Decision Optimization

Decision Optimization

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


#Analytics
#DecisionOptimization
#DecisionOptimization
 View Only
  • 1.  Node id sequence at a particular node in the tree

    Posted 09/05/17 05:17 PM

    Originally posted by: raswalgul


    I'm using callbacks to try to get information on the partial solution at a particular node in the branch and bound tree. By partial solution, I mean the incomplete solution formed by the branches created at all the ancestors of the given node. But I'm unable to directly retrieve a partial solution at a node.

     

    One way to do this if I maintain a datastructure that stores which variable is branched on at what node_id. To do this, one must know the sequence of node_ids that would constitute a path in the tree. OR, one must at least be able to retrieve the node_id of the parent node of the given node. But I can't retrieve this information either.

     

    Can any one help me with this?


    #CPLEXOptimizers
    #DecisionOptimization


  • 2.  Re: Node id sequence at a particular node in the tree

    Posted 09/06/17 12:45 AM

    CPLEX does not maintain the tree structure in a way that can be queried. However, you can track the tree structure of the branching tree my means of using a branch callback and user data attached to each node. How to do this has been discussed on this forum several times.

    However, if I understand correctly, then what you are really looking for is the set of fixed variables at a node. There is an easier way to do this: you can query the local bounds from CPLEX. Fixed variables will have identical lower and upper bounds. See functions CPXgetcallbacknodelb,CPXgetcallbacknodeub for C or the ControlCallback's getLB() and getUB() methods in the object oriented APIs.


    #CPLEXOptimizers
    #DecisionOptimization


  • 3.  Re: Node id sequence at a particular node in the tree

    Posted 09/06/17 12:55 AM

    Originally posted by: raswalgul


    Thank you for the response. It looks like the easier way you've mentioned would solve my problem!


    #CPLEXOptimizers
    #DecisionOptimization