Re: Partial Schur decomposition in Matlab
- From: Dirk Diggler <dirk_f_diggler@xxxxxxxxx>
- Date: Wed, 31 Oct 2007 17:51:27 +0000 (UTC)
On Wed, 31 Oct 2007 16:50:58 +0000, Peter Spellucci wrote:
In article <fg9dbf$eo8$1@xxxxxxxxxxxxxxx>,
Dirk Diggler <dirk_f_diggler@xxxxxxxxx> writes:
>Arpack (routine XXaupd.f) naturally calculates orthogonal basis for a
>selected invariant subspace of a linear operator, thus producing a matrix
>with (real) orthonormal columns which can be directly used to obtain a
hardly if the operator isn't self adjoined (hermitian)
and in order to make sense the question implies that an
nonhermitian case is wanted _ otherwise the question had been a joke
peter
Sorry, my mistake. Xneupd does that. And of course, nonhermitian case is
wanted. For example, this is a part of dneupd.f:
c\BeginDoc
c
c\Name: dneupd
c
c\Description:
c
c This subroutine returns the converged approximations to eigenvalues
c of A*z = lambda*B*z and (optionally):
c
c (1) The corresponding approximate eigenvectors;
c
c (2) An orthonormal basis for the associated approximate
c invariant subspace;
c
c (3) Both.
c
c There is negligible additional cost to obtain eigenvectors. An orthonormal
c basis is always computed. There is an additional storage cost of n*nev
c if both are requested (in this case a separate array Z must be supplied).
c
.
- References:
- Partial Schur decomposition in Matlab
- From: Dirk Diggler
- Re: Partial Schur decomposition in Matlab
- From: Peter Spellucci
- Re: Partial Schur decomposition in Matlab
- From: Dirk Diggler
- Re: Partial Schur decomposition in Matlab
- From: Peter Spellucci
- Partial Schur decomposition in Matlab
- Prev by Date: Re: Non-linear least squares fitting problem
- Next by Date: Re: bessel J in Fortran
- Previous by thread: Re: Partial Schur decomposition in Matlab
- Next by thread: need Ebook "fortran 95/2003 explained"
- Index(es):
Relevant Pages
|
|
