Opinion Dynamics of Social Networks with Stubborn Agents via Group Gossiping with Random Participants

Abstract

Several recent models of opinion dynamics utilize gossip-based methods as an alternative to deterministic classical models. This approach is meant to be a more realistic representation of real-world communications by using random pairwise interactions. Our previous work extended the process of gossip-based models by enabling agents to communicate with a random subset of their neighbors. In this paper, we apply this idea to networks with stubborn agents. While the opinions in this model tend to oscillate, its expected dynamics is convergent, and the expected opinions and time-averaged opinions coincide over time.