采用广义线性局部交互协议的二阶多智能体网络一致性  

Consensus of networked multi-agent systems with double-integrator dynamics using generalized linear local interaction protocols

作　　者：肖丽[1,2] 廖晓峰[1] 韦鹏程[2] 李华青[1] 

机构地区：[1]重庆大学计算机学院,重庆400044 [2]重庆第二师范学院数学与通信工程系,重庆400067

出　　处：《清华大学学报（自然科学版）》2012年第8期1061-1065,共5页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金资助项目(60703035);重庆市自然科学基金资助项目(2009BBB2227);重庆市教委资助项目(KJ091501;KJ091502;KJ121507)

摘　　要：多智能体系统的主要研究目的在于探索由个体之间的相互作用所产生的群体协调现象的内在机制和原理。该文设计出一种广义局部线性交互协议,对固定有向网络拓扑条件下的二阶多智能体网络系统的一致性行为进行了深入研究,得到了一个能保证多智能体网络达到一致的充分必要条件。并且,还确定了多智能体网络系统在达到一致时,最终的收敛状态由协议的线性增益、有向网络拓扑和所有智能体的初始值位移状态共同决定。最后,采用5个智能体组成的网络拓扑进行数值仿真,阐释了理论分析结果。The main purpose of research on the multi-agent systems is to explore the internal mechanisms and principles of the interaction between the individual agents. The convergence of networked multi-agent systems with double-integrator dynamics was investigated by designing generalized linear local interaction protocols, with a sufficient and necessary condition for the agents in the networks obtained to reach consensus. Theoretical results show that if the second-order consensus is reached, the convergence is commonly determined by the initial position states of agents, the linear gains in the protocols and the fixed directed network topology. A numerical example of five agents is given to illustrate the theoretical results.

关 键 词：协议 多智能体系统 一致性 拓扑 收敛 

分 类 号：TP273.5[自动化与计算机技术—检测技术与自动化装置]