期刊
IEEE TRANSACTIONS ON CYBERNETICS
卷 51, 期 1, 页码 64-76出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TCYB.2020.2972403
关键词
Controllability; Topology; Complexity theory; Graphics; Process control; Cybernetics; Systematics; Complete graphic characterization; destructive nodes; local interactions; multiagent systems; topology structure
类别
资金
- National Natural Science Foundation of China [61873136, 61903210, 61603288]
- Science Foundation of Shandong Province for Distinguished Young Scholars [JQ201419]
The study focuses on the main difficulty of establishing the graph-based criterion for the complete graphic characterization of multiagent controllability, and presents a systematic design process using controllability destructive nodes. The results are not only applicable to any leader's selection but also reveal the complexity in studying complete controllability graphic characterizations.
Establishing the graph-based criterion for the selection of any location and any number of leaders is the main difficulty in the complete graphic characterization of multiagent controllability. This greatly increases the complexity of the study, compared with the results derived for only one or several classes of leaders. Through a detailed analysis of graphs of six nodes, this article presents a systematic design and identification process for the complete graphic characterization by taking advantage of controllability destructive nodes. The topologies obtained by the proposed method allow directly determining controllability at the network topology level. The results are not only applicable to any leader's selection but also reveal the difficulty and complexity in the study of complete controllability graphic characterizations. Moreover, by comparing graphs composed of five and six nodes, the results reveal the graph-theory-based controllability complexity caused by adding only one node. Finally, results are derived to show how to design topology structures to ensure the controllability under any selection of leaders.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据