Opinion Formation with the Evolution of Network

We analyze users’ behavior in Tianya online forum, and find great fluctuations exist in the amount of agents that participate in interactions. Therefore we present a model that includes the coevolving dynamics of network and opinions. New nodes are added to network gradually, promoting the total size of the system. On the other side, old nodes which feel tired of the topic may withdraw interactions. The rule of continuous opinion exchanges is applied in the dynamics, and only active agents can update their opinions and build connections with new nodes. Simulation results show topological and individual dissipative features of our model are analogous to the Tianya network. Our network is a scale-free and small-world network. Unlike other opinion models, if the tolerance parameter is small, more disordering will be caused in our model with the time elapsed. For a large tolerance parameter, the number of clusters declines, and only one macroscopic-size cluster exists at last. We also realize a first-order phase transition where the second-largest cluster is incorporated into the largest one.