Examination of Average Consensus with Maximum-degree Weights and Metropolis-Hastings Algorithm in Regular Bipartite Graphs

Abstract

Expressing extensive raw data in a transparent ag-gregated form is a necessary process in many topical multi-agent systems. In this paper, we analyze the performance of the average consensus algorithm with the Maximum-degree weights and the Metropolis-Hastings algorithm in regular bipartite graphs, where both the algorithms for aggregating data diverge. We examine the evolution of their inner states and the mean square error in these critical graphs with various connectivity. Besides, these results are compared to the performance of the examined algorithms in non-regular non-bipartite topologies. The goal of our contribution presented in this paper is to identify whether the studied algorithms are also applicable in regular bipartite graphs despite their divergence.