High Execution Time
Awaiting user input
Hello Gurobi Team,
I am trying to execute the following model but it is taking too much time to get a solution. I don't know if in this case is too big or the solution is unfeasible.
Is there any suggestion to relax the constraints or to tune any parameter inside the model?
Set parameter TimeLimit to value 3600
Set parameter FeasibilityTol to value 1e-07
Set parameter IntFeasTol to value 1e-07
Set parameter OptimalityTol to value 1e-07
Set parameter Presolve to value 1
Gurobi Optimizer version 11.0.0 build v11.0.0rc2 (mac64[arm] - Darwin 23.3.0 23D60)
CPU model: Apple M2
Thread count: 8 physical cores, 8 logical processors, using up to 8 threads
Optimize a model with 1075119 rows, 478908 columns and 21506181 nonzeros
Variable types: 79818 continuous, 399090 integer (0 binary)
Coefficient statistics:
Matrix range [1e+00, 2e+04]
Objective range [4e+00, 4e+00]
Bounds range [1e+00, 2e+04]
RHS range [1e+00, 2e+04]
Presolve removed 846138 rows and 377362 columns (presolve time = 5s) ...
Presolve removed 900840 rows and 401291 columns
Presolve time: 5.66s
Presolved: 174279 rows, 77617 columns, 1432424 nonzeros
Variable types: 9822 continuous, 67795 integer (64018 binary)
Deterministic concurrent LP optimizer: primal simplex, dual simplex, and barrier
Showing barrier log only...
Root barrier log...
Ordering time: 2.73s
Barrier statistics:
Dense cols : 872
AA' NZ : 4.443e+06
Factor NZ : 2.887e+07 (roughly 300 MB of memory)
Factor Ops : 5.506e+10 (less than 1 second per iteration)
Threads : 5
Objective Residual
Iter Primal Dual Primal Dual Compl Time
0 -1.46998918e+10 -6.16816563e+10 1.98e+05 1.14e+03 1.23e+07 27s
1 -5.88360457e+09 -4.61395062e+10 8.85e+04 2.30e+03 5.51e+06 28s
2 -2.07475440e+09 -2.91954604e+10 4.05e+04 1.00e+03 2.53e+06 28s
3 -6.82003772e+08 -1.73464959e+10 2.30e+04 4.30e+02 1.41e+06 29s
4 3.27184639e+08 -6.82964368e+09 1.03e+04 9.80e+01 6.19e+05 29s
5 6.58537990e+08 -3.32129537e+09 6.30e+03 4.36e+01 3.72e+05 30s
6 8.51860060e+08 -1.57644774e+09 3.95e+03 2.17e+01 2.31e+05 31s
7 9.54768872e+08 -6.17334483e+08 2.73e+03 1.08e+01 1.57e+05 31s
8 1.04524464e+09 -5.02439245e+07 1.66e+03 5.36e+00 9.56e+04 32s
9 1.12206942e+09 5.24145943e+08 6.56e+02 1.64e-02 3.89e+04 32s
10 1.14420938e+09 8.67667346e+08 1.27e+02 1.16e-10 8.70e+03 33s
11 1.12404035e+09 9.98863782e+08 2.14e+01 1.82e-10 2.07e+03 34s
12 1.12006347e+09 1.03094125e+09 1.20e+01 1.57e-10 1.31e+03 35s
13 1.11693142e+09 1.07211004e+09 2.09e+00 1.04e-10 5.24e+02 35s
14 1.11507946e+09 1.09767658e+09 5.16e-01 7.73e-11 1.77e+02 36s
15 1.11429197e+09 1.11153032e+09 9.71e-02 8.00e-11 2.80e+01 37s
16 1.11412334e+09 1.11369942e+09 2.71e-02 5.82e-11 4.69e+00 38s
17 1.11406926e+09 1.11400620e+09 1.53e-03 1.69e-10 5.92e-01 39s
18 1.11406663e+09 1.11406430e+09 7.27e-05 1.53e-10 2.26e-02 39s
19 1.11406648e+09 1.11406647e+09 2.02e-04 3.06e-10 2.74e-04 40s
20 1.11406647e+09 1.11406649e+09 4.35e-04 7.42e-10 5.43e-05 41s
21 1.11406647e+09 1.11406649e+09 4.25e-04 3.55e-10 2.80e-05 41s
22 1.11406647e+09 1.11406649e+09 4.35e-04 5.71e-10 1.27e-05 42s
23 1.11406647e+09 1.11406649e+09 9.62e-04 3.27e-10 1.23e-05 43s
24 1.11406647e+09 1.11406649e+09 1.72e-03 5.39e-10 1.09e-05 44s
25 1.11406647e+09 1.11406649e+09 1.63e-03 6.53e-10 3.77e-07 45s
26 1.11406647e+09 1.11406649e+09 8.62e-04 1.06e-09 1.15e-07 46s
27 1.11406647e+09 1.11406649e+09 5.64e-04 1.48e-09 1.19e-08 47s
28 1.11406647e+09 1.11406649e+09 3.95e-04 6.48e-10 2.63e-08 47s
29 1.11406647e+09 1.11406649e+09 1.38e-03 2.01e-09 4.97e-09 48s
30 1.11406647e+09 1.11406649e+09 1.64e-03 3.49e-10 4.27e-09 48s
31 1.11406647e+09 1.11406649e+09 1.61e-03 3.42e-10 4.03e-09 49s
32 1.11406647e+09 1.11406649e+09 2.33e-03 7.65e-09 2.89e-09 50s
33 1.11406647e+09 1.11406649e+09 2.06e-03 5.11e-09 2.53e-09 50s
34 1.11406647e+09 1.11406649e+09 7.74e-04 2.24e-09 3.62e-10 51s
35 1.11406647e+09 1.11406649e+09 1.16e-03 2.27e-09 1.98e-10 52s
36 1.11406648e+09 1.11406649e+09 7.19e-04 2.04e-09 1.36e-10 52s
37 1.11406648e+09 1.11406649e+09 2.55e-03 2.02e-09 1.34e-10 53s
38 1.11406648e+09 1.11406649e+09 3.38e-03 7.13e-10 1.26e-09 54s
39 1.11406648e+09 1.11406649e+09 2.08e-03 7.19e-10 4.36e-10 54s
40 1.11406647e+09 1.11406649e+09 2.65e-03 9.31e-10 8.16e-11 55s
41 1.11406647e+09 1.11406649e+09 1.95e-03 4.66e-10 6.31e-11 56s
42 1.11406648e+09 1.11406649e+09 1.66e-03 7.95e-10 5.97e-10 56s
43 1.11406647e+09 1.11406649e+09 1.03e-03 9.31e-10 2.92e-10 57s
44 1.11406647e+09 1.11406649e+09 1.07e-03 6.20e-10 2.78e-11 58s
Barrier performed 44 iterations in 57.98 seconds (101.47 work units)
Sub-optimal termination - objective 1.11406647e+09
Root crossover log...
9597 DPushes remaining with DInf 0.0000000e+00 58s
0 DPushes remaining with DInf 8.1590755e-06 58s
13549 PPushes remaining with PInf 6.1511943e-03 58s
4023 PPushes remaining with PInf 6.0980772e-03 60s
0 PPushes remaining with PInf 6.0979840e-03 63s
Push phase complete: Pinf 6.0979840e-03, Dinf 1.3641931e-03 63s
Root simplex log...
Iteration Objective Primal Inf. Dual Inf. Time
20602 1.1140665e+09 0.000000e+00 1.359488e-03 63s
20705 1.1140665e+09 0.000000e+00 0.000000e+00 63s
20706 1.1140665e+09 0.000000e+00 0.000000e+00 63s
Extra simplex iterations after uncrush: 1
Concurrent spin time: 24.90s (can be avoided by choosing Method=3)
Solved with barrier
Root relaxation: objective 1.114067e+09, 20706 iterations, 65.65 seconds (78.45 work units)
Total elapsed time = 88.32s (DegenMoves)
Total elapsed time = 693.57s (DegenMoves)
Nodes | Current Node | Objective Bounds | Work
Expl Unexpl | Obj Depth IntInf | Incumbent BestBd Gap | It/Node Time
0 0 1.1141e+09 0 10672 - 1.1141e+09 - - 696s
0 0 1.1141e+09 0 10550 - 1.1141e+09 - - 719s
0 0 1.1141e+09 0 10860 - 1.1141e+09 - - 808s
0 0 1.1141e+09 0 10706 - 1.1141e+09 - - 1078s
-
Hi Carlos,
It's not uncommon for models with many integer variables (which are not binary variables in disguise) to struggle with finding solutions. What are the lower and upper bounds on your integer variables?
As a first step can you try running your model with our No Relaxation (NoRel) heuristic? Set NoRelHeurTime to an hour or two (perhaps longer if you are able to), and see how it goes. How successful it is (if at all) can determine next steps. When you do this also write out a mst file as it will probably be useful later to store any solution that NoRel finds.
- Riley
0 -
Thanks Riley!
I have 6 BoolVar and 2 NumVar.
The NumVar (both) have a [0,20000] bound.
So did what you suggested and got the following results after a 3-hour run.
Nodes | Current Node | Objective Bounds | Work
Expl Unexpl | Obj Depth IntInf | Incumbent BestBd Gap | It/Node Time
0 0 1.1141e+09 0 10538 - 1.1141e+09 - - 1476s
0 0 1.1141e+09 0 10566 - 1.1141e+09 - - 1527s
0 0 1.1141e+09 0 10401 - 1.1141e+09 - - 1694s
0 0 1.1141e+09 0 9851 - 1.1141e+09 - - 1745s
0 0 1.1143e+09 0 11528 - 1.1143e+09 - - 3063s
0 0 1.1143e+09 0 10205 - 1.1143e+09 - - 3128s
0 0 1.1145e+09 0 9990 - 1.1145e+09 - - 3147s
0 0 1.1145e+09 0 9292 - 1.1145e+09 - - 3254s
0 0 1.1145e+09 0 10241 - 1.1145e+09 - - 3347s
0 0 1.1145e+09 0 10156 - 1.1145e+09 - - 3395s
0 0 1.1145e+09 0 9859 - 1.1145e+09 - - 3506s
0 0 1.1145e+09 0 9830 - 1.1145e+09 - - 3554s
0 0 1.1145e+09 0 10019 - 1.1145e+09 - - 3661s
0 0 1.1145e+09 0 10384 - 1.1145e+09 - - 3777s
0 0 1.1149e+09 0 8820 - 1.1149e+09 - - 4532s
0 0 1.1149e+09 0 9511 - 1.1149e+09 - - 4571s
0 0 1.1178e+09 0 9998 - 1.1178e+09 - - 5363s
0 0 - 0 - 1.1178e+09 - - 10800s
Cutting planes:
Gomory: 3
Cover: 4
Implied bound: 1167
Clique: 585
MIR: 186
Flow cover: 224
Zero half: 43
RLT: 146
Relax-and-lift: 69
BQP: 29
PSD: 1
Explored 1 nodes (3692511 simplex iterations) in 10800.04 seconds (22706.96 work units)
Thread count was 8 (of 8 available processors)
Solution count 0
Time limit reached
Best objective -, best bound 1.117773368396e+09, gap -0 -
Hi Carlos,
I have 6 BoolVar and 2 NumVar
I'm not sure what this means. These are not Gurobi terms.
How did the NoRel heuristic perform? That's the part of the log that's interesting at the moment.
- Riley
0
Please sign in to leave a comment.
Comments
3 comments