A multigrid method for the generalized symmetric eigenvalue problem: Part II—performance evaluation
✍ Scribed by T. Hwang; I. D. Parsons
- Publisher
- John Wiley and Sons
- Year
- 1992
- Tongue
- English
- Weight
- 987 KB
- Volume
- 35
- Category
- Article
- ISSN
- 0029-5981
No coin nor oath required. For personal study only.
✦ Synopsis
Abstract
The behaviour of the multigrid method is studied by solving some simple test problems. Optimum choices for some of the parameters are discussed, together with effective techniques for solving the coarse mesh correction equation. The effects of ill‐conditioning on the performance of the algorithm are examined. In particular, thin shells and non‐uniform meshes are observed to slow convergence. The solution of practical, large scale problems demonstrates the utility and speed of the proposed multigrid method. For example, the first 10 eigensolutions of a stiffened plate problem with 193 536 degrees‐of‐freedom were computed in 1.6 CPU hours using 42 Mbytes of memory on a Convex C240.
📜 SIMILAR VOLUMES
A new iterative method based on a Newton correction vector for extension of the Krylov subspace, its diagonal, and band versions are proposed for calculation of selected lowest eigenvalues and corresponding eigenvectors of the generalized symmetric eigenvalue problem. Additionally, diagonal and band
A new version of the program MEIGEN is presented for the eigenvalue problem of Sturm-Liouville-type linear equations in Milne's method. Use of the spline function and the WKB approximation provide a high-speed method for calculating eigenvalues and eigenfunctions avoiding divergence problems.
## Abstract In the numerical modelling of mechanical systems, eigenvalue problems occur in connection with the evaluation of resonance frequencies, buckling modes and other more esoteric calculations. The matrices whose eigenvalues are sought sometimes have a skew‐symmetric component and the presen