[Dip] SubProblem Infeasibility

[Matthew Galati]

isCutoff just checks if it was solved to optimally or not. 

If a subproblem is found infeasible then the original problem (or node) must be infeasible. There should be a check for that. 

I'll need to check the code tomorrow. 

Sent from my iPhone

On Apr 9, 2012, at 6:47 PM, Siddhartha Jain <sj1 at cs.cmu.edu> wrote:

> To be more precise,
> 
> After solveRelaxed is called on a subproblem, there's an if condition
> 
> if (solveResult.m_isCutoff)
>   mostNegRCvec[b] = min(mostNegRCvec[b],0.0);
> 
> From what I understand if this condition is true, then that means that a subproblem was infeasible (as the dual problem was unbounded). Is this true? If so, why does the loop keep exploring other subproblems rather than returning? If this is not true, then where does the infeasibility check happen?
> 
> Thanks
> 
> On Mon, Apr 9, 2012 at 12:31 AM, Siddhartha Jain <sj1 at cs.cmu.edu> wrote:
> Hi,
> 
> As far as I can tell, the generateVarsFea method is used to generate the columns from the subproblems. It calls the solveRelaxed method on the subproblems to generate the columns. But what happens if one of the subproblems is infeasible? There doesn't seem to be a check which breaks the process of generating columns if a subproblem turns out to be infeasible.
> 
> 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/20120409/e8562044/attachment.html>