An accelerated subspace iteration method for generalized eigenproblems
โ Scribed by Xuelin Wang; Ji Zhou
- Publisher
- Elsevier Science
- Year
- 1999
- Tongue
- English
- Weight
- 211 KB
- Volume
- 71
- Category
- Article
- ISSN
- 0045-7949
No coin nor oath required. For personal study only.
โฆ Synopsis
An accelerated subspace iteration for generalized eigenproblems is proposed by combining the repeated inverse iteration with the over-relaxation technique. Two schemes are developed to obtain an over-relaxation factor. Numerical results show that the proposed acceleration is ecient and numerically stable for eigenproblems with a large number of eigenpairs required or with close eigenvalues.
๐ SIMILAR VOLUMES
For a heavily damped system, either viscous or hysteretic or both, the homogeneous solution constitutes a generalized complex symmetric eigenproblem [A][x] = l[B]{x}, where [A] and [B] are sparse complex symmetric matrices. The general method to solve the transformed eigenproblem [B] -1 [A]{x} = l{x
An ecient and stable technique to remove the limitation in choosing a shift in the subspace iteration method with shifting is presented. A major diculty of the subspace iteration method with shifting is that, because of the singularity problem, a shift close to an eigenvalue cannot be used, resultin