Opinion spreading and agent segregation on evolving networks

We study a stochastic model where the distribution of opinions in a population of agents coevolves with their interaction network. Interaction between agents is enhanced or penalized according to whether they succeed at reaching an agreement or not. The system evolves towards a state where the network's structure and the opinion distribution is frozen, and the population is divided into disconnected communities. The structural properties of the population in the final state vary considerably with the control parameters. By means of numerical simulations, we give a detailed account of such properties, as well as of the final opinion distribution. We also provide approximate analytical results which explain some of the numerical results and clarify their origin. (c) 2006 Elsevier B.V. All rights reserved.