Decision Optimization

Decision Optimization

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


#Analytics
#DecisionOptimization
#DecisionOptimization
 View Only
  • 1.  CPLEX 12.5 with CPX_PARAM_OBJULIM / CPX_PARAM_OBJLLIM

    Posted 01/21/13 09:38 AM

    Originally posted by: SystemAdmin


    In CPLEX 12.5 to get the default, min, max values for upper and lower limits on the objective value I use

    int CPXXinfodblparam(CPXCENVptr env, int whichparam, double * defvalue_p, double * minvalue_p, double * maxvalue_p)

    For whichparam = 1026 (CPX_PARAM_OBJULIM) and whichparam = 1025 (CPX_PARAM_OBJULIM), the output defaults are +1e+75 and -1e75 (which are correct), whereas the min and max values are 0.0.

    Please correct this trivial bug.
    #CPLEXOptimizers
    #DecisionOptimization


  • 2.  Re: CPLEX 12.5 with CPX_PARAM_OBJULIM / CPX_PARAM_OBJLLIM

    Posted 01/21/13 10:56 AM

    Originally posted by: SystemAdmin


    That is not a bug.
    If the function returns 0 for both the min and max value then this means that the parameter has no limits. See the reference documentation.
    #CPLEXOptimizers
    #DecisionOptimization


  • 3.  Re: CPLEX 12.5 with CPX_PARAM_OBJULIM / CPX_PARAM_OBJLLIM

    Posted 01/21/13 03:15 PM

    Originally posted by: SystemAdmin


    Nice, it seems to be even worse than a "trivial bug". First, how do you check two doubles for zero. Second, what do you mean with "the parameter has no limit". Please remember that it should work for all platforms.
    #CPLEXOptimizers
    #DecisionOptimization


  • 4.  Re: CPLEX 12.5 with CPX_PARAM_OBJULIM / CPX_PARAM_OBJLLIM

    Posted 01/22/13 01:27 AM

    Originally posted by: SystemAdmin


    I'm afraid I don't see your problem here :-(

    > Fish2013 wrote:
    > Nice, it seems to be even worse than a "trivial bug". First, how do you check two doubles for zero.
    >
    You can simply use the '==' operator to do that. Since these are constants there is no round-off impossible and using '==' on floating point numbers is fine here.

    > Second, what do you mean with "the parameter has no limit".
    >
    It means that you can specify any double precision floating point value here, including positive and negative infinity -- though using a NaN here might result in unexpected behavior.

    > Please remember that it should work for all platforms.
    >
    I don't see how any of the above would depend on the platform (for the platforms that CPLEX supports).
    #CPLEXOptimizers
    #DecisionOptimization


  • 5.  Re: CPLEX 12.5 with CPX_PARAM_OBJULIM / CPX_PARAM_OBJLLIM

    Posted 01/22/13 02:38 AM

    Originally posted by: SystemAdmin


    Do you mean I should use minvalue_p == 0.0 && maxvalue_p == 0.0?
    #CPLEXOptimizers
    #DecisionOptimization


  • 6.  Re: CPLEX 12.5 with CPX_PARAM_OBJULIM / CPX_PARAM_OBJLLIM

    Posted 01/22/13 04:11 AM

    Originally posted by: SystemAdmin


    > Fish2013 wrote:
    > Do you mean I should use minvalue_p == 0.0 && maxvalue_p == 0.0?
    >
    Yes, that is what I meant.
    #CPLEXOptimizers
    #DecisionOptimization


  • 7.  Re: CPLEX 12.5 with CPX_PARAM_OBJULIM / CPX_PARAM_OBJLLIM

    Posted 01/22/13 07:15 AM

    Originally posted by: SystemAdmin


    Hi, Daniel. Thanks for your comments.
    #CPLEXOptimizers
    #DecisionOptimization