[Help-glpk] weird problem and patches (glpsol&limits)

[Livio Bertacco]

Hello,

I've attached a couple of patches and an lp problem on which glpk seems to
have some difficulties.

The "difficulties" on strano.lp are that I get inconsistent results
depending on compiler and options. I'm running glpsol on a Windows XP
machine with intel Pentium-M cpu. I'm using two versions of glpsol, one
(glpsolg) compiled with gcc 3.3.1 and the other (glpsolm) compiled with
VisualC++v7.1.

I get the following results:
glpsolg --nomip enters endless loop.
glpsolg --nomip --nopresol says infeasible
glpsolm --nomip finds OPTIMAL solution
glpsolm --nomip --nopresol says infeasible
So these are very inconsistent...(BTW cplex finds optimal solution)

Therefore it seems that GCC and MSVC handle floating point operations
slightly differently. Does anyone have any insight on this?
(I used just option -O2 for gcc)

The patches are:

1) It_and_tm_lim.patch
This fixes the handling of time and iteration limits in the mip solver.
Previously the mip solver didn't set any iteration/time limit on subproblems
(nodes) so the solution of the subproblem could take longer(more iterations)
than limits allowed. Even worse, if the subproblem didn't terminate (e.g.
for numerical instability) than the mip itself could run forever even with a
tm_lim or it_lim set. 

2) glpsol.patch
This adds the following options to the glpsol program:
a) --wbas and --rbas, read and write a warmstart basis
b) --itlim: set an iteration limit
c) -p or --param: set any other GLPK parameter (not already covered by
specific options)

Finally, what's the meaning of lines 410-411 of glpmip1.c:
>/* delay output from the simplex method for first 10 seconds */
>lpx_set_real_param(lp, LPX_K_OUTDLY, 10.0);
I mean, why would you want to hide what's going on for 10 sec?

Thank you,
Livio Bertacco
PhD Candidate
University of Padua

patches.tar.gz
Description: Binary data

strano.lp.gz
Description: Binary data