Evolution of social power over influence networks containing antagonistic interactions (Q2053816)

This paper studies opinion dynamics with social power in networks containing antagonistic connections. The opinion formation process involves a sequence of issues to be discussed among the individuals. A reflected appraisal mechanism is introduced in the opinion interactions in the network. A weight is given to each agent by its neighbors based on the appraisals, namely the weights are proportional to the impression of its neighbors. The group influence is given in the framework of social power in social networks with signed edges. The paper studies the convergence property and stability of the proposed model. Differential Lyapunov theory is used to investigate the stability of the system. It shows that the group members forget the initial perception of social influence at an exponential rate. The long term social powers are influenced fully by the appraisal network. Some simulation examples are provided to illustrate the theory.