[SciPy-user] fmin_cobyla hangs

[dmitrey]

W/o code sample we can't say anything.
Try scikits.openopt ALGENCAN or lincher solver (latter requires CVXOPT 
installed http://www.ee.ucla.edu/~vandenbe/cvxopt/, former requires 
ALGENCAN with Python connection installed, see 
http://www.ime.usp.br/~egbirgin/tango/py.php)

Get OpenOpt from svn:
svn co http://svn.scipy.org/svn/scikits/trunk/openopt openopt
(then see help(scikits.openopt.NLP))
However, afaik scipy server is down for now.

Regards, D.

Christian K. wrote:
> Hi,
> sometimes when running fmin_cobyla it never returns but keeps the cpu load high.
> The last output (with iprint=2) is the following:
>
>    The initial value of RHO is 1.000000E+00  and PARMU is set to zero.
>
>    NFVALS =    1   F = 1.635286E+03    MAXCV = 5.335792E-02
>    X =-2.000000E-02   9.600000E-02   2.120000E-01   3.280000E-01   4.440000E-01
>        5.600000E-01
>
>    Increase in PARMU to 1.344559E+03
>
>    Increase in PARMU to 4.872087E+04
>
> Does anybody know enough of the internals of fmin_cobyla to imagine what is
> going on? In case it helps, I will try to build a sample script/data package.
>
> Christian
>
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
>
>
>