The Krause-Hegselmann Consensus Model with Discrete Opinions

TL;DR

This paper extends the Krause-Hegselmann consensus model to discrete opinions, introduces a faster algorithm, and analyzes how network structure and number of opinions affect consensus formation.

Contribution

It presents a new discrete-opinion version of the Krause-Hegselmann model with improved computational efficiency and explores consensus dynamics on different network topologies.

Findings

Consensus is always reached for Q<=7 in complete societies.

The number of surviving opinions S is roughly constant for Q>7 and Q<N.

On Barabasi-Albert networks, the consensus threshold depends on node outdegree.

Abstract

The consensus model of Krause and Hegselmann can be naturally extended to the case in which opinions are integer instead of real numbers. Our algorithm is much faster than the original version and thus more suitable for applications. For the case of a society in which everybody can talk to everybody else, we find that the chance to reach consensus is much higher as compared to other models; if the number of possible opinions Q<=7, in fact, consensus is always reached, which might explain the stability of political coalitions with more than three or four parties. For Q>7 the number S of surviving opinions is approximately the same independently of the size N of the population, as long as Q<N. We considered as well the more realistic case of a society structured like a Barabasi-Albert network; here the consensus threshold depends on the outdegree of the nodes and we find a simple scaling law for S, as observed for the discretized Deffuant model.