An improved method for general perturbation eigenvalue problems
Sun Hui
Abstract
Sun Hui
Abstract
An improved perturbation method for general eigenvalue problems is presented. The proposed method is universally applicable to all the three cases of eigenvalue problems: distinct, repeated and nearly equal eigenvalues. The improvement is made to the orthogonal and normality condition and the solving condition of the eigenvector coefficient matrix adopted in the existing methods. Illustrative examples are given to demonstrate the precision of the improved technique.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
An improved perturbation method for general eigenvalue problems is presented. The proposed method is universally applicable to all the three cases of eigenvalue problems: distinct, repeated and nearly equal eigenvalues. The improvement is made to the orthogonal and normality condition and the solving condition of the eigenvector coefficient matrix adopted in the existing methods. Illustrative examples are given to demonstrate the precision of the improved technique.
Key concepts: Eigenvalues and eigenvectors, Eigenvalue perturbation, Normality, Perturbation (astronomy), Mathematics, Applied mathematics, Divide-and-conquer eigenvalue algorithm, Inverse iteration