home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!sun4nl!cwi.nl!dik
- From: dik@cwi.nl (Dik T. Winter)
- Newsgroups: sci.math
- Subject: Re: Discrete Fields
- Message-ID: <7952@charon.cwi.nl>
- Date: 20 Nov 92 01:17:10 GMT
- References: <1992Nov19.153057.18919@news.lrz-muenchen.de>
- Sender: piet@cwi.nl
- Organization: CWI, Amsterdam
- Lines: 19
-
- In article <1992Nov19.153057.18919@news.lrz-muenchen.de> stpa@nws.e-technik.tu-muenchen.de (St. Paul) writes:
- > I'm looking for literature on matrix computation on
- > discrete fields (SVD, eigenvalues, QR-decomposition etc. ..).
-
- EVD requires an algebraically complete field. This is easy to verify (given
- a polynomial it is simple to write down a matrix using the coefficients of
- the polynomial that has the roots of the polynomial as its eigenvalues).
- QR and SVD can be done if you are able to extract the square root of every
- number in the field (and this is required). You simply follow the standard
- routines for matrices over R. But be aware that methods for SVD are
- iterative, so you need some metric to determine convergence, and also
- some singular values will be in the algebraic completion of the field.
-
- If the matrix has eigenvalues representable in the field in question,
- there are methods that might be able to approximate them, but again,
- you need a metric.
- --
- dik t. winter, cwi, kruislaan 413, 1098 sj amsterdam, nederland
- home: bovenover 215, 1025 jn amsterdam, nederland; e-mail: dik@cwi.nl
-