[R] non linear estimation

[Rubén Roa]

> -----Mensaje original-----
> De: r-help-bounces at r-project.org 
> [mailto:r-help-bounces at r-project.org] En nombre de JamesHuang
> Enviado el: jueves, 29 de abril de 2010 3:38
> Para: r-help at r-project.org
> Asunto: Re: [R] non linear estimation
> 
> 
> any suggestion? actually I just wanna know if there is a 
> package for non linear estimation with restriction, thanks. I 
> am a new for R....

I do not know if there is any specific package for optimization with restrictions, but you can use optim with method="L-BFGS-B"
This only lets you set bounds of single parameters, so for restrictions such as a+b<19 in
Y=a+(b+c*x)*exp(-d*x)
you could deduce your restrictions in terms of single parameters (for example, in your original mail you put that a>10, a+b<19, and b<3, so the restriction a+b>19 is actually redundant), or else you could think of some re-parameterization that would put a+b (and all other multi-par restrictions) as a single parameter.

Wait, is this a homework?

____________________________________________________________________________________ 

Dr. Rubén Roa-Ureta
AZTI - Tecnalia / Marine Research Unit
Txatxarramendi Ugartea z/g
48395 Sukarrieta (Bizkaia)
SPAIN