Browsing by Author "Li, Yung-Ta"

Now showing 1 - 2 of 2

A semiorthogonal generalized Arnoldi method and its variations for quadratic eigenvalue problems
(2013-03-01)
Huang, Wei-Qiang
;
Li, Tiexiang
;
Li, Yung-Ta
;
Lin, Wen-Wei
;
應用數學系
In this paper, we are concerned with the computation of a few eigenpairs with smallest eigenvalues in absolute value of quadratic eigenvalue problems. We first develop a semiorthogonal generalized Arnoldi method where the name comes from the application of a pseudo inner product in the construction of a generalized Arnoldi reduction for a generalized eigenvalue problem. The method applies the RayleighRitz orthogonal projection technique on the quadratic eigenvalue problem. Consequently, it preserves the spectral properties of the original quadratic eigenvalue problem. Furthermore, we propose a refinement scheme to improve the accuracy of the Ritz vectors for the quadratic eigenvalue problem. Given shifts, we also show how to restart the method by implicitly updating the starting vector and constructing better projection subspace. We combine the ideas of the refinement and the restart by selecting shifts upon the information of refined Ritz vectors. Finally, an implicitly restarted refined semiorthogonal generalized Arnoldi method is developed. Numerical examples demonstrate that the implicitly restarted semiorthogonal generalized Arnoldi method with or without refinement has superior convergence behaviors than the implicitly restarted Arnoldi method applied to the linearized quadratic eigenvalue problem. Copyright (c) 2012 John Wiley & Sons, Ltd.
A Structured Quasi-Arnoldi procedure for model order reduction of second-order systems
(2012-04-15)
Li, Yung-Ta
;
Bai, Zhaojun
;
Lin, Wen-Wei
;
Su, Yangfeng
;
應用數學系
Existing Krylov subspace-based structure-preserving model order reduction methods for the second-order systems proceed in two stages. The first stage is to generate a basis matrix of the underlying Krylov subspace. The second stage is to employ an explicit subspace projection to obtain a reduced-order model with a moment-matching property. An open problem is how to avoid explicit projection so that it will be efficient for truly large scale systems. In addition, it is also desired that a structure-preserving reduced system of order n matches maximum 2n moments. In this paper we propose a new procedure to compute a so-called Structured Quasi-Arnoldi (SQA) decomposition. Once the SQA decomposition is computed, a structure-preserving reduced-order model can be defined immediately from the decomposition without the explicit subspace projection. Furthermore, the reduced model of order n matches maximum 2n moments. Numerical examples demonstrate that the transpose-free SQA-based reduced model is compatible with the two-sided structure-preserving explicit projection methods and is more accurate than the one-sided structure-preserving explicit projection methods due to the higher number of matched moments. (C) 2011 Elsevier Inc. All rights reserved.