Nonlinear AgentBased Dynamics: Which Metric Mitigates Polarization?

Abstract

This article provides a stochastic agentbased model to exhibit the role of aggregation metrics in order to mitigate polarization in a complex society. Our sociophysics model is based on interacting and nonlinear Brownian agents, which allow us to study the emergence of collective opinions. The opinion of an agent, xi(t) is a continuous positive value in an interval [0,1]. We find: (i) most agentmetrics display similar outcomes, (ii) The middlemetric and noisymetric obtain new opinion dynamics either toward assimilation or fragmentation, and (iii) We show that a developed twostage metric provide new insights about convergence and equilibria. In summary, our simulation demonstrates the power of institutions, which affect the emergence of collective behavior. Consequently, opinion formation in a decentralized complex society is reliant to the individual information processing and rules of collective behavior.