[SciPy-user] Questions about scipy.optimize.fmin_cobyla

fdu.xiaojf@gmai... fdu.xiaojf@gmai...
Tue Jul 17 02:08:01 CDT 2007


Hi,
Dominique Orban wrote:
 > Anne Archibald wrote:

 >
 > Re-parametrizing your problem is, of course, another option.
 >

I have reparameterized my problem, fmin_cobyla can run with out "math
domain error". But I get following message during minimization:

#--------begin of message---------------
Tolerance of 3.5527136788e-015 reached
Tolerance of 3.5527136788e-015 reached
Tolerance of 3.5527136788e-015 reached
Tolerance of 3.5527136788e-015 reached
Tolerance of -3.5527136788e-015 reached
Tolerance of -3.5527136788e-015 reached

    Return from subroutine COBYLA because the MAXFUN limit has been reached.

    NFVALS = 1000   F =-2.111828E+06    MAXCV = 6.311133E-07
    X = 8.497382E-01  -5.764801E+00  -7.701660E+00   5.986411E-01  -3.959199E+00
       -9.925761E-01  -6.380951E+00   1.034551E+00  -2.976923E+00   1.046551E-01
#--------end of message---------------

What does "Tolerance of 3.5527136788e-015 reached" mean ? Does it mean
some error occured or fmin_cobyla doesn't perform minimization well ?

Regards,

Xiao Jianfeng



More information about the SciPy-user mailing list