Topology Switching for Optimal Leader-Following Consensus of Multi-Agent Systems

Abstract

This brief studies the optimal topology switching problem of multi-agent systems (MASs), analyzing the optimization of the quadratic cost function through the active switching mechanism of topologies when the linear consensus protocol is applied. The leader is designed as a one-dimensional constant, while accordingly, the followers model is described as a single integrator or a first-order difference equation. For three scenarios, namely, continuous and discrete-time MAS under connected topologies and continuous-time MAS under jointly connected topologies, switching conditions based on the system error are respectively given by the classical optimal control theory and dynamic programming. Numerical simulation verifies the system performance improvement by the switching algorithm.