The Combination of Pairwise and Group Interactions Promotes Consensus in Opinion Dynamics

Abstract

In complex systems, agents often interact with others in two distinct types of interactions, pairwise interaction and group interaction. The Deffuant–Weisbuch model adopting pairwise interaction and the Hegselmann–Krause model adopting group interaction are the two most widely studied opinion dynamics. In this study, we propose a novel opinion dynamics by combining pairwise and group interactions for agents and study the effects of the combination on consensus in the population. In the model, we introduce a parameter $\alpha$ to control the weights of the two interactions in the dynamics. Through numerical simulations, we find that there exists an optimal $\alpha$ , which can lead to a highest probability of complete consensus and minimum critical bounded confidence for the formation of consensus. Furthermore, we show the effects of $\alpha$ on opinion formation by presenting the observations for opinion clusters. Moreover, we check the robustness of the results on different network structures and find the promotion of opinion consensus by $\alpha$ not limited to a complete graph.