Decision Optimization

Decision Optimization

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


#Analytics
#DecisionOptimization
#DecisionOptimization
 View Only
  • 1.  how to verify that variable priority list is being used

    Posted 06/13/19 06:33 AM

    Originally posted by: srinit34


    hi

    I am using dynamic search and also setting branching variable priorities by using  IloCplex.setpriority () method.

    Is there any way to ensure , from the node log, that my priority order is actually being used ?


    #CPLEXOptimizers
    #DecisionOptimization


  • 2.  Re: how to verify that variable priority list is being used

    Posted 06/13/19 03:26 PM

    Originally posted by: EdKlotz


    First, keep in mind that CPLEX ultimately solves the presolved model, not your original one.    So if you have two variables x1 and x2 with x1 having higher priority, you may never see any branches on x1 and will see x2 first.  You can write out a .pre file that will contain the presolved model to help identify variables that were removed by presolve.

    Second, dynamic search doesn't print out branching selections, so test your orders by running with traditional branch and cut search.    The code that handles branching is not really influenced by whether traditional or dynamic search is in play, so if you can confirm that your orders are working with traditional search, they will also work with dynamic search.

    Third, log files with multiple threads can be tricky in that a branching variable with a lower priority in one subtree might appear before a variables with a higher priority in another subtree.   So do your tests with the threads parameter set to 1.


    #CPLEXOptimizers
    #DecisionOptimization


  • 3.  Re: how to verify that variable priority list is being used

    Posted 06/13/19 04:34 PM

    Originally posted by: srinit34


    Thanks for the clarification. I am assigning the presolved MIP to Cplex.

    I will check with traditional search.


    #CPLEXOptimizers
    #DecisionOptimization