【摘要】 <正> 设A,B=A+E∈C^n×n。本文讨论A扰动以E后,其特征向量的扰动界。以（λ,x）,（λ（B）,x（B））分别表示A,B的特征对,当λ为单特征值时,有且若A为可正规化阵,还可得（1）的精确界。但是,特征向量可以相差模为1的常数因子,此时需重议（1）。再者,若λ为重特征值,相应的特征向量未必唯一确定,有时也是更多还原

【Abstract】 Some better perturbation bounds of eigenvalues and eigenvectors for arbitrary matrices are established.The results that the relative perturbation bounds of eigenvectors associated both simple and multiple eigenvalue have the same order of magnitude as that for associated eigenvalues arc proved.The perturbation bounds of the Schur decomposition for normal matrices and the SVD for complex matrices arc discussed as applicatoins of perturbation bounds for eigenvectors mentioned above.更多还原