On the existence of common Lyapunov functions for consensus algorithms based on averaging

This paper addresses distributed deterministic consensus algorithms based on averaging. We relate the conditions for achieving consensus to the existence of a common norm for a set of row-stochastic matrices associated with the original set of averaging matrices. For a system to achieve consensus, it is shown that this associated set of matrices should have a Common Lyapunov Function, even if the original set might not have one.