This paper studies opinion formation in adaptive networks. The networks are adapted based on opinions of the nodes. If two neighboring nodes have dissimilar opinion values (i.e., the distance between their opinions is larger than a threshold), the link between them may be rewired to a link between nodes with similar opinions. As the process of rewiring proceeds, the nodes also update their opinions based on bounded confidence model, i.e., if the distance between the opinions of two neighbors is less than a certain threshold, they update their opinions based on a specific rule. As model networks, we consider preferential attachment scale-free and Erdos-Renyi pure random networks. Our extensive numerical simulations on model networks show that the final rewired networks (at the end of the simulation process) have different topological properties than the original scale-free or random networks. The rewired networks are more heterogeneous (with higher entropy of degree distribution) than the original networks. The rewired networks also exhibit increased assortativity and local efficiency and decreased global efficiency.
↧