[R] Method of L-BFGS-B of optim evaluate function outside of box constraints

[Adelchi Azzalini]

On Wednesday 20 August 2003 15:44, Shengqiao Li wrote:
> Hi, R guys:
>
> I'm using L-BFGS-B method of optim for minimization problem. My function
> called besselI function which need non-negative parameter and the besselI
> will overflow if the parameter is too large. So I set the constraint box
> which is reasonable for my problem. But the point outside the box was
> test, and I got error. My program and the error follows. This program
> depends on CircStats package.
>
>
> Anyone has any idea about this?
>

No idea... I can only say that a  similar behaviour of optim
with  method="L-BFGS-B" (i.e. evaluation of the function 
outside the lower/upper limits) has occurred to me too. 
It is a rare but possible behaviour.

Last June, I have sent a full description of the problem to 
r-bugs at r-project.org

Regards,

Adelchi Azzalini


-- 
Adelchi Azzalini  <azzalini at stat.unipd.it>
Dipart.Scienze Statistiche, Università di Padova, Italia
http://azzalini.stat.unipd.it/
(please, no ms-word/ms-excel/alike attachments)