Decision Optimization

Decision Optimization

Delivers prescriptive analytics capabilities and decision intelligence to improve decision-making.


#Analytics
#DecisionOptimization
#DecisionOptimization
 View Only
  • 1.  How can status 118 happen for pure binary programs?

    Posted 03/12/13 04:57 PM

    Originally posted by: T_O


    Hi,

    we have a pure binary program where we do a lot of bound changes (fixing / unfixing variables to 0/1). After some changes are done, we reoptimize. This works quite well, but after some time, CPLEX returns status code 118 (CPXMIP_UNBOUNDED). If I write a .prm and .sav file and read them into the interactive optimizer, everything is fine, nothing is umbounded, I can obtain a solution. The kappastats say that all bases are stable.

    
    CPLEX> dis sol qual Incumbent solution: MILP objective                                -1.3900000000e+002 MILP solution norm |x| (Total, Max)           2.80000e+002 1.00000e+000 MILP solution error (Ax=b) (Total, Max)       0.00000e+000 0.00000e+000 MILP x bound error (Total, Max)               0.00000e+000 0.00000e+000 MILP x integrality error (Total, Max)         0.00000e+000 0.00000e+000 MILP slack bound error (Total, Max)           0.00000e+000 0.00000e+000   Branch-and-cut subproblem optimization: Max condition number:                    4.8820e+006 Percentage (number) of stable bases:     100.00%   (95) Percentage (number) of suspicious bases:   0.00%   (0) Percentage (number) of unstable bases:     0.00%   (0) Percentage (number) of ill-posed bases:    0.00%   (0)
    


    
    CPLEX> dis prob stat Problem name: debug.sav Variables            :    6068  [Binary: 6068] Objective nonzeros   :      16 Linear constraints   :    9211  [Less: 8423,  Greater: 762,  Equal: 26] Nonzeros           :   25257 RHS nonzeros       :    1244   Variables            : Min LB: 0.0000000        Max UB: 1.000000 Objective nonzeros   : Min   : 1.000000         Max   : 144.0000 Linear constraints   : Nonzeros           : Min   : 1.000000         Max   : 256.0000 RHS nonzeros       : Min   : 1.000000         Max   : 5.000000
    


    Any ideas?

    Best regards,
    Thomas
    #CPLEXOptimizers
    #DecisionOptimization


  • 2.  Re: How can status 118 happen for pure binary programs?

    Posted 03/13/13 02:19 AM

    Originally posted by: T_O


    Here is an example, where it also happens when reading the sav-file.

    It does not happen, if I disable repeated presolve.

    Best regards,
    Thomas
    #CPLEXOptimizers
    #DecisionOptimization


  • 3.  Re: How can status 118 happen for pure binary programs?

    Posted 03/13/13 02:34 AM

    Originally posted by: SystemAdmin


    I cannot reproduce the problem here. When I just solve your SAV file with CPLEX 12.4 or CPLEX 12.5 then I get -139 as optimal objective value.
    What version of CPLEX do you use? What machine/OS? Do you use any non-default parameter settings? Could you please attach a log file for the offending solve?
    #CPLEXOptimizers
    #DecisionOptimization


  • 4.  Re: How can status 118 happen for pure binary programs?

    Posted 03/13/13 02:47 AM

    Originally posted by: T_O


    CPLEX 12.5, Windows 7 64-bit, 16 GB Ram, i7 3520M @ 2,9GHz

    I attached the log file. In this example, i did not set any parameters. Just
    r debug
    o
    


    I get -139 as objective, but it says unbounded:

    MIP - Integer unbounded:  Objective = -1.3900000000e+002
    


    Best regards,
    Thomas
    #CPLEXOptimizers
    #DecisionOptimization


  • 5.  Re: How can status 118 happen for pure binary programs?

    Posted 03/13/13 06:44 AM

    Originally posted by: SystemAdmin


    OK, I reproduced the problem here. The issue is exactly the same as already discussed a few days ago in this thread.
    The workaround is to disable repeated presolve or fix up the status manually, see also in that thread.
    #CPLEXOptimizers
    #DecisionOptimization


  • 6.  Re: How can status 118 happen for pure binary programs?

    Posted 03/13/13 08:30 AM

    Originally posted by: T_O


    Ok, thank you. I disabled repeated presolve.

    Best regards,
    Thomas
    #CPLEXOptimizers
    #DecisionOptimization