A symmetric band Lanczos process based on coupled recurrences and some applications (Q2780539)
From MaRDI portal
| This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use this page instead for the normal view: A symmetric band Lanczos process based on coupled recurrences and some applications |
scientific article; zbMATH DE number 1729154
| Language | Label | Description | Also known as |
|---|---|---|---|
| English | A symmetric band Lanczos process based on coupled recurrences and some applications |
scientific article; zbMATH DE number 1729154 |
Statements
15 April 2002
0 references
band Lanczos process
0 references
symmetric matrix
0 references
block-Krylov subspace
0 references
multiple starting vectors
0 references
orthogonal basis
0 references
reduced-order modeling
0 references
passivity
0 references
circuit simulation
0 references
eigenvalue problem
0 references
numerical examples
0 references
0.85426784
0 references
0.85045856
0 references
0.84681207
0 references
0.83920723
0 references
0.83779824
0 references
0.8377523
0 references
A symmetric band Lanczos process based on coupled recurrences and some applications (English)
0 references
The authors propose a new variant of the band Lanczos process for symmetric matrices and multiple starting vectors. In the standard version an \(n \times n\) projection \(T_n^s\) of the given symmetric matrix onto the \(n\)-dimensional subspace spanned by the first \(n\) Lanczos vectors is computed directly. In the new version coupled recurrences involving two sets of basis vectors are used to produce the factors of an \(LDL^T\) factorization of an \(n \times n\) matrix \(T_n\) which is closely related to \(T_n^s\). NEWLINENEWLINENEWLINEApplications of the proposed method to reduced-order modeling of large electronic circuits and to generalised symmetric eigenvalue problems are discussed and numerical examples are presented. The numerical results show that the new variant is more robust and accurate than the standard version.
0 references