R help - May 2005 - constraints

Is there a package in R that handles general linear (in-)equality + box
constrained optimization?
If it is not there, could anyone advise me which way to go?
And/or point me to packages that solve these problems partially?
best, ingmar
-- 
Ingmar Visser
Department of Psychology, University of Amsterdam
Roetersstraat 15, 1018 WB Amsterdam
The Netherlands
http://users.fmg.uva.nl/ivisser/
tel: +31-20-5256735

On Sunday 22 May 2005 03:18, Ingmar Visser wrote:
> Is there a package in R that handles general linear (in-)equality + box
> constrained optimization?
R> help.search("constrained optimisation")
constrOptim(stats)      Linearly constrained optimisation

Best wishes,
Arne
> If it is not there, could anyone advise me which way to go?
> And/or point me to packages that solve these problems partially?
> best, ingmar
-- 
Arne Henningsen
Department of Agricultural Economics
University of Kiel
Olshausenstr. 40
D-24098 Kiel (Germany)
Tel: +49-431-880 4445
Fax: +49-431-880 1397
ahenningsen at agric-econ.uni-kiel.de
http://www.uni-kiel.de/agrarpol/ahenningsen/

On Sunday 22 May 2005 21:32, you wrote:
> Hi Arne,
> Thanks. Unfortunately constrOptim does not handle combined equality and
> inequality constraints, or perhaps not in a way that I understand. Do you
> have suggestions as to how to handle such combinations of constraints?
> best, ingmar
You can impose the equality constraints directly by parameter substitution. 
See
   https://stat.ethz.ch/pipermail/r-help/2005-May/070481.html
and the 3 answers.
Then you can impose the inequality constraints in constrOptim.
I don't know if it possible to impose an equality constraint, say a = b, in 
constrOptim by two inequality constraints:
   a - b >= 0
and
   b - a >= 0
However, I guess that the solver prefers the first way ;-)

Best wishes,
Arne
> On 5/22/05 3:28 PM, "Arne Henningsen" <ahenningsen at
agric-econ.uni-kiel.de>
>
> wrote:
> > On Sunday 22 May 2005 03:18, Ingmar Visser wrote:
> >> Is there a package in R that handles general linear (in-)equality
+ box
> >> constrained optimization?
> >
> > R> help.search("constrained optimisation")
> > constrOptim(stats)      Linearly constrained optimisation
> >
> > Best wishes,
> > Arne
> >
> >> If it is not there, could anyone advise me which way to go?
> >> And/or point me to packages that solve these problems partially?
> >> best, ingmar
-- 
Arne Henningsen
Department of Agricultural Economics
University of Kiel
Olshausenstr. 40
D-24098 Kiel (Germany)
Tel: +49-431-880 4445
Fax: +49-431-880 1397
ahenningsen at agric-econ.uni-kiel.de
http://www.uni-kiel.de/agrarpol/ahenningsen/

On Sun, 22 May 2005, Arne Henningsen wrote:
> constrOptim by two inequality constraints:
>   a - b >= 0
> and
>   b - a >= 0
> However, I guess that the solver prefers the first way ;-)
It is not possible.  The starting value must be in the interior of the 
feasible region (because the constraints are implemented with log 
barriers).  This approach gives a feasible region with no interior.

 	-thomas