Re: [ESA-SoC] Generalized eigenvalue

[András Mihálykó]

Hi, I'm almost done with my BSc thesis, so now, I'll have more time for this
problem. 
I've read throught this whole thread, and realized, that there are such
matrices, where it can be seen, that balancing is harmful: for example
these, http://www.math.wsu.edu/faculty/watkins/pdfiles/balbad.pdf
In this case now the eig() uses dgeev (or dgeevx), so the balancing is done
automatically. If we could decide if an M matrix is such, that nobalance is
better, than eig() could automatically use the 'nobalance' option, so that
users, that don't know which option is better, could have better result. 
Of course if you call eig(M,'balance') for such a matrix, it would use
balancing, just I think we could make the better result deafult, not always
balancing.

András



--
View this message in context: 
http://octave.1599824.n4.nabble.com/ESA-SoC-Generalized-eigenvalue-tp4669870p4670469.html
Sent from the Octave - Maintainers mailing list archive at Nabble.com.