Study of bounded confidence models of opinion dynamics on networks

Abstract

Using the Deffuant model of opinion dynamics, in which the agents in a network can only influence each other if the discrepancy in their opinions is less than the confidence bound parameter, we observe the formation of clusters during the dynamic. In this process there is a fraction of agents that stay isolated and don’t change their opinion or they form really small clusters that barely change their opinion from the start and stay in the extreme opinions. This work is centered on studying this fraction of isolated agents, how the probability changes for different topologies, dynamics and stages.