Decision Optimization

Decision Optimization

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


#Analytics
#DecisionOptimization
#DecisionOptimization
 View Only
  • 1.  CPLEX 12.1 vs CPLEX 12.2

    Posted 09/03/11 04:28 AM

    Originally posted by: SystemAdmin


    Hi everybody,

    I had the problem with CPLEX 12.1 that (using the Python API) problems could not be solved during the runtime of my code by the "del"-statement of python. I was told that this bug was fixed by 12.1.1. Unfortunately, as a Acadamic Initiative User, I cannot access this fix. So I am trying to use the 12.2 version, but was a bit disappointed right now:

    1.) How can I delete a problem for memory in CPLEX 12.2 using the Python API? The del statement apparently still does not work.

    2.) In my Branch-and-Price framework CPLEX 12.2 computes a different solution than CPLEX 12.1 - using the same code and the same problem data. How can this be? Are there significant changes concering the accuracy or something?

    3.) In Cplex 12.2 I am not able to turn of the display output of the MIP optimizer completly (even setting the parameter to zero). When the optimal solution is found, I always get a message in the console. How can I avoid this?

    Thanks a lot and best regards
    #CPLEXOptimizers
    #DecisionOptimization


  • 2.  Re: CPLEX 12.1 vs CPLEX 12.2

    Posted 09/03/11 04:56 AM

    Originally posted by: SystemAdmin


    Update: When I turn off the heuristics in the mip solving procedure I get the same results in both cases. Quite strange!
    #CPLEXOptimizers
    #DecisionOptimization


  • 3.  Re: CPLEX 12.1 vs CPLEX 12.2

    Posted 09/06/11 01:17 PM

    Originally posted by: SystemAdmin


    Hello,

    Regarding 1: Do you find that when using CPLEX 12.2, your program still terminates with an error condition that seems to be due to running out of memory? If so, could you please either post your code as an attachment to this forum thread, or send it to pstarhill (at) us (dot) ibm (dot) com?

    Regarding 2: Do you find that there is a different optimal objective function value, or simply a different solution vector?

    Regarding 3: You should be able to disable output of the optimizer by using code like the following:
    # set up an instance of the Cplex class
    my_cplex_object.set_results_stream(None)
    my_cplex_object.set_log_stream(None)
    

    You can also redirect these streams to any object with "write" and "flush" methods, or pass in a string to create a new file and write the results to disk. There is also a warning stream and an error stream, but you probably don't want to disable those entirely.
    Philip Starhill
    CPLEX Research Engineer
    #CPLEXOptimizers
    #DecisionOptimization