文献类型：期刊文献

英文题名：Opinion separation in leader-follower coopetitive social networks

作者：Liang, Haili[1];Yuan, Fanli[1];Zhou, Zhao[2];Su, Housheng[3]

机构：[1]Shanghai Univ, Sch Mechatron Engn & Automat, Shanghai Key Lab Power Stn Automat Technol, Shanghai 200444, Peoples R China;[2]East China Univ Sci & Technol, Key Lab Adv Control & Optimizat Chem Proc, Minist Educ, Shanghai 200237, Peoples R China;[3]Huazhong Univ Sci & Technol, Sch Automat, Key Lab, Minist Educ Image Proc & Intelligent Control, Luoyu Rd 1037, Wuhan, Peoples R China

年份：2021

卷号：434

起止页码：90

外文期刊名：NEUROCOMPUTING

收录：;EI(收录号：20210509858391);WOS:【SCI-EXPANDED(收录号:WOS:000632639200008)】；

基金：This work is sponsored by the National Natural Science Foundation of China under Grant Nos. 61703261 and 61703164, and the National Key R&D Program of China (2018YFB1701103), and the Fundamental Research Funds for the Central Universities, Junior Fellowship of CAST Advanced S&T Thinktank Program DXBZKQN-2017-028.

语种：英文

外文关键词：Opinion dynamics; Multi-agent systems; Leader?follower control; Coopetitive networks

摘要：Opinion dynamics on social networks describe the time evolution of the opinions within the leader-follower coopetitive (cooperative-competitive) interactions. When does opinion formation result in opinion separation with the single leader? This paper investigates a continuous-time model of such opinion separation problems with signed networks. Our model combines the positive/negative weighting rule of Degroot to describe the impact of their neighbors and the reflected leader-follower mechanism of Jadbabaie to describe trust/distrust relationships from followers to the leader. When the leader is stationary, we establish a necessary and sufficient condition, which ensure all opinions finally converge to bipartite consensus as long as the network topology is structurally balanced. When the leader is dynamic, a sufficient condition is proposed to ensure the opinions of followers move with the opinion of leader or the opposite value in the same velocity. Subsequently, we introduce Laplacian schemes, the framework of signed networks and matrix analysis to develop the stability of networked systems. Numerical simulation examples are given to illustrate the performance of the proposed leader-follower coopetitive systems. (c) 2021 Elsevier B.V. All rights reserved.