Decision Optimization

Decision Optimization

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

 View Only
  • 1.  why the LP iterations stays the same?

    Posted Tue September 30, 2014 05:48 PM

    Originally posted by: Uonly


    Iteration log . . .
    Iteration:     1   Dual objective     =        794618.792788
    Iteration:    38   Dual infeasibility =             0.000042
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788
    Iteration:    47   Dual objective     =        794618.792788

     

    It is clear that it is doing some pricing and pivoting, but why the iteration number is the same?

     

    Thanks

     


    #CPLEXOptimizers
    #DecisionOptimization


  • 2.  Re: why the LP iterations stays the same?

    Posted Thu October 02, 2014 02:26 AM

    Originally posted by: RWunderling


    This looks indeed strange - I don't recall ever seeing this before.  Could you please PM me a sav file of the model and detailed instructions on how to reproduce this behaviour (including port, version, parameter settings)?

    Thanks,

    Roland


    #CPLEXOptimizers
    #DecisionOptimization


  • 3.  Re: why the LP iterations stays the same?

    Posted Thu October 02, 2014 03:21 AM

    Originally posted by: T_O


    My first guess was that it is doing some "mini-iterations" of the bound flipping ratio test, but of course, I do not know the implementation details in CPLEX. Maybe, it is something completely different.


    #CPLEXOptimizers
    #DecisionOptimization


  • 4.  Re: why the LP iterations stays the same?

    Posted Thu October 02, 2014 10:02 AM

    Originally posted by: Uonly


    Yes I was wondering why. I am too bold to suspect that could be "bug" in cplex. when I was setting the SIMDISPLAY to be 2. I can observe many iteratoins of price in and out when the iterations number is 47.

    Btw, would you please provide your email address so that I can send you the code and explain how to repeat the issue?

    Thanks again Roland.!

     

     


    #CPLEXOptimizers
    #DecisionOptimization


  • 5.  Re: why the LP iterations stays the same?

    Posted Fri October 03, 2014 02:28 AM

    Originally posted by: RWunderling


    Please PM to roland<dot>wunderling<at>at<dot>ibm<dot>com (no typo with the <at>at).
    Thanks,
    Roland


    #CPLEXOptimizers
    #DecisionOptimization


  • 6.  Re: why the LP iterations stays the same?

    Posted Mon October 06, 2014 05:41 AM

    Originally posted by: RWunderling


    Thanks for providing the sample.  This is indeed a minor bug, whereby the iteration count is not increased when dual superbasic variables are pivoted out of the basis.  It will be fixed in the next release.


    #CPLEXOptimizers
    #DecisionOptimization