[Dip] Purpose of m_auxSI

[Matthew Galati]

It was just an option I was playing around with in an attempt to try to
prune earlier. I did not have time to research it much - but at first try,
it didn't seem to help at all. And, clearly in cases where the compact
formulation is very big - this is a bad idea.

   //solve compact formulation first before starting PhaseI
   //  hopefully identify infeasibiity in tree quicker
   int    InitCompactSolve;


   //---
   //--- construct auxillary compact lp interface
   //---
   if(m_param.InitCompactSolve){
      //TODO: would be nice if we could utilize IP presolve here?
      m_auxSI = new OsiLpSolverInterface();
      assert(m_auxSI);
      loadSIFromModel(m_auxSI);
   }


It should be off by default. If it is not, please let me know.

Matt




On Sat, Feb 25, 2012 at 5:13 PM, Siddhartha Jain <sj1 at cs.cmu.edu> wrote:

> Hi,
>
> In the phaseInit method of DecompAlgoPC, the LP m_auxSI is being solved.
> After doing some digging, it seems like m_auxSI represents the original
> undecomposed problem. Is this true?
>
> If so, then why is the original undecomposed problem being solved in an
> algorithm meant to take advantage of decomposition?
>
> Thanks,
> --Sid
>
> _______________________________________________
> Dip mailing list
> Dip at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/dip
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/dip/attachments/20120225/4bef2ee1/attachment.html>