-------- Forwarded Message --------
From: [email protected]
To: [email protected]
Subject: problem with fpump option
Date: Thu, 17 Mar 2016 16:11:45 +0000

Hi,

 

I try to tune parameters of glpsol.exe for improving optimization
performance.

I noticed that option fpump often gave a huge boost by finding quickly a
good solution.

My problem is that for some data, glpsol.exe met an unexpected error
just before leaving.

 

I am using Glpk 4.59 x86 on Windows 7 (x64). I compiled it with Visual
Studio 2010.

The command line I use is the following one : glpsol.exe -m Modele.mod
-d Modele.dat -o solution.txt --presol --fpump --gomory --cover –clique

The end of the log is :

Model has been successfully generated

GLPK Integer Optimizer, v4.59

2245 rows, 1466 columns, 5164 non-zeros

301 integer variables, all of which are binary

Preprocessing...

846 constraint coefficient(s) were reduced

984 rows, 779 columns, 2856 non-zeros

268 integer variables, all of which are binary

Scaling...

A: min|aij| = 1.000e-003  max|aij| = 1.000e+004  ratio = 1.000e+007

GM: min|aij| = 2.563e-001  max|aij| = 3.901e+000  ratio = 1.522e+001

EQ: min|aij| = 6.770e-002  max|aij| = 1.000e+000  ratio = 1.477e+001

2N: min|aij| = 5.674e-002  max|aij| = 1.653e+000  ratio = 2.913e+001

Constructing initial basis...

Size of triangular part is 960

Solving LP relaxation...

GLPK Simplex Optimizer, v4.59

984 rows, 779 columns, 2856 non-zeros

      0: obj = -3.774482640e+010 inf =  2.812e+005 (310)

    288: obj = -4.407173251e+010 inf =  1.282e-011 (0) 1

*   457: obj = -7.671615844e+010 inf =  1.612e-011 (0) 2

OPTIMAL LP SOLUTION FOUND

Integer optimization begins...

Gomory's cuts enabled

Cover cuts enabled

Clique cuts enabled

Constructing conflict graph...

Conflict graph has 1 + 1 = 2 vertices

+   457: mip =     not found yet >=              -inf        (1; 0)

Applying FPUMP heuristic...

Pass 1

Solution found by heuristic: -71614455133.5

Pass 1

Error: basis matrix is singular to working precision (cond = 1.12e+016)

Warning: glp_simplex returned 5

Cuts on level 0: gmi = 2;

Cuts on level 45: gmi = 2;

+   557: >>>>> -7.328310401e+010 >= -7.347180970e+010   0.3% (46; 0)

Cuts on level 42: gmi = 2;

+   695: >>>>> -7.338423438e+010 >= -7.340799383e+010 < 0.1% (51; 70)

Cuts on level 36: gmi = 2;

+   757: >>>>> -7.338510397e+010 >= -7.340756805e+010 < 0.1% (32; 157)

+   918: mip = -7.338510397e+010 >=     tree is empty   0.0% (0; 251)

INTEGER OPTIMAL SOLUTION FOUND

Time used:   0.7 secs

Memory used: 3.6 Mb (3797151 bytes)

Model has been successfully processed

Writing MIP solution to 'solution.txt'...

Error: 1 memory block(s) were lost

Error detected in file ..\examples\glpsol.c at line 1494

 

In fact, the optimization problem has been solved and the solution file
has been written.

So, the bug occurs just before leaving and I obtain a pop-up window
allowing to debug or close the program.

 

I know that if I delete fpump option from command line, it works.

But for some other data, I need to find quickly (1 min) a solution and
without fpump, no solution has been obtained before time limit.

 

If anyone has any idea, it would be great !

 

Nicolas

 



ENGIE Mail Disclaimer:
http://www.engie.com/disclaimer/disclaimer-fr.html


GLPSOL: GLPK LP/MIP Solver, v4.59
Parameter(s) specified in the command line:
 -m Modele.mod -d Modele.dat -o solution.txt --presol --fpump --gomory --cover
 --clique
Reading model section from Modele.mod...
106 lines were read
Reading data section from Modele.dat...
211 lines were read
Generating C...
Generating Ct1...
Generating Ct2...
Generating Ct3...
Generating Ct4...
Generating Ct5...
Generating Ct6...
Generating Ct7...
Generating Ct8...
Generating Ct9...
Generating Ct10...
Generating Ct11...
Generating Ct12...
Generating Ct13...
Generating Ct14...
Generating Ct15...
Generating Ct16...
Generating Ct17...
Generating Ct18...
Generating Ct19...
Generating Ct20...
Generating Ct21...
Generating Ct22...
Generating Ct23...
Generating Ct24...
Model has been successfully generated
GLPK Integer Optimizer, v4.59
2245 rows, 1466 columns, 5164 non-zeros
301 integer variables, all of which are binary
Preprocessing...
846 constraint coefficient(s) were reduced
984 rows, 779 columns, 2856 non-zeros
268 integer variables, all of which are binary
Scaling...
 A: min|aij| = 1.000e-003  max|aij| = 1.000e+004  ratio = 1.000e+007
GM: min|aij| = 2.563e-001  max|aij| = 3.901e+000  ratio = 1.522e+001
EQ: min|aij| = 6.770e-002  max|aij| = 1.000e+000  ratio = 1.477e+001
2N: min|aij| = 5.674e-002  max|aij| = 1.653e+000  ratio = 2.913e+001
Constructing initial basis...
Size of triangular part is 960
Solving LP relaxation...
GLPK Simplex Optimizer, v4.59
984 rows, 779 columns, 2856 non-zeros
      0: obj = -3.774482640e+010 inf =  2.812e+005 (310)
    288: obj = -4.407173251e+010 inf =  1.282e-011 (0) 1
*   457: obj = -7.671615844e+010 inf =  1.612e-011 (0) 2
OPTIMAL LP SOLUTION FOUND
Integer optimization begins...
Gomory's cuts enabled
Cover cuts enabled
Clique cuts enabled
Constructing conflict graph...
Conflict graph has 1 + 1 = 2 vertices
+   457: mip =     not found yet >=              -inf        (1; 0)
Applying FPUMP heuristic...
Pass 1
Solution found by heuristic: -71614455133.5
Pass 1
Error: basis matrix is singular to working precision (cond = 1.12e+016)
Warning: glp_simplex returned 5
Cuts on level 0: gmi = 2;
Cuts on level 45: gmi = 2;
+   557: >>>>> -7.328310401e+010 >= -7.347180970e+010   0.3% (46; 0)
Cuts on level 42: gmi = 2;
+   695: >>>>> -7.338423438e+010 >= -7.340799383e+010 < 0.1% (51; 70)
Cuts on level 36: gmi = 2;
+   757: >>>>> -7.338510397e+010 >= -7.340756805e+010 < 0.1% (32; 157)
+   918: mip = -7.338510397e+010 >=     tree is empty   0.0% (0; 251)
INTEGER OPTIMAL SOLUTION FOUND
Time used:   0.7 secs
Memory used: 3.6 Mb (3797151 bytes)
Model has been successfully processed
Writing MIP solution to 'solution.txt'...
Error: 1 memory block(s) were lost
Error detected in file ..\examples\glpsol.c at line 1494

Attachment: Modele.dat
Description: Modele.dat

Attachment: Modele.mod
Description: Modele.mod

_______________________________________________
Help-glpk mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/help-glpk

Reply via email to