首页 | 本学科首页   官方微博 | 高级检索  
     检索      


Emergence and evolution of multiple clusters of attracting agents
Authors:Dirk Aeyels
Institution:SYSTeMS Research Group, Department of Electrical Energy, Systems and Automation, Ghent University, Technologiepark Zwijnaarde 914, 9052 Zwijnaarde, Belgium
Abstract:We investigate a multi-agent system with a behavior akin to the cluster formation in systems of coupled oscillators. The saturating attractive interactions between an infinite number of non-identical agents, characterized by a multimodal distribution of their natural velocities, lead to the emergence of clusters. We derive expressions that characterize the clusters, and calculate the asymptotic velocities of the agents and the critical value for the coupling strength under which no clustering can occur. The results are supported by mathematical analysis.For the particular case of a symmetric and unimodal distribution of the natural velocities, the relationship with the Kuramoto model of coupled oscillators is highlighted. While in the generic case the emergence of a cluster corresponds to a second-order phase transition, for a specific choice of the natural velocity distribution a first-order phase transition may occur, a phenomenon recently observed in the Kuramoto model. We also present an example for which the clustering behavior is quantitatively described in terms of the coupling strength.As an illustration of the potential of the model, we discuss how it applies to the dynamic process of opinion formation.
Keywords:Multi-agent systems  Phase transition  Self-organization  Kuramoto model  Clustering
本文献已被 ScienceDirect 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号