constraints. Further, we can show a graph theoretic approach for guaranteeing a bound on the state deviations stemming from external disturbances. We can also show that this may be verified using local information only, thus yielding a fully scalable method. We exemplify this on a fictitious multi-terminal DC network.