[R] Solving an optimization problem: selecting an "optimal" subset

[Hans W Borchers]

Dimitri Shvorob wrote:
> 
> Same request to Hans:
> I am afraid I need a little more spoon-feeding following 
> 
>> I sent a GAMS script modeling this problem to the NEOS solvers
> 
> Thanks a lot!
> 

If you have access to CPLEX (I mean the commercial program, not Rcplex
which is just an interface to it), then I would suggest to follow Erwin's
proposal.
My GAMS file looks similar to the one Erwin posted. For instructions how to
send it to the NEOS solvers, see <http://neos.mcs.anl.gov/neos/solvers/>.

To replace an absolute value by two binary variables is an old trick in
optimization modeling. Here it will transform a subset sum problem into an
MILP task, solvable in R with lpSolve, Rglpk or Rsymphony.
But with a 100 binary variables it may be too much for these packages.
I will give it a try today.

Hans Werner

-- 
View this message in context: http://n4.nabble.com/Solving-an-optimization-problem-selecting-an-optimal-subset-tp1446084p1457982.html
Sent from the R help mailing list archive at Nabble.com.