Alan_Weiss <aweiss@mathworks.com> wrote in message <kmtpr0$q8t$1@newscl01ah.mathworks.com>...
> On 5/14/2013 11:48 AM, Huanxin wrote:
> > dear all, I used a code in matlab 2008a, but when I used in matlab
> > r2012b, the matlab says : "The LevenbergMarquardt option is no longer
> > valid."
> > the code is : "options =
> > optimset('Display','off','TolX',eps,'TolFun',eps,'LargeScale','on',...
> > 'LevenbergMarquardt');"
> > so what can I use to change this option? I tried comment this
> > LevenbergMarquardt, it gives the same result for my currently using
> > data, but I am not sure other case it will works or not.
> > Thank you!
>
> See the R2008b release notes:
> http://www.mathworks.com/help/optim/release-notes.html
>
> In short, use the Algorithm option in lsqnonlin or lsqcurvefit, and
> don't use the LargeScale option at all. If you want to use the
> Levenberg-Marquardt algorithm, set the Algorithm option to
> 'levenberg-marquardt'.
>
> And by the way, I think you have questionable settings for your other
> options, too. As explained in
> http://www.mathworks.com/help/optim/ug/tolerances-and-stopping-criteria.html
> it is usually not a good idea to set tolerances smaler than 1e-14,
> unless you have a very good reason for doing so.
>
> Alan Weiss
> MATLAB mathematical toolbox documentation
I used : options = optimset('Display','off','TolX',eps,'TolFun',eps,'LargeScale','on',...
'Algorithm','Levenberg-Marquardt'); it works, I decide to use this. thank you!![]()
> On 5/14/2013 11:48 AM, Huanxin wrote:
> > dear all, I used a code in matlab 2008a, but when I used in matlab
> > r2012b, the matlab says : "The LevenbergMarquardt option is no longer
> > valid."
> > the code is : "options =
> > optimset('Display','off','TolX',eps,'TolFun',eps,'LargeScale','on',...
> > 'LevenbergMarquardt');"
> > so what can I use to change this option? I tried comment this
> > LevenbergMarquardt, it gives the same result for my currently using
> > data, but I am not sure other case it will works or not.
> > Thank you!
>
> See the R2008b release notes:
> http://www.mathworks.com/help/optim/release-notes.html
>
> In short, use the Algorithm option in lsqnonlin or lsqcurvefit, and
> don't use the LargeScale option at all. If you want to use the
> Levenberg-Marquardt algorithm, set the Algorithm option to
> 'levenberg-marquardt'.
>
> And by the way, I think you have questionable settings for your other
> options, too. As explained in
> http://www.mathworks.com/help/optim/ug/tolerances-and-stopping-criteria.html
> it is usually not a good idea to set tolerances smaler than 1e-14,
> unless you have a very good reason for doing so.
>
> Alan Weiss
> MATLAB mathematical toolbox documentation
I used : options = optimset('Display','off','TolX',eps,'TolFun',eps,'LargeScale','on',...
'Algorithm','Levenberg-Marquardt'); it works, I decide to use this. thank you!