Re: [Help-glpk] Re: XyPron: shadow prices in MIP problem

[glpk xypron]

Hello Ali,

See as example function glp_mpl_postsolve() in
http://glpk.dyndns.org/viewvc/svn/glpk/branches/glpk-4.38-dot/src/glpapi16.c?revision=1&view=markup

Best regards

Xypron
-------- Original-Nachricht --------
> Datum: Fri, 11 Jun 2010 15:51:52 +0200
> Betreff: Re: [Help-glpk] Re: XyPron: shadow prices in MIP problem

> No, it is not.
> 
> Could you please send us your code so we have the chance to help?
> 
> Ali
> 
> On Fri, Jun 11, 2010 at 10:41 AM, Fabio <address@hidden> wrote:
> > Is it normal that solving the problem again with fixed integer variables
> at
> > the optimal values takes nearly as long as solving the original problem?
> > (76sec vs. 72sec)
> >
> > The problem has:
> > 34536 rows, 34512 columns, 86232 non-zeros
> > 17256 integer variables, all of which are binary
> > On May 31, 2010, at 11:33 AM, Andrew Makhorin <address@hidden> wrote:
> >
> >> Just a follow-up question to the original question (dual values for
> >> the continious values in a mixed integer problem):
> >
> >> So I basically have to solve the problem again but with fixed
> >> integer values taken from the MIP solution. For the java interface
> >> this means:
> >> - retrieve the solutions of the integer variables
> >> - delete old problem (MIP)
> >> - create new problem (LP, old integer vars are fixed)
> >> - solve again
> >
> >> Or is there a better way?
> >
> > You don't need to destroy the old problem object. Just fix integer
> > variables at the optimal values found by the mip solver and call the
> > simplex solver.
> >
> >
> > _______________________________________________
> > Help-glpk mailing list
> > address@hidden
> > http://lists.gnu.org/mailman/listinfo/help-glpk
> >
> >
> 
> _______________________________________________
> Help-glpk mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/help-glpk

-- 
GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT!
Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01