网络化Lagrange系统的分群一致性

发布时间：2018-03-19 15:04

本文选题：分群一致　切入点：Lagrange网络　出处：《上海大学》2016年博士论文　论文类型：学位论文

【摘要】：近年来具有Lagrange个体动力学结构和特征的网络化动力学系统协调控制己引起人们的广泛关注：这主要是由于它能描述在复杂集成化生产进程中许多物理和力学对象,其中具有灵活性、机动性、可靠性和可操作性是可期待甚至是必要的特征.另一方面,与完全同步或者一致相比,有些情况下分群一致能更好地处理复杂多智能体系统协调控制问题.本文研究了Lagrange网络分群一致(同步)控制及相关问题,主要工作有以下几点：一.不确定Lagrange网络的自适应分群一致.在无循环划分和平衡耦合划分两种有向分群拓扑结构下,给出了参数不确定Lagrange网络的分布式自适应分群一致控制策略.基于几何图论和矩阵理论,提出了一种新的分解方法,并在此方法基础上讨论了受控Lagrange系统的稳定性,阐述了Lagrange网络实现分群一致的充分必要条件.所给的控制算法有以下两个特点：在无循环划分网络下,网络的无循环结构就能够保证实现分群一致；而在平衡耦合划分拓扑下,简单的代数准则能够实现系统的分群目标.进一步地,当网络拓扑是有向无循环划分网络时,通过设计一种积分控制器,系统的分群一致状态能够显式地表达出来,因而提供了一种实现预定分群一致目标的系统的控制方法.二.受扰动的不确定Lagrange网络的区域分群一致.在无循环划分分群网络结构下,研究了网络化参数不确定Lagrange系统自适应分群区域一致控制问题,并对没有引导者和具有引导者这两种情况分别提出了相应的自适应区域分群一致控制算法.对于任意预先给定的一致误差界,本文给出的充分条件总是能够保证达到期望的区域分群一致.与已有的结果比较,所给的控制算法能够实现全局的稳定性,即对任意的有限初值状态,所给的算法都能够实现系统的区域分群一致.三.基于动力学基本方程的Lagrange网络的分群同步.利用动力学基本方程理论,从分析动力学的角度提出了一种网络化Lagrange系统的优化控制方法框架.提出的控制算法的主要特点是在控制要求(约束)中引入了网络的拓扑结构.所给的算法在结构上可以分成两部分：第一部分描述了网络的结构,第二部分是系统的主动反馈控制.基于图论和矩阵理论,给出了网络化Lagrange系统实现分群同步的充分条件.最后,由陀螺仪构成的网络系统作为数值模拟对象来说明了这种优化控制算法中各个参数的作用,并验证了所给理论结果的有效性.
[Abstract]:In recent years, the coordinated control of networked dynamic systems with individual dynamics structure and characteristics of Lagrange has attracted wide attention: this is mainly because it can describe many physical and mechanical objects in complex integrated production processes. Flexibility, mobility, reliability and maneuverability are desirable and even necessary features. On the other hand, compared to complete synchronization or consistency, In some cases, clustering uniformity can better deal with the coordinated control problem of complex multi-agent systems. In this paper, cluster consistent control and related problems in Lagrange networks are studied. The main works are as follows: 1. The adaptive clustering of uncertain Lagrange networks is consistent. A distributed adaptive cluster consistent control strategy for Lagrange networks with uncertain parameters is presented. Based on geometric graph theory and matrix theory, a new decomposition method is proposed, and the stability of controlled Lagrange systems is discussed. In this paper, the necessary and sufficient conditions for clustering uniformity in Lagrange networks are described. The proposed control algorithm has the following two characteristics: in the case of non-cyclic partition networks, the no-cyclic structure of the network can guarantee the clustering uniformity, while in the balanced coupling partition topology, Moreover, when the network topology is directed uncircularly partitioned network, by designing an integral controller, the cluster uniform state of the system can be expressed explicitly. Therefore, this paper provides a control method for the system to achieve the consistent target of predetermined clustering. Secondly, the region clustering of disturbed uncertain Lagrange networks is consistent. Under the network structure of non-cyclic partitioning clustering, In this paper, the problem of adaptive cluster region consistent control for networked parameter uncertain Lagrange systems is studied. The corresponding adaptive region clustering consistent control algorithm is proposed for the two cases of no guide and one with guide. For any given consistent error bound, The sufficient conditions given in this paper can always ensure that the desired regional clustering is consistent. Compared with the existing results, the proposed control algorithm can achieve global stability, that is, for arbitrary finite initial value states, The proposed algorithms can realize the regional cluster consistency of the system. Thirdly, the cluster synchronization of Lagrange networks based on the fundamental equations of dynamics is used. An optimal control method framework for networked Lagrange systems is proposed from the point of view of analysis dynamics. The main feature of the proposed control algorithm is the introduction of the network topology into the control requirements (constraints). The structure can be divided into two parts: the first part describes the structure of the network, The second part is the active feedback control of the system. Based on graph theory and matrix theory, the sufficient conditions for the realization of cluster synchronization in networked Lagrange systems are given. The network system composed of gyroscopes is used as a numerical simulation object to illustrate the function of the parameters in the optimal control algorithm and verify the validity of the theoretical results given.
【学位授予单位】：上海大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：O231

【参考文献】