Re: Estimating the condition of an upper triangular matrix
From: Peter Spellucci (nospamspellucci_at_fb04373.mathematik.tu-darmstadt.de)
Date: 06/17/04
- Next message: Ioannis Botonakis: "Re: FEM for heat transfer in cylindrical coordinates"
- Previous message: Peter Spellucci: "Re: Combining linear regularization methods"
- In reply to: Thomas Schmelzer: "Estimating the condition of an upper triangular matrix"
- Next in thread: Michael Fairbank: "Re: Estimating the condition of an upper triangular matrix"
- Messages sorted by: [ date ] [ thread ]
Date: Thu, 17 Jun 2004 09:03:44 +0000 (UTC)
In article <2jcssiFvv1joU1@uni-berlin.de>,
"Thomas Schmelzer" <thomas-schmelzer@gmx.de> writes:
>Dear experts,
>given a small (10 \times 10) matrix in upper triangular form I am seeking an
>algorithm to estimate the condition (i.e. the ratio of the largest and
>smallest singular value) of that matrix. I want to avoid any kind of
>overestimation. What is the state of the art on that field?
>cheers
>Thomas
>
>
>
>
>
if you want to avoid any kind of overestimation then compute it, using svd or
the computation of the largest and the lowest eigenvalue of R'*R
(von Mises for the largest and inverse iteration with shift zero for the
smallest). since you habve already the factorization, no problem.
hth
peter
- Next message: Ioannis Botonakis: "Re: FEM for heat transfer in cylindrical coordinates"
- Previous message: Peter Spellucci: "Re: Combining linear regularization methods"
- In reply to: Thomas Schmelzer: "Estimating the condition of an upper triangular matrix"
- Next in thread: Michael Fairbank: "Re: Estimating the condition of an upper triangular matrix"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|