Re: [Help-gsl] How to constrain parameters in NonLinear Least-Squares Fitting?

[Patrick Alken]

I don't have a link, but using the example in the manual, in the exp_f() 
function, you could check if the parameters are in a certain range and 
if not, add a large value to the corresponding residual, say 1000.0.

Alternatively, the GPL levmar library 
(http://users.ics.forth.gr/~lourakis/levmar) contains direct support for 
constrained nonlinear least squares optimization.

Patrick

On 06/30/2014 12:10 PM, Eduardo N. Hering wrote:
> Thanks for your answer, Patrick.
>
> Can you send me a link with an example on where the
> penalty function should be called, and which format
> should it have?
>
> Thanks a lot.
>
> Eduardo.
>
>
> > On 30 Jun 2014, at 18:59, Patrick Alken <address@hidden> wrote:
> >
> > This is not currently possible in GSL, unless you check the parameter values 
> > yourself and increase the penalty function accordingly.
> >
> > On 06/24/2014 11:03 AM, Eduardo N Hering wrote:
> > > Hello.
> > >
> > > Is it possible to constrain parameters in nonlinear least-squares fitting?
> > >
> > > Thanks for your attention,
> > >
> > >
> > >
> > > E.N.Hering, D.Sc.
> > >
> > > address@hidden
> > > http://www.enhering.com
> > > _______________________________________
> > >
> > > I support PETA.org.
> > > "Animals are NOT ours to eat, wear,
> > > experiment on, use for entertainment or
> > > abuse in any way” PETA.org
> > > _______________________________________
> > >
> > > CONFIDENTIALITY:
> > > This e-mail and any attachments are
> > > confidential and may be privileged.
> > >
> > > Reproduction of this message, or any
> > > part of it, is not authorised without
> > > explicit permission of the sender.
> > >
> > > If you are not a named recipient,
> > > please notify the sender immediately
> > > and:
> > > 1- do not disclose the contents to
> > > another person;
> > > 2- do not use it for any
> > > purpose;
> > > 3- do not store or copy the information
> > > in any medium.
> E.N.Hering, D.Sc.
>
> address@hidden
> http://www.enhering.com
> _______________________________________
>
> I support PETA.org.
> "Animals are NOT ours to eat, wear,
> experiment on, use for entertainment or
> abuse in any way” PETA.org
> _______________________________________
>
> CONFIDENTIALITY:
> This e-mail and any attachments are
> confidential and may be privileged.
>
> Reproduction of this message, or any
> part of it, is not authorised without
> explicit permission of the sender.
>
> If you are not a named recipient,
> please notify the sender immediately
> and:
> 1- do not disclose the contents to
> another person;
> 2- do not use it for any
> purpose;
> 3- do not store or copy the information
> in any medium.