Decision Optimization

 View Only

  • 1.  Modify the constraint for infeasiblity

    Posted Tue December 14, 2021 02:39 AM
    Hi
    I have a model in which constraints are very hard and due to which I am getting infeasible solution and the model (CPLEX) is taking huge amount of time to Solve it.
    C0#0: A.0#0 + A.1#1 + A.2#2 + A.3#3 + A.4#4 + A.5#5 + A.6#6 + A.7#7 + A.8#8 = 1
    C#1: A.0#0 + A.1#1 + A.2#2 + A.3#3 + A.4#4 + A.5#5 + A.6#6 + A.7#7 + A.8#8  = 2

    Is it possible to formulate these constraint in a alternate way that CPLEX can able to find the infeasibility quickly. I tried to use an indicator variable to formulate these equality to inequality constraint but didnt get much difference.




    ------------------------------
    sandeep singh chauhan
    ------------------------------

    #DecisionOptimization


  • 2.  RE: Modify the constraint for infeasiblity

    Posted Wed December 15, 2021 10:27 AM
    I'm not sure I understand your need (and the attached model is not clear also), but did you try the feasibility pump heuristic ?https://www.ibm.com/docs/en/icos/20.1.0?topic=parameters-feasibility-pump-switch

    ------------------------------
    Vincent Beraudier
    ------------------------------

    Original Message


  • 3.  RE: Modify the constraint for infeasiblity

    Posted Mon December 20, 2021 01:45 AM
    Hi 
    Thanks for your response. I am not looking for feasibility pump. Is it possible to find the infeasible solution by changing the constraint. If a + b + c = 2 is modified as (a+b+c) - 2M >= 0 where M is binary variable.

    ------------------------------
    sandeep singh chauhan
    ------------------------------

    Original Message