Re: What's the algorithm used in 'inverse matrix' of Octave?

[CdeMills]

Minkyung Kersti Son wrote:
> 
> Hello, users~!
> I wonder what the algorithm is used in the calculation of inverse matrix
> of Octave.
> I got a inverse matrix (complex number) with octave, however I failed to
> get a inverse matrix with my fortran code using umfpack.The code of
> umfpack yielded non-exact inverse matrix, as the dimension of matrix
> becomes to large.UMFPACK employed a LU decomposition, but I don't know
> what the kind of LU decomposition is used.
> Where can I read the source code or algorithm of a calculation of inverse
> matrix of Octave?
> Very thanks.
> 
> 
Maybe your matrix is basically ill-conditionned ? Are there specific
properties you can use to improve the accuracy, like your matrix being
symmetric, positive definite, can be factored, ... ? If the answer to one of
those questions is 'yes', then there are numeric transformations which can
improve the accuracy. 

Regards

Pascal

 

--
View this message in context: 
http://octave.1599824.n4.nabble.com/What-s-the-algorithm-used-in-inverse-matrix-of-Octave-tp3732054p3733575.html
Sent from the Octave - General mailing list archive at Nabble.com.