Incorrect solution of a conic QCP
Awaiting user input
For the following conic quadratic model:
Minimize
0.375 x + 5e+06 y
Subject To
q: [ x * y ] >= 1
Bounds
End
Gurobi 10.0.2 (Linux x86_64) returns an incorrect solution:
x = 3651.48
y = 3.29228e-08
x*y = 0.000120217
0
-
Hi,
I cannot reproduce this so far. Could you post a complete log? Are you using any non-default parameter settings?
Silke
0 -
Hi,
here is the log and solution file when run on M1 with default settings from LP file:
$gurobi_cl ResultFile=socp_09.sol socp_09.lp
Set parameter Username
Set parameter LogFile to value "gurobi.log"
Academic license - for non-commercial use only - expires 2024-01-06
Using license file /Users/bg307/gurobi.lic
Gurobi Optimizer version 10.0.2 build v10.0.2rc0 (mac64[arm])
Copyright (c) 2023, Gurobi Optimization, LLC
Read LP format model from file socp_09.lp
Reading time = 0.00 seconds
: 0 rows, 2 columns, 0 nonzeros
CPU model: Apple M1 Pro
Thread count: 8 physical cores, 8 logical processors, using up to 8 threads
Optimize a model with 0 rows, 2 columns and 0 nonzeros
Model fingerprint: 0x14e9b579
Model has 1 quadratic constraint
Coefficient statistics:
Matrix range [0e+00, 0e+00]
QMatrix range [1e+00, 1e+00]
Objective range [4e-01, 5e+06]
Bounds range [0e+00, 0e+00]
RHS range [0e+00, 0e+00]
QRHS range [1e+00, 1e+00]
Presolve time: 0.00s
Presolved: 3 rows, 5 columns, 7 nonzeros
Presolved model has 1 second-order cone constraint
Ordering time: 0.00s
Barrier statistics:
AA' NZ : 3.000e+00
Factor NZ : 6.000e+00
Factor Ops : 1.400e+01 (less than 1 second per iteration)
Threads : 1
Objective Residual
Iter Primal Dual Primal Dual Compl Time
0 2.24972646e+07 0.00000000e+00 2.04e+00 7.57e+04 3.82e+06 0s
1 1.35934623e+06 3.49014663e+05 4.73e-02 4.66e+04 2.01e+05 0s
2 5.54359788e+05 4.73152089e+05 5.21e-08 3.12e+04 5.96e+04 0s
3 5.09022687e+05 1.43919195e+05 1.77e-12 2.27e+03 7.88e+04 0s
4 5.67724773e+04 3.34171094e+04 8.81e-13 4.55e+02 1.00e+04 0s
5 3.10166010e+04 3.83782321e+04 2.31e-09 1.61e+02 1.62e+03 0s
6 2.00097928e+04 2.78064479e+04 1.85e-09 7.67e+01 6.01e+02 0s
7 1.55891507e+04 2.14983360e+04 7.45e-09 4.40e+01 4.11e+02 0s
8 1.22909382e+04 1.76224137e+04 1.34e-08 2.96e+01 2.98e+02 0s
9 1.07575734e+04 1.49199117e+04 1.07e-07 2.24e+01 3.64e+02 0s
10 9.06579188e+03 1.08815350e+04 4.53e-07 1.36e+01 7.14e+02 0s
11 4.43182866e+03 6.30544512e+03 4.36e-07 3.68e+00 1.92e+02 0s
12 3.38529420e+03 4.57435023e+03 8.80e-06 1.29e+00 2.67e+01 0s
13 3.25035365e+03 3.92708258e+03 3.22e-05 7.39e-01 1.58e+01 0s
14 3.02760435e+03 3.09791519e+03 1.54e-05 1.99e-01 2.68e+01 0s
15 2.79219996e+03 2.43202351e+03 1.49e-06 5.16e-01 7.11e+01 0s
16 1.17428611e+03 2.72118979e+03 3.21e-04 6.79e-02 8.09e+00 0s
17 1.84092937e+03 2.21974423e+03 1.81e-04 4.03e-02 4.41e+01 0s
18 1.13569150e+03 2.22998944e+03 3.31e-04 6.26e-03 5.19e-01 0s
19 1.11603865e+03 2.23019322e+03 3.36e-04 2.26e-02 4.68e-03 0s
20 1.11551530e+03 2.23031637e+03 6.88e-02 4.40e+01 5.00e-05 0s
21 1.11551070e+03 2.23031637e+03 1.41e-01 7.37e+01 4.94e-05 0s
22 1.11550088e+03 2.23031637e+03 2.33e-01 1.53e+02 4.83e-05 0s
23 1.11534095e+03 2.23031637e+03 1.23e-01 3.16e+02 3.76e-05 0s
24 1.11516642e+03 2.23031641e+03 5.28e-04 1.52e+02 1.89e-06 0s
25 1.11516642e+03 2.23031641e+03 9.49e-04 1.19e+01 1.89e-06 0s
26 1.11516618e+03 2.23031641e+03 1.76e-03 9.41e-01 1.91e-06 0s
27 1.11516618e+03 2.23031641e+03 3.53e-03 8.26e+00 1.91e-06 0s
28 1.11516617e+03 2.23031641e+03 4.45e-04 1.72e+01 1.91e-06 0s
29 1.11516617e+03 2.23031641e+03 7.94e-04 3.57e+01 1.91e-06 0s
30 1.11516617e+03 2.23031641e+03 1.50e-03 7.41e+01 1.91e-06 0s
31 1.11516617e+03 2.23031641e+03 7.45e-04 1.39e+02 1.91e-06 0s
32 1.11516617e+03 2.23031641e+03 2.40e-04 1.16e+01 1.90e-06 0s
33 1.11516616e+03 2.23031641e+03 4.37e-04 2.41e+01 1.90e-06 0s
34 1.11516615e+03 2.23031641e+03 2.42e-04 4.99e+01 1.90e-06 0s
35 1.11516615e+03 2.23031641e+03 3.76e-04 1.10e+00 1.90e-06 0s
36 1.11516615e+03 2.23031641e+03 2.49e-04 7.58e+00 1.90e-06 0s
37 1.11516615e+03 2.23031641e+03 4.61e-04 1.35e+01 1.90e-06 0s
38 1.11516615e+03 2.23031641e+03 2.25e-04 2.83e+01 1.90e-06 0s
39 1.11516615e+03 2.23031641e+03 4.16e-04 5.79e+01 1.86e-06 0s
40 1.11516615e+03 2.23031641e+03 7.41e-04 1.19e+02 1.86e-06 0s
41 1.11516615e+03 2.23031641e+03 3.20e-04 2.54e+02 1.86e-06 0s
42 1.11516615e+03 2.23031641e+03 2.61e-04 1.95e+02 1.86e-06 0s
43 1.11516615e+03 2.23031641e+03 6.23e-04 1.32e+01 1.86e-06 0s
44 1.11516615e+03 2.23031641e+03 1.06e-03 3.77e-01 1.86e-06 0s
Barrier solved model in 44 iterations and 0.00 seconds (0.00 work units)
Optimal objective 1.11516615e+03Solution file:
# Objective value = 1.1151661531133648e+03
x 2.9737764067526346e+03
y 1.1622535448529273e-13Is the objective coefficient 5e+06 so criminal?
0 -
Thank you! I can reproduce this using gurobi_cl.
We recommend that objective coefficients be no bigger than 1e4, but this violation is still surprising to me.
0
Please sign in to leave a comment.
Comments
3 comments