[Dip] solveRelaxed Method

[Matthew Galati]

Yes.

> -----Original Message-----
> From: Kipp Martin [mailto:kmartin at chicagobooth.edu]
> Sent: Thursday, September 09, 2010 10:26 AM
> To: Matthew Galati
> Cc: dip at list.coin-or.org
> Subject: Re: [Dip] solveRelaxed Method
> 
> Hi Matt:
> 
> > That is correct. The user also is responsible to setBlockId after creating the
> variable with var->setBlockId( ).
> > For some reason (I forget), that is not part of the constructor, so it has to
> be set afterwards.
> 
> Yes, I am setting that. In my code I have
> 
> var->setBlockId( whichBlock);
> 
> where whichBlock is the argument in
> 
> DecompSolverStatus OSDipApp::solveRelaxed(const int whichBlock,
> 		const double * redCostX, const double convexDual,
> 		list<DecompVar*> & vars)
> 
> This is the correct setting, right?
> 
> Thanks
> 
> >
> > I can probably make a version of DecompVar that just needs
> > solInd/solEls - since the framework can deduce the rest with a little work -
> which reduce the burden on the user.
> >
> > I plan to do that, when I have some free time, under this ticket:
> >    https://projects.coin-or.org/Dip/ticket/50
> >
> >> -----Original Message-----
> >> From: dip-bounces at list.coin-or.org
> >> [mailto:dip-bounces at list.coin-or.org]
> >> On Behalf Of Kipp Martin
> >> Sent: Thursday, September 09, 2010 5:56 AM
> >> To: dip at list.coin-or.org
> >> Subject: [Dip] solveRelaxed Method
> >>
> >> Hi:
> >>
> >> In the solveRelaxed() method one adds a Decomp var using the
> constructor:
> >>
> >> DecompVar(solInd, solEls, varRedCost - convexDual, varOrigCost);
> >>
> >> based on examining example code it seems like:
> >>
> >> 1.) varRedCost should be the optimal objective function value of the
> >> subproblem corresponding to whichBlock using the cost vector redCostX
> >> for objective function coefficients, and
> >>
> >> 2.) varOrigCost should be the value you get by taking the optimal
> >> solution to the subproblem corresponding to whichBlock using the cost
> >> vector redCostX for objective function coefficients but then
> >> evaluating the objective function based on the original cost vector.
> >>
> >> Is this correct?
> >>
> >> Thanks
> >>
> >>
> >>
> >> --
> >> Kipp Martin
> >> Professor of Operations Research
> >> and Computing Technology
> >> Booth School of Business
> >> University of Chicago
> >> 5807 South Woodlawn Avenue
> >> Chicago, IL 60637
> >> 773-702-7456
> >> kmartin at chicagobooth.edu
> >> http://www.chicagobooth.edu/faculty/bio.aspx?person_id=12825325568
> >> http://projects.coin-or.org/OS
> >>
> >> _______________________________________________
> >> Dip mailing list
> >> Dip at list.coin-or.org
> >> http://list.coin-or.org/mailman/listinfo/dip
> >
> 
> 
> --
> Kipp Martin
> Professor of Operations Research
> and Computing Technology
> Booth School of Business
> University of Chicago
> 5807 South Woodlawn Avenue
> Chicago, IL 60637
> 773-702-7456
> kmartin at chicagobooth.edu
> http://www.chicagobooth.edu/faculty/bio.aspx?person_id=12825325568
> http://projects.coin-or.org/OS