We analyse how the interplay between several sources of heterogeneity in agent’s bias, namely plurality and polarization, shapes the emergence of different regimes of collective opinion in multi-agent systems. We do so by considering a kinetic-like model of opinion dynamics with restrictive pairwise interactions — which are modeled by a smooth bounded confidence — in a networked population endued with quenched plurality and polarization. The former is understood as gradation around some leading opinion whereas the latter means formal disagreement. Our results are able to reproduce the so-called ‘social hysteresis’ phenomenon, i.e., the outcome of a given collective decision process depends on the history of that society, as well as heterogeneity-assisted ordering, that is to say, the achievement of an agreement assumed in the form of a collective stance. In addition, we show that those results are qualitatively independent of the type of network architecture considered, specifically random, small-world or scale-free.

我们分析了代理人偏见的几种异质性源之间的相互作用，即多元性和两极分化，如何影响多代理人系统中不同集体意见制度的出现。为此，我们考虑了在具有猝灭的多元和两极分化的网络人口中，具有受限的成对交互作用的类似动力学的意见动态模型，该模型具有平滑的有界置信度建模。前者被理解为围绕一些主要观点的等级，而后者则意味着形式上的分歧。我们的结果能够重现所谓的“社会滞后”现象，即，给定的集体决策过程的结果取决于那个社会的历史以及异质性辅助的排序，也就是说，以集体立场的形式达成的协议的达成。此外，我们证明了这些结果在质量上与所考虑的网络体系结构类型无关，尤其是随机的，小规模的或无规模的。