Analysis of Opinion Dynamics in Social Networks Subject to Time-Varying Topologies

Abstract

For social networks, the interactions among agents and the relative self-confidence of each agent compared with the effects of its neighbors generally change over time. This requires using time-varying signed digraphs to describe the opinion forming processes of agents, where the positive and negative edges can represent cooperations and antagonisms, respectively. In this paper, an improved opinion dynamics model instead of the conventional Laplacian-type model is exploited with allowance of the potential variation of relative self-confidence of each agent, which can be reflected by the diagonal dominance degree. It is shown that both the structural characteristics of social networks and the diagonal dominance degrees determine the opinion forming performances, and some sufficient conditions related to these two factors are proposed to establish the bipartite consensus and stability results of agents. Two simulation examples are provided to illustrate the obtained opinion forming behaviors.