FINITE-TIME CLUSTER CONSENSUS FOR NETWORKS VIA DISTRIBUTED ITERATIONS
作者单位：Department of Mathematics Tongji University
会议名称：《第十届动力学与控制学术会议》
会议日期：2016年
学科分类：07[理学] 070104[理学-应用数学] 0701[理学-数学]
基 金：国家自然科学基金青年基金(11505127) 上海市浦江人才计划(15PJ1408300) 中央高校基本科研业务费专项资金(0800219319)
关 键 词：complex networks cluster consensus multi-agent systems distributed algorithm finite-time
摘 要：This paper studies the finite-time cluster average consensus in bidirectional networks. By designing distributed linear iterations involving stochastic matrices consistent with a network topology, we show that cluster average consensus can always be achieved in finite time for bidirectional networks containing cluster-spanning trees. In particular, for a network with n nodes, the cluster average consensus can be achieved in less than n(n-1)/2 steps regardless of the number of different clusters. We develop a detailed algorithm to implement the proposed finite-time cluster average consensus. Numerical simulations on a realistic social network are presented to illustrate the availability of the obtained theoretical results. Moreover, we compare the current approach with the past asymptotical consensus protocols, including those involving external intra-cluster identical inputs. The advantage of finite-time consensus is prominent.