On the Lidskii-Mirsky-Wielandt Theorem, preprint, available at httpwww.wm.eduCASMINEQmprepr.pdfVIP

下载本文档

7
0
约6.96万字
约 32页
2017-04-05 发布于江苏
举报
版权申诉

On the Lidskii-Mirsky-Wielandt Theorem, preprint, available at httpwww.wm.eduCASMINEQmprepr.pdf

1、原创力文档（book118）网站文档一经付费（服务费），不意味着购买了该文档的版权，仅供个人/单位学习、研究之用，不得用于商业用途，未经授权，严禁复制、发行、汇编、翻译或者网络传播等，侵权必究。。
2、本站所有内容均由合作方或网友上传，本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺！文档内容仅供研究参考，付费前请自行鉴别。如您付费，意味着您自己接受本站规则且自行承担风险，本站不退款、不进行额外附加服务；查看《如何避免下载的几个坑》。如果您已付费下载过本站文档，您可以点击这里二次下载。
3、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等，请点击“版权申诉”（推荐），也可以打举报电话：400-050-0827(电话支持时间：9:00-18:30)。
4、该文档为VIP文档，如果想要下载，成为VIP会员后，下载免费。
5、成为VIP后，下载本文档将扣除1次下载权益。下载后，不支持退款、换文档。如有疑问请联系我们。
6、成为VIP后，您将拥有八大权益，权益包括：VIP文档下载权益、阅读免打扰、文档格式转换、高级专利检索、专属身份标志、高级客服、多端互通、版权登记。
7、VIP文档为合作方或网友上传，每下载1次，网站将根据用户上传文档的质量评分、类型等，对文档贡献者给予高额补贴、流量扶持。如果你也想贡献VIP文档。上传文档

On the Lidskii-Mirsky-Wielandt Theorem, preprint, available at httpwww.wm.eduCASMINEQmprepr

The Lidskii-Mirsky-Wielandt Theorem {Additive and Multiplicative VersionsChi-Kwong Li and Roy MathiasAbstractWe use a simple matrix splitting technique to give an elementary new proof ofthe Lidskii-Mirsky-Wielandt Theorem and to obtain a multiplicative analog of theLidskii-Mirsky-Wielandt Theorem, which we argue is the fundamental bound in thestudy of relative perturbation theory for eigenvalues of Hermitian matrices and singularvalues of general matrices. We apply our bound to obtain numerous bounds on thematching distance between the eigenvalues and singular values of matrices. Our resultsstrengthen and generalize those in the literature.1 IntroductionGiven an n n Hermitian matrix A let 1(A) 2(A) n(A) denote its orderedeigenvalues. The singular values of an m n matrix A, are dened byi(A) = qi(AA); i = 1; 2; : : : ; q minfm;ng:An m n matrix can have at most q minfm;ng non-zero singular values so we shall onlyconsider its largest q singular values.The Lidskii-Mirsky-Wielandt theorem (e.g., see [19, Theorem IV.4.8] or [2, Theorem 9.4])states:Theorem 1.1 Let A and E be n n Hermitian matrices. Then for any indices 1 i1 i2 ik n we have kXj=1[ij (A+ E) ij (A)] kXj=1 j(E): (1.1)This is a very useful result in matrix theory. It is a majorization relation that implies abound on the matching distance between the eigenvalues of A and those of A+ E, namely,kXj=1 jij(A+ E) ij (A)j kXj=1 j(E): (1.2)Department of Mathematics, The College of William and Mary, Williamsburg, Virginia 23187, USA.E-mail: ckli@math.wm.edu, mathias@math.wm.edu. Both authors were supported by grants from the Na-tional Science Foundation. 1 This result can also be stated in terms of norms. A norm kk on Rn is called a symmetricnorm if it is both permutation invariant, that is,kPxk = kxk; 8 x 2 Rn; permutation matrices Pand absolute, that is, k(xi)ni=1k = k(jxij)ni=1k; 8 x 2 Rn:Symmetric norms are also sometimes called symmetric gauge functions. For ev