Re: [Bug-glpk] [Help-glpk] --wfreemps

[Andrew Makhorin]

> the problem seems not to be writing of the mps file, but the reading.
> 
> The integers are interpreted as binaries.
> 

Correct. Glp_write_mps has a bug: it does not write 'PL' to the BOUNDS
section for integer non-negative (i.e. having no upper bound) variables,
so glp_read_mps reads such variables as binary.

Raniere, thank you for your report.

You can fix the bug by changing the fragment (file glpmps.c, lines
1341-1344):

bnds: /* write BOUNDS section */
      for (j = P->n; j >= 1; j--)
         if (!(P->col[j]->type == GLP_LO && P->col[j]->lb == 0.0))
            break;

to the following one:

bnds: /* write BOUNDS section */
      for (j = P->n; j >= 1; j--)
         if (!(P->col[j]->kind == GLP_CV && 
               P->col[j]->type == GLP_LO && P->col[j]->lb == 0.0))
            break;

> -------- Original-Nachricht --------
> > Datum: Mon, 17 Oct 2011 23:37:48 -0200
> > Betreff: [Help-glpk] --wfreemps
> 
> > Hi,
> > I trying to convert one problem on gmpl format to free mps format, both in
> > attached, and the solve it.
> > Unfortunately, when solve the mps file I get an answer different from the
> > answer that I get when solve using the gmpl file.
> > 
> > I was using the following commands:
> > $ glpsol --math exemp01.mod
> > $ glpsol --math exemp01.mod --check --wfreemps exemp01.mps
> > $ glpsol --freemps exemp01.mps
> > 
> > 
> > Raniere Gaia Costa da Silva
>