Re: matrix inversion

[David Bateman]

Gorazd Brumen wrote:
> Hi David,
>
> Yes, it is actually qute easy to prove that
> the inverse is a series of t NxN full matrix inversions.
> Since running dmperm gives me the following result
>
> dmperm (A)
> error: dmperm: not available in this version of Octave
>
> (could it be that I am running octave 2.9.7 with forge for (up to)
> 2.9.6???)
>
> I checked it in matlab though and it is as I have written:
> t NxN inversions, you just got confused because in your
> example t and N are the same.
> I was astonished to there is already a procedure to do what I was
> doing by hand. I will use it, it simplifies
> my code significantly.
>
>
> If I get dmperm running in Octave, I would like to try your code, and
> once again, thanks for the kind replies.
>
>   
dmperm was introduced in octave in version 2.9.5.... The error message
you have means that you don't have CXSparse from UFSparse (SparseSuite)
installed correctly... If possible I'd suggest using the debian or
fedora octave packages as all of these dependencies are correctly
configured...

Regards
David


-- 
David Bateman                                address@hidden
Motorola Labs - Paris                        +33 1 69 35 48 04 (Ph) 
Parc Les Algorithmes, Commune de St Aubin    +33 6 72 01 06 33 (Mob) 
91193 Gif-Sur-Yvette FRANCE                  +33 1 69 35 77 01 (Fax) 

The information contained in this communication has been classified as: 

[x] General Business Information 
[ ] Motorola Internal Use Only 
[ ] Motorola Confidential Proprietary