[Scip] scip reports optimal solution along with violated constraints

[Martin Bergner]

Hi Tobias,

I am using SCIP in the version 1.2.0.8

> Is there anything strange in the SCIP output? For example, what does
> presolving say about the size of the model? What numbers appear in the
> node log, in particular the "rows" and 
> "cols" column? Is the model solved directly at the root node?
No, it looks good, here is an excerpt for subproblem 1 first from the
code, then from the command line:

----------8<----------8<----------8<----------8<---------- 
presolving (2 rounds):
 6 deleted vars, 0 deleted constraints, 0 tightened bounds, 0 added holes, 0 changed sides, 0 changed coefficients
 0 implications, 0 cliques
presolved problem has 51 variables (0 bin, 0 int, 0 impl, 51 cont) and 14 constraints

 time | node  | left  |LP iter|LP it/n| mem |mdpt |frac |vars |cons |cols |rows |cuts |confs|strbr|  dualbound   | primalbound  |  gap   
t 0.0s|     1 |     0 |     0 |     - | 187k|   0 |   - |  51 |  14 |   0 |   0 |   0 |   0 |   0 |      --      | 0.000000e+00 | 100.00%
  0.0s|     1 |     0 |     0 |     - | 187k|   0 |   - |  51 |  14 |   0 |   0 |   0 |   0 |   0 | 0.000000e+00 | 0.000000e+00 |   0.00%

SCIP Status        : problem is solved [optimal solution found]
Solving Time (sec) : 0.00
Solving Nodes      : 1
Primal Bound       : +0.00000000000000e+00 (3 solutions)
Dual Bound         : +0.00000000000000e+00
Gap                : 0.00 %
  [linear] <c593>: -1<x58>[C] -1<x66>[C] -1<x74>[C] -1<x82>[C] -1<x90>[C] -1<x98>[C] -1<x106>[C] +1<x514>[C] == 0;
violation: right hand side is violated by 74
best solution is not feasible in original problem
----------8<----------8<----------8<----------8<----------

Compare the command line with presolving turned off:
----------8<----------8<----------8<----------8<---------- 
SCIP> r prob_1.lp 
original problem has 57 variables (6 bin, 0 int, 0 impl, 51 cont) and 14 constraints
SCIP> o

presolving:
presolving (0 rounds):
 0 deleted vars, 0 deleted constraints, 0 tightened bounds, 0 added holes, 0 changed sides, 0 changed coefficients
 0 implications, 0 cliques
presolved problem has 57 variables (6 bin, 0 int, 0 impl, 51 cont) and 14 constraints
     14 constraints of type <linear>
Presolving Time: 0.00

 time | node  | left  |LP iter|LP it/n| mem |mdpt |frac |vars |cons |cols |rows |cuts |confs|strbr|  dualbound   | primalbound  |  gap   
  0.0s|     1 |     0 |    10 |     - | 221k|   0 |   4 |  57 |  14 |  57 |  14 |   0 |   0 |   0 | 4.625000e+00 |      --      |    Inf 
r 0.0s|     1 |     0 |    10 |     - | 222k|   0 |   - |  57 |  14 |  57 |  14 |   0 |   0 |   0 | 4.625000e+00 | 4.625000e+00 |   0.00%
  0.0s|     1 |     0 |    10 |     - | 222k|   0 |   - |  57 |  14 |  57 |  14 |   0 |   0 |   0 | 4.625000e+00 | 4.625000e+00 |   0.00%

SCIP Status        : problem is solved [optimal solution found]
Solving Time (sec) : 0.01
Solving Nodes      : 1
Primal Bound       : +4.62500000000000e+00 (1 solutions)
Dual Bound         : +4.62500000000000e+00
Gap                : 0.00 %
----------8<----------8<----------8<----------8<----------

The only differences are in the presolving. If I allow scip to presolve
from the command line, it eliminates all variables and constraints. The
objective function and solution, however, is correct either way from the
command line.

Regards, 
Martin