[Cbc] Incorrect status in solution file for an integer infeasible problem on releases/2.8.10 r2045 and stable/2.8 r2056

[John Forrest]

Problem found to be infeasible in slightly different place - caused problem.

Fixed in stable/2.8

John Forrest


On 18/08/14 20:21, Rémy Roy wrote:
> Hello,
>
> I'm using the PuLP library to run Cbc. On my latest installation, I've 
> decided to try Cbc 2.8.10. Running the PuLP test suite, I've noticed 
> that the pulpTest060 test failed on both the 2.8.10 release and the 
> current stable/2.8 version.
>
> The mps file that results from the pulpTest060 test is:
>
> *SENSE:Minimize
> NAME          MODEL
> ROWS
>  N  OBJ
>  L  C0000000
>  G  C0000001
>  E  C0000002
> COLUMNS
>     X0000000  OBJ        1.00000e+00
>     MARK      'MARKER'                 'INTORG'
>     X0000001  C0000001   1.00000e+00
>     X0000001  C0000000   1.00000e+00
>     MARK      'MARKER'                 'INTEND'
>     MARK      'MARKER'                 'INTORG'
>     X0000002  C0000000   1.00000e+00
>     X0000002  C0000002  -1.00000e+00
>     MARK      'MARKER'                 'INTEND'
>     MARK      'MARKER'                 'INTORG'
>     X0000003  C0000001   1.00000e+00
>     X0000003  C0000002   1.00000e+00
>     MARK      'MARKER'                 'INTEND'
> RHS
>     RHS       C0000000   5.20000e+00
>     RHS       C0000001   1.03000e+01
>     RHS       C0000002   7.40000e+00
> BOUNDS
>  FX BND       X0000000   0.00000e+00
>  UP BND       X0000001   4.00000e+00
>  LO BND       X0000002  -1.00000e+00
>  UP BND       X0000002   1.00000e+00
>  UP BND       X0000003   1.00000e+01
> ENDATA
>
> The command line executed for running this test is:
>
> cbc /tmp/27068-pulp.mps branch printingOptions rows solution 
> /tmp/27068-pulp.sol
>
> The stdout for this problem and releases/2.8.10 r2045 is:
>
> Welcome to the CBC MILP Solver
> Version: 2.8.10
> Build Date: Aug 18 2014
> Revision Number: 2056
>
> command line - ./cbc /tmp/27068-pulp.mps branch printingOptions rows 
> solution /tmp/27068-pulp.sol (default strategy 1)
> At line 2 NAME          MODEL
> At line 3 ROWS
> At line 8 COLUMNS
> At line 22 RHS
> At line 26 BOUNDS
> At line 32 ENDATA
> Problem MODEL has 3 rows, 4 columns and 6 elements
> Coin0008I MODEL read with 0 errors
> Continuous objective value is 0 - 0.00 seconds
> Cgl0004I processed model has 1 rows, 2 columns (2 integer) and 2 elements
> 1 infeasibilities
> Problem is infeasible!
> Option for printingOptions changed from normal to rows
> Total time (CPU seconds):       0.00   (Wallclock seconds):       0.00
>
> The solution file for this problem and releases/2.8.10 r2045 is:
>
> Optimal - objective value 0.00000000
>       0 C0000000               5                      0
>       1 C0000001            12.4                      0
>       2 C0000002             7.4                      0
>       0 X0000000               0                      1
>       1 X0000001               4                      0
>       2 X0000002               1                      0
>       3 X0000003             8.4                      0
>
> The stdout and solution file are almost the same for stable/2.8 r2056 
> except for the version value.
>
> The expected result is that the solution file should start with 
> "Infeasible" or "Unbounded" instead of "Optimal" for this specific 
> problem.
>
> Regards,
>
> Rémy
>
>
> _______________________________________________
> Cbc mailing list
> Cbc at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/cbc

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/cbc/attachments/20140819/24a47081/attachment-0001.html>