矩阵特征值的几个扰动定理

1 引言 设A∈C~(n×m),B∈C~(m×m)(m≤n),它们的特征值分别为{λ_k}_(k=1)~n和{μ_k}_(k=1)~m.令 R=AQ-QB (1)这里Q∈C~(n×m)为列满秩矩阵.Kahan研究了矩阵A在C~(n×m)上的Rayleigh商的性质,证明了下列定理:设A为Hermite矩阵,Q为列正交矩阵,即Q~HQ=I,而B=Q~HAQ,则存在 1,2,… ,n的某个排列π,使得 {sum from j=1 to m │μ_j-λ_(π(j))│~2}~(1/2)≤2~(1/2)‖R‖_F (2)其中R如(1)所示,‖·‖_F为矩阵的Frobenius范数.刘新国在[2]中将此定理推广到B为可对角化矩阵的情形,并且还建立了较为一般的扰动定理:设A为正规矩阵,B为可对角化矩阵;存在非奇异矩阵G,使得G~(-1)BG为对角阵,则存在1,2,…,n的某个排列π,使得 │μ_j-λ_(π(j))│≤2(2~(1/2))nK(G)_(σ_m~(-1))‖R‖_F,j=1,2,…,m. (3)     

关于Wielandt-Hoffman定理

关于正规矩阵的任意扰动,有下述定理成立. 定理1.设A为n阶正规矩阵,C为n阶任一矩阵.A的特征值为λ_1,…,λ_n,C的特征值为μ_1…,μ_n.C~H表示C的转置共轭,||·||_2与||·||_F分别表示矩阵的谱范数与Frobenius范数.记     

线性流形上亚半正定阵的一类逆特征值问题

1　引言与引理设 Rm× n表示所有 m× n实矩阵集合 ,m=n时 ,Rm× n简记为 Rm;Rm0 表示所有 m阶亚半正定阵集合 ,即 Rm0 ={ A∈Rm× m|YTAY≥ 0 , Y∈Rm× 1 } ;ORm表示 m阶正交矩阵集合 ;A+表示矩阵 A的 Moore-Penrose广义逆 ;‖·‖表示 Frobenius范数 .In 表示 n阶单位阵 ,有时令SE={ A∈ Rm× m|‖ AE -F‖ =min,E,F∈ Rm× k} ,(1 .1 )则 SE是线性流形 .文 [1 ] ,[2 ]分别研究了 SE上实对称矩阵及实对称半正定阵的逆特征值问题 ,本文将进一步研究 SE上亚半正定阵的一类逆特征值问题 ,具体叙述如下 :问题 　给定 X,B∈R…     

关于矩阵范数的几个不等式

正1引言文中,用M_n表示n×n复矩阵全体,用‖·‖表示任意的酉不变范数,分别用|λ_n(A)|≤…≤|λ_1(A)|,s_n(A)≤…≤s1(A)来表示矩阵A的特征值和奇异值,用|A|=(A~*A)~(1/2)表示A的绝对值算子.     

关于径向矩阵与谱矩阵

§1.引言与记号 设A∈C~(s×n),则称 ‖A‖=‖AX‖/‖X‖ 为A的谱模(谱范数),其中‖X‖表示向量X∈C~(n×1)的Euclid范数。即当X=(x_1,…,x_n)~(?)时,‖X‖=(XX)~1/2=sum from i=1 to n(|X_1|~2)~1/2;‖AX‖为向量AX的Euclid范数。 如众周知,我们有如下结论: 引理 1[1]、设A、B∈C~(n×n),则谱模满足范数的三个条件: 1>.恒正性:‖A‖≥0且‖A‖=0 A=0; 2>.齐次性:若α∈C,则‖αA‖=|α|·‖A‖; 3>.三角不等式:‖A+B‖≤‖A‖+‖B‖。     

关于Kahan定理的一个注记

设A∈C~(n×n),B∈C~(k×k)均为Hermite矩阵,它们的特征值分别为{λ_j}_(j=1)~n和{μ_j}_(j=1)~k(k≤n);Q∈~(n×k)为列满秩矩阵.令 (1) 则存在A的k个特征值λ_(j_2),λ_(j_2),…,λ_(j_k),使得 (2) 其中σ_k为Q的最小奇异值,||·||_2表示矩阵的谱范数.这是著名的Kahan定理·1996年曹志浩等在[2]中将(2)加强为 (3) 这是Kahan的猜想.在本文中,我们讨论将Kahan定理中“B为k阶Hermite矩阵”改为B为k阶(任意)方阵后,特征值的扰动估计,有以下结果. 定理 设A∈C~(n×n)为Hermite矩阵,其特征值为{λ_j}_(j=1)~n,B∈C~(k×k)的特征值为{μ_j}_(j=1)~k,而Q∈C~(n×k)为列满秩矩阵.则存在A的k个特征值λ_(j_1),λ_(j_2),…,λ_(j_k),使得     

满足AB=αBA的方阵A,B和AB的特征值的性质及其联系

设m阶方阵A,B满足AB=αBA,其中α=e~(2kπi/n),k,n为互素整数且n≥2.证明了σ(AB)■{α~(j-((n-1)/2))λ_AλB|λA∈σ(A),λB∈σ(B),j=0,1,…,n-1}及其它相关的结果,其中σ(A)表示方阵A的所有特征值的集合.     

Hermite矩阵与可对称化矩阵特征值之间的扰动上界

正1引言矩阵特征值的扰动问题,就是研究矩阵元素的改变对矩阵特征值的影响.设矩阵A,B为n阶复矩阵,矩阵B为矩阵A经过扰动之后的矩阵,且λ(A)={λ_i},λ(B))={μ_i},研究矩阵特征值的扰动就是研究λ(A)与λ(B)之间的差距,一般用2范数和Frobenius范数来描述它们之间的差距.矩阵特征值问题是由于处理数据时存在误差而引起的,使得到的特征值往往是经过     

任意矩阵特征值扰动的估计

一、引言 设A,B是n×n方阵,E≡A-B,矩阵特征值扰动问题的一种提法是[2]:给定B的特征值μ,估计|μ-λ_i|的极小值的上界,这里λ_i是A的某一个特征值。 Bauer和Fikl在1969年给出如下定理:     

具有左、右特征向量及特征值约束下逆特征值问题

§1 问题的提法R~(n×m)表示所有 n×m 阶实阵集合,(A)表示矩阵 A 的列空间,A~+表示 A 的 Moore-Penrose 广义逆,P_A=AA~+表示到(A)的正交投影核子;I_n 表示 n 阶单位阵,‖·‖_F 表示 Frobenius 范数。问题Ⅰ给定X,Y∈~(n×m),Λ=diag(λ_1,λ_2,…,λ_m)∈R~(m×m),找 A∈R~(n×m),使得问题Ⅱ给定 A~*∈R~(n×n),找∈S_E,使得‖A~*-‖_F=‖A~*-A‖_F,其中 S_E是问题Ⅰ的集合。本文讨论问题Ⅰ有解的充分与必要条件,且求出 S_E的表达式,同时给出的表达式。     

关于正规矩阵对广义特征值新的扰动界限

本文考虑了正规矩阵对的任意扰动时广义特征值的变化情况,给出了正规矩阵对任意扰动的Hoffman-Wielandt型扰动界,推广了正规矩阵对的相应的扰动结果.     

一类矩阵对的广义特征值的扰动界限

关于矩阵特征值的扰动,下面的结果是熟知的:若A与C皆为n阶正规矩阵,它们的特征值分别为α_1,…,α_n与γ_1,…,γ_n,则据Wielandt-Hoffman定理,存在1,…,n的一个排列k_1,…,k_n,使得     

多参数结构特征二阶灵敏度

提出了一种有效计算多参数结构特征值与特征向量二阶灵敏度矩阵--Hessian矩阵的方法.将特征值和特征向量二阶摄动法转变为多参数形式,推导出二阶摄动灵敏度矩阵,由此得到特征值和特征向量的二阶估计式.该法解决了无法用直接求导法计算特征值和特征向量二阶灵敏度矩阵的问题.数值算例说明了该算法的应用和计算精度.     

Perturbation bounds for means of eigenvalues and invariant subspaces

When a matrix is close to a matrix with a multiple eigenvalue, the arithmetic mean of a group of eigenvalues is a good approximation to this multiple eigenvalue. A theorem of Gershgorin type for means of eigenvalues is proved and applied as a perturbation theorem for a degenerate matrix.For a multiple eigenvalue we derive bounds for computed bases of subspaces of eigenvectors and principal vectors, relating them to the spaces spanned by the last singular vectors of corresponding powers of the matrix. These bounds assure that, provided the dimensionalities are chosen appropriately, the angles of rotation of the subspaces are of the same order of magnitude as the perturbation of the matrix.A numerical example is given.     

Outliers in the spectrum of iid matrices with bounded rank perturbations

It is known that if one perturbs a large iid random matrix by a bounded rank error, then the majority of the eigenvalues will remain distributed according to the circular law. However, the bounded rank perturbation may also create one or more outlier eigenvalues. We show that if the perturbation is small, then the outlier eigenvalues are created next to the outlier eigenvalues of the bounded rank perturbation; but if the perturbation is large, then many more outliers can be created, and their law is governed by the zeroes of a random Laurent series with Gaussian coefficients. On the other hand, these outliers may be eliminated by enforcing a row sum condition on the final matrix.     

一种通用的复模态矩阵摄动法

对于非自伴随系统,提出了一种通用的复模态矩阵摄动法.该法能同时适用于孤立特征值,重特征值及密集(相近)特征值三种复特征值情况.由复特征子空间缩聚技术求解低阶摄动项,高阶摄动项则由逐次逼近过程求得.三个计算实例表明,该通用方法合理可靠,精度高.     

Sylvester方程在矩阵扰动分析中的应用

§1.引言 矩阵扰动分析的研究对于矩阵论的发展及数值分析问题计算结果的分析和处理都有重要意义.有关特征值、广义特征值及最小二乘问题的主要研究结果均含于[1]中,[5]运用二次方程根的判别法通过对代数Ricatti方程的解的估计给出了QR分解因子及Cholesky因子的扰动分析,但论证方法及所得结果都比较复杂且所求条件很强.[3]和     

On the closeness of eigenvalues and singular values for almost normal matrices

It is proved that a matrix is almost normal if and only if its singular values are close to the absolute values of its eigenvalues. In the special case when the spectral norm and spectral radius are close, it is proved that the dominating eigenvalue is well conditioned. A refinement of a perturbation theorem by Henrici is proved, and its numerical behavior is compared with adaptations of the Gerschgorin theorem. It is specially devised for almost triangular matrices.     

Perturbation theory of selfadjoint matrices and sign characteristics under generic structured rank one perturbations

For selfadjoint matrices in an indefinite inner product, possible canonical forms are identified that arise when the matrix is subjected to a selfadjoint generic rank one perturbation. Genericity is understood in the sense of algebraic geometry. Special attention is paid to the perturbation behavior of the sign characteristic. Typically, under such a perturbation, for every given eigenvalue, the largest Jordan block of the eigenvalue is destroyed and (in case the eigenvalue is real) all other Jordan blocks keep their sign characteristic. The new eigenvalues, i.e. those eigenvalues of the perturbed matrix that are not eigenvalues of the original matrix, are typically simple, and in some cases information is provided about their sign characteristic (if the new eigenvalue is real). The main results are proved by using the well known canonical forms of selfadjoint matrices in an indefinite inner product, a version of the Brunovsky canonical form and on general results concerning rank one perturbations obtained.     

A perturbation method for optimizing matrix stability

We present the first practical perturbation method for optimizing matrix stability using spectral abscissa minimization. Using perturbation theory for a matrix with simple eigenvalues and coupling this with linear programming, we successively reduce the spectral abscissa of a matrix until it reaches a local minimum. Optimality conditions for a local minimizer of the spectral abscissa are provided and proved for both the affine matrix problem and the output feedback control problem. Experiments show that this novel perturbation method is efficient, especially for a matrix with the majority of whose eigenvalues are already located in the left half of the complex plane. Moreover, unlike most available methods, the method does not require the introduction of Lyapunov variables. The method is illustrated for a small size matrix from an affine matrix problem and is then applied to large matrices actually arising from more sophisticated control problems used in the design of the Boeing 767 jet and a nuclear powered turbo-generator.