Pairwise Attraction-Repulsion on Multilayer Social Networks

TL;DR

This paper introduces a probabilistic attraction-repulsion model for opinion dynamics on multilayer social networks, demonstrating conditions for consensus, metastability, and polarization through theoretical analysis and numerical experiments.

Contribution

It generalizes classical opinion models to multilayer networks with heterogeneous influences and establishes rigorous conditions for consensus and other collective behaviors.

Findings

Almost sure global consensus under attractive interactions and sufficient connectivity

Existence of regimes with metastability and polarization

Numerical simulations confirm theoretical dynamical regimes

Abstract

We introduce a probabilistic pairwise \emph{attraction--repulsion} model for opinion dynamics on multilayer social networks, in which agents hold layer-specific states and interact through random matchings that couple multiple, time-varying layers. At each time step, interacting pairs update their layer-specific states using layer-dependent, time-varying interaction rates and a random sign (attractive or repulsive), and the resulting updates are averaged across layers. This framework generalizes classical gossip and Deffuant-type models while capturing heterogeneous cross-layer influences and antagonistic interactions. Under mild graph-theoretic and moment assumptions, we establish almost sure global consensus. Specifically, when the expected net effect of interactions is strictly attractive and random matchings ensure sufficient cross-layer connectivity, all agents' layer states converge almost surely to the global average. We further identify a purely attractive regime in which consensus holds even under intermittent connectivity and without any moment assumptions on the initial states. Numerical experiments illustrate the dynamical regimes predicted by the theory, including consensus, metastability, and polarization. Together, these results provide a rigorous foundation for understanding how multilayer structure, stochastic interactions, and mixed-sign influence shape collective outcomes in social and engineered networked systems.