Diversity and Disorder in the Voter Model with Delays

TL;DR

This paper introduces a modification to opinion models that promotes diversity by preventing rapid state changes, leading to a globally disordered state in the voter model, with implications for understanding opinion dynamics.

Contribution

It proposes a generic relaxation period after opinion changes that sustains diversity without freezing the system, applicable to various agent-based models.

Findings

In the thermodynamic limit, the system always reaches a disordered state.

Finite size effects can cause transitions between coexistence and consensus.

The model was tested for multiple opinions and supported by mean field calculations.

Abstract

In this work, we investigate interactions that simultaneously order a system locally, while keeping it globally disordered. The study is done in the context of the emergence of diversity in opinion propagation models with interactions rooted in conformity, but some suggestions on how this could be extended to other topics (like ecology and neuroscience) are also made. We do this by introducing a generic modification that can be added to different opinion propagation models (and other agent based models) and that seems to introduce a global tendency towards diversity without leading the system to a frozen state, even in the absence of thermal noises, contrarian agents or cyclic interactions. This modification consists of effectively introducing a relaxation period right after an agent changes its state, during which it cannot change its state again. We tested this modification for the voter model in a square lattice and verified that in the thermodynamic limit, the only attractor is the completely disordered state, where all opinions coexist in the same proportion. For fixed lattice sizes, finite size effects cause a transition with the lenght of the relaxation period between coexistence and consensus. We made simulations for 2, 3, 4, 5 and 6 opinions in a square lattice and mean field calculations for an arbitrary number of opinions.