Marcus,
Thanks. I looked it up this morning after doing some Web searches, but it
appeared to be a tool with a (financial) cost. What worries me is that your
E-mail address indicates that you work for TOMLAB; hence, your advice is
subjective and I must dispense it.
Roy
"Marcus M. Edvall" <medvREMall@tomREMlab.biREMz> wrote in message
news:eed4be3.0@webx.raydaftYaTP...
> Dear Roy,
>
> Use an efficient optimization routine instead. TOMLAB /SNOPT is most
> likely a good option. <http://tomlab.biz>
>
> Best wishes, Marcus
>
> Roy S Schestowitz wrote:
> >
> >
> > FMINSEARCH uses Nelder-Mead simplex to find the minimum and
> > according to its
> > documentation it accepts only the following options: Display, TolX,
> > TolFun,
> > MaxFunEvals, and MaxIter.
> >
> > How do I define the volume it searches and its starting position? I
> > need it
> > to probe a very high (~50 dims) space and it scales quite badly at
> > present.
> >
> > TIA
> >
> > Roy
> >
|