Impact of topological characteristics on consensus building in multiagent systems

Abstract

This paper represents a network of nodes as a graph and investigates the impact of topological characteristics on consensus building. The properties evaluated are the algebraic connectivity ratio, average path, average cluster coefficient, average matching index, modularity, and the average participation. The consensus is measured by the average number of state updates that the system requires to get the global scope of the task at hand. Factor analysis and regression analysis are applied to analyze the impact of the topological characteristics, to understand the dependencies among them and to reduce the dimensionality of the analysis. A theoretical study is provided to support the results obtained by the statistical method. The insights from the analysis are discussed.