Dynamic behavior for social networks with state-dependent susceptibility and antagonistic interactions (Q2665146)

This paper studies a continuous-time generalized nonlinear opinion dynamics model with state-dependent susceptibility to persuasion as well as antagonistic interactions. The communication network is considered as a weighted signed graph, where edges carries positive or negative weights to indicate the potential cooperative or antagonistic relationship. Three types of state-dependent susceptibility are studied, including stubborn positives, stubborn neutrals and stubborn extremists. It is shown that the system matrix is diagonally dominant for the cases of stubborn positives and stubborn neutrals. In the case of stubborn extremists, two types of adjacency matrices are studied. They include the adjacency matrix that is eventually positive and the adjacency matrix that is not eventually positive. Using the Perron-Frobenius properties, sufficient conditions of state convergence are presented in these cases. Some numerical results have also been provided.