Decision Optimization

Decision Optimization

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

 View Only
Expand all | Collapse all

Branching preferences and time to optimality gap in OPL

  • 1.  Branching preferences and time to optimality gap in OPL

    Posted Fri February 07, 2014 12:46 PM

    Originally posted by: mbam


    Hello,

    I have two questions:

    1) Is there a way to control the variables that are branched on in a MIP? For example, if there are two sets of variables, x and y, I want to give priority to x.

    2) How can I output the time it took to solve to an optimality gap of 10%, 5%, 1%, and 0% in the same run?

     


    #DecisionOptimization
    #OPLusingCPLEXOptimizer


  • 2.  Re: Branching preferences and time to optimality gap in OPL

    Posted Sat February 08, 2014 07:54 AM

    Hi

     

    1) You should use

    priority : Accesses the priority level of the decision variable (CPLEX models only).

    Example:

    dvar int x;

    execute

    {

    x.priority=2;

    }

     

    2) You should use a main (flow control)  and do different runs with different epgap

    main

    {

    thisOplModel.generate();

    cplex.epgap=0.1;

    cplex.solve();

    cplex.epgap=0.05;

    cplex.solve();

    cplex.epgap=0.01;

    cplex.solve();

    cplex.epgap=0.000001;

    cplex.solve();

    }

    regards

     

     


    #DecisionOptimization
    #OPLusingCPLEXOptimizer


  • 3.  Re: Branching preferences and time to optimality gap in OPL

    Posted Mon February 10, 2014 11:12 AM

    Originally posted by: mbam


    Thank you very much for your reply. I will try the branching priority.

    For the optimality gap, I need to be able to get the time for all those gaps within one run. I have been doing what you suggested above, but I get inconsistent results. So I need to get the time within one run. Do you have any idea how to do that?

     


    #DecisionOptimization
    #OPLusingCPLEXOptimizer


  • 4.  Re: Branching preferences and time to optimality gap in OPL

    Posted Mon February 24, 2014 02:56 PM

    Originally posted by: mbam


    I am still struggling with the second question. Any help would be much appreciated.

    Thank you.


    #DecisionOptimization
    #OPLusingCPLEXOptimizer