Naming Game on small-world networks: the role of clustering structure

TL;DR

This paper studies how the clustering structure of small-world networks influences the dynamics of the Naming Game, revealing that higher clustering slows convergence and affects agents' memory.

Contribution

It demonstrates the significant impact of clustering coefficient on Naming Game dynamics and establishes a quantitative relationship between them.

Findings

Higher clustering coefficient slows convergence.

Clustering affects maximum memory of agents.

A quantitative relationship between clustering and memory is identified.

Abstract

Naming Game is a recently proposed model for describing how a multi-agent system can converge towards a consensus state in a self-organized way. In this paper, we investigate this model on the so-called homogeneous small-world networks and focus on the influence of the triangular topology on the dynamics. Of all the topological quantities, the clustering coefficient is found to play a significant role in the dynamics of the Naming Game. On the one hand, it affects the maximum memory of each agent; on the other hand, it inhibits the growing of clusters in which agents share a common word, i.e., a larger clustering coefficient will cause a slower convergence of the system. We also find a quantitative relationship between clustering coefficient and the maximum memory.