二阶积分型多自主体系统的均方线性双向趋同
发布时间：2018-12-13 02:27
【摘要】：在多自主体系统中,每个自主体之间不只是合作关系,还存在竞争关系.针对这类系统,本文研究了通讯噪声影响下多自主体系统的双向趋同问题.我们在双向趋同协议的设计中首次引入了时变趋同性增益.借助状态转移矩阵和代数图论等工具,分析了闭环系统解的性质,并分别针对连续时间和离散时间二阶积分型多自主体系统,给出了所设计的双向趋同协议为均方线性双向趋同协议的充要条件.研究表明,表征自主体间通讯关系的有向符号图结构平衡并且含有一棵生成树是保证协议为均方线性双向趋同协议的最弱拓扑条件.特别地,如果有向符号图还是加权平衡的,则所设计的控制协议为均方平均双向趋同协议,即每个自主体的位置均方收敛于一个模相同符号不同的期望为初值的加权平均的随机变量.
[Abstract]:In multi-agent systems, each self-agent is not only a cooperative relationship, but also a competition relationship. In this paper, the problem of bidirectional convergence of multi-agent systems under the influence of communication noise is studied. For the first time, we introduce time-varying convergence gain in the design of bidirectional convergence protocol. By means of state transition matrix and algebraic graph theory, the properties of the solution of the closed-loop system are analyzed. A necessary and sufficient condition for the proposed bidirectional convergence protocol to be a mean-square linear bidirectional convergence protocol is given. It is shown that the balanced structure of the directed symbol graph which represents the communication relationship between agents and the existence of a spanning tree are the weakest topological conditions to ensure that the protocol is a mean-square linear bidirectional convergence protocol. In particular, if the directed symbol graph is weighted and balanced, the proposed control protocol is the mean square average bidirectional convergence protocol. That is, the mean square of each self-principal position converges to a random variable with a weighted average expected to be the initial value with different symbols of the same module.
【学位授予单位】：曲阜师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O175.5
本文编号：2375707
[Abstract]:In multi-agent systems, each self-agent is not only a cooperative relationship, but also a competition relationship. In this paper, the problem of bidirectional convergence of multi-agent systems under the influence of communication noise is studied. For the first time, we introduce time-varying convergence gain in the design of bidirectional convergence protocol. By means of state transition matrix and algebraic graph theory, the properties of the solution of the closed-loop system are analyzed. A necessary and sufficient condition for the proposed bidirectional convergence protocol to be a mean-square linear bidirectional convergence protocol is given. It is shown that the balanced structure of the directed symbol graph which represents the communication relationship between agents and the existence of a spanning tree are the weakest topological conditions to ensure that the protocol is a mean-square linear bidirectional convergence protocol. In particular, if the directed symbol graph is weighted and balanced, the proposed control protocol is the mean square average bidirectional convergence protocol. That is, the mean square of each self-principal position converges to a random variable with a weighted average expected to be the initial value with different symbols of the same module.
【学位授予单位】：曲阜师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O175.5
【参考文献】
相关期刊论文 前2条
1 ;CONSENSUS CONTROL FOR LEADER-FOLLOWING MULTI-AGENT SYSTEMS WITH MEASUREMENT NOISES[J];Journal of Systems Science & Complexity;2010年01期
2 ;Sampled-data based average consensus with measurement noises: convergence analysis and uncertainty principle[J];Science in China(Series F:Information Sciences);2009年11期
，本文编号：2375707
本文链接：https://www.wllwen.com/kejilunwen/yysx/2375707.html
