点击次数：

论文类型：期刊论文

发表时间：2017-09-01

发表刊物：AUTOMATICA

收录刊物：Scopus、SCIE、EI

卷号：83

页面范围：116-123

ISSN号：0005-1098

关键字：Delayed information; Optimal tradeoff; Convergence rate; Consensus algorithms

摘要：We consider a distributed consensus problem over a network, where at each time instant every node receives two pieces of information from disjoint neighboring sets: a weighted average of current states of neighbors from a primary network, and a weighted average of one-hop delayed states of neighbors from a secondary network. The proposed algorithm makes each node update its state to a weighted average of these individual averages. We show that convergence to consensus is guaranteed with non-trivial weights. We also present an explicit formula for the weights allocated to each piece of the information for the optimal rate of convergence, when the secondary network is the complement of the primary network. Finally numerical examples are given to explore the case when the neighbor sets of the agents do not cover the whole network. (C) 2017 Elsevier Ltd. All rights reserved.