Multi-agent system consensus in packet-switched networks

In this paper, we consider the problem of consensus reaching in multi-agent systems exchanging information over digital networks. We first show how to model stochastic packet delays in such networks in a distributed delay framework. We discuss the validity of the approximation and we give conditions that justify this transformation. Subsequently, we present a control law for multi-agent systems (MAS) that guarantees consensus over networks with stochastic packet delays for both fixed and switching topologies. For fixed topologies, the underlying graph has to be connected and undirected. For switching topologies, we require in addition that the graphs be regular.