求解控制系统部分特征值配置问题新方法  

New algorithms of partial eigenvalue assignment in control system

作　　者：贾仲孝[1] 曹继超[2] 

机构地区：[1]清华大学数学科学系,北京100084 [2]大连理工大学应用数学系,辽宁大连116024

出　　处：《大连理工大学学报》2004年第5期763-768,共6页Journal of Dalian University of Technology

基　　金：国家重点基础研究专项基金资助项目(G19990328).

摘　　要：在控制理论领域里,特征值配置问题是一个经典问题.提出了新的通过部分特征值配置来使大型单输入时不变控制系统稳定化的算法.该算法建立在隐式重新启动的精化Arnoldi方法基础上,适合那些需要对一小部分特征值重新配置的控制系统.同时对配置问题进行了理论分析,证明算法的精度越高,配置后的系统越稳定.与已有的基于隐式重新启动的Arnoldi方法进行比较体现出新方法的优越性.The eigenvalue assignment problem is a classical problem in control systems. New algorithms for the stabilization of large single-input time-invariant control systems by partial eigenvalue assignment are presented. The algorithms are based on the implicitly restarted refined Arnoldi method, and well suited for control systems that require the reassignment of a few eigenvalues. The theoretical analysis for the eigenvalue assignment problem is also presented, and it shows that the higher accuracy of the algorithms, the further stability of the assignmented problem. Finally, the advantage of the new algorithms is shown by comparing with the known algorithms based on the implicitly restarted Arnoldi method.

关 键 词：特征值 隐式 求解 证明 算法 控制理论 领域 重新启动 配置 控制系统 

分 类 号：O151.21[理学—数学] O241.81[理学—基础数学]