Probabilistic Switching and Convergence Rate in Consensus Problems

P. Hovareshti, J. S. Baras and V. Gupta

Proceedings of the Allerton Conference on Communication, Control and Computing, pp. 423-430, Chicago, Illinois, September 26-28, 2007

 

Abstract

We consider the convergence problem of consensus seeking agents and the effect of probabilistic switching on the convergence rate. Using a probabilistic setting, we develop a framework to study the convergence rate for different classes of graph topologies, including Small World networks. We show that by making the effective diameter of the underlying graph small, probabilistic switching can be helpful in the design of fast consensus algorithms.