Multi-Agent Leader-Follower Tracking Over Random Structurally Balanced Networks

In the research of multi-agent systems (MASs), structurally balanced networks (SBNs) have been widely used to explain the cooperation and competition between agents. At present, most of the research focuses on the model over SBNs with deterministic information channels. The highlight of this paper is to examine leader-follower tracking of general linear MASs over random SBNs, in which the existence of information channel between agents at each time instant is probabilistic and independent of other channels. By describing the random interaction process by the products of sub-stochastic matrices, a sufficient condition is established even with a strictly unstable system matrix, and the upper bound of eigenvalue modules closely related to the stability of the system matrix is further given. The observed leader-follower tracking performance over SBNs is verified by simulation experiments.