Weighted Harmonic Projection Method for Computing the Eigenvalues of Nonsymmetric Matrix

Abstract

The harmonic projection method can be used to find interior eigenpairs of large matrices. Given a target point or shift t to which the needed interior eigenvalues are close, the desired interior eigenpairs are the eigenvalues nearest t and the associated eigenvectors. In this paper, we present a new algorithm, which is called weighted harmonic projection algorithm for computing the eigenvalues of a nonsymmetric matrix. The implementation of the algorithm has been tested by numerical examples, the results show that the algorithm converges fast and works with high accuracy