On 03/28/2013 04:32 PM, Steven G. Johnson wrote: > > On Jan 23, 2013, at 6:08 AM, Agnes Meyder <[email protected]> wrote:= >> The input consists of e.g x=11 values (11 dimensions), each of them >> constrained by lower and upper bonds. When I observe the development of >> these 11 values over the course of the optimization, I can observe the >> same consecutive changes in x until step 97. Based on the same score for >> the same x, nlopt-COBYLA chooses diverging changes in step 98 for x in >> the fifth etc. digit after the dot. (See example below.) > > Maybe there is some slight change in the roundoff errors in your code, which > is affecting COBLYA? That is, are you sure that the floating-point values > are *exactly* the same between runs? > Sadly, I do not have this version of the code anymore. The scores could have been different on the very last digits not printed out.
Luckily BOBYQA outperforms COBYLA etc in terms of speed and accuracy for my problem, so I don't have to rely on COBYLA anymore. Thanks for your answer! Agnes -- Dipl.-Inform. (Bioinf.) Agnes Meyder 040-42838-7324 -- [email protected] Research Group for Computational Molecular Design Center for Bioinformatics Hamburg -- http://www.zbh.uni-hamburg.de _______________________________________________ NLopt-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/nlopt-discuss
