Dynamic Properties of the Opinion Formation Model in a Multiagent System with a Time-Varying Structure of Links under Information Pressure Conditions

A model of opinion dynamics in a multiagent system is proposed that represents a change in agent’s opinion as a result of two processes. The first is an avalanche-like distribution of information between agents through the links existing in the system, during which the agents participating in the di...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Journal of experimental and theoretical physics 2021-11, Vol.133 (5), p.612-625
Hauptverfasser: Savitskaya, N. E., Fedorova, T. A.
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:A model of opinion dynamics in a multiagent system is proposed that represents a change in agent’s opinion as a result of two processes. The first is an avalanche-like distribution of information between agents through the links existing in the system, during which the agents participating in the distribution can copy the opinions of neighbors under the information pressure from them. The second is an independent change of opinion by agents who do not participate in the information avalanche but are under excessive information pressure. The probabilities of the agent’s participation in a particular process are determined by the main parameters of the model: the activity a of agents, which determines the probability for the agent to establish links with his neighbors, and the probability p of an independent change in the agent’s opinion under the pressure of the information obtained. It is shown that, depending on the relationship between these parameters, one of three dynamical regimes (phases) can be realized in the system. The first is a state close to dynamic parity that is characterized by a unimodal probability density function for the system-averaged opinion m with a maximum m 0 = 0. In the second regime, the system spends most of its time in states close to consensus. This regime is described by a bimodal probability density function for the system-averaged opinion with maxima m 1, 2 = ±1. The third dynamical phase is transient, all values of the system-averaged opinion in this phase are almost equiprobable, and the probability density function of this quantity is close to that for a uniform distribution. It is also shown that the proposed model makes it possible to study the opinion formation process in multiagent systems under conditions that could not be considered in previously developed models. As an example, the effect of the lifetime of the links established in the system on the implementation of various dynamical regimes in it is considered. It is shown that, as this time decreases, the range of parameter values at which the bimodal regime is realized narrows; i.e., it becomes difficult to achieve consensus in the system.
ISSN:1063-7761
1090-6509
DOI:10.1134/S1063776121100125