Decision Optimization

Decision Optimization

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

 View Only
  • 1.  Weird node id found in cplex log

    Posted Tue January 17, 2012 06:47 AM

    Originally posted by: far123


    In the following Cplex log, I found that the NodeID of a node is always in the same parity with its Parent except the direct "children" of the root node. For example, the NodeID is 2 for the node #1 and its Parent's NodeId is 0, the NodeID is 3 for the node #5 and its Parent's NodeId is 1.
    
    Nodes                                         Cuts/  Node  Left     Objective  IInf  Best Integer    Best Bound    ItCnt     Gap         Variable B NodeID Parent  Depth 0     2  1390228.6373   244                1390228.6373     6999                                 0             0 Elapsed real time =  13.89 sec. (tree size =  0.01 MB, solutions = 0) 1     3  1390281.1144   245                1390228.6373     7070                    x1623 D      2      0      1 2     4  1390287.6935   171                1390286.8699     7124                    x1625 D      4      2      2 3     5  1402206.4736   163                1390286.8699     7186                    x1623 U      1      0      1 4     6  1401993.9676   176                1401961.4707     7275                   x66611 D      6      4      3 5     7  1402206.5425   142                1401961.4707     7291                   x39696 U      3      1      2 6     8  1402919.4733   261                1401994.0276     7352                   x56550 U      8      6      4 7     9  1402209.9869   235                1401994.0276     7396                   x56447 D      5      3      3 8    10  1402098.2509   172                1402014.9800     7416                   x56550 D     10      6      4 9    11  1403132.8412   232                1402014.9800     7444                   x56550 U      7      5      4 10    12  1402098.8226   121                1402098.8142     7464                   x65025 D     12     10      5 ...... ...... 792   794  1431419.0226    78                1402098.9676    28014                   x48761 D    896    894    341 *   793   791      integral     0  1405281.1161  1402098.9676    28024    0.23%          x69305 U    689    687    254 794   792  1431419.0501    92  1405281.1161  1402098.9676    28031    0.23%          x30559 D    898    896    342 795   793  1403452.3180   192  1405281.1161  1402098.9676    28173    0.23%          x66557 D    691     12      6 796   794  1455210.1849     1  1405281.1161  1402098.9676    28212    0.23%          x30559 U    900    896    342 797   795  1432473.8286    76  1405281.1161  1402098.9676    28268    0.23%          x68988 U    902    898    343 798   253  1403479.5874   213  1405281.1161  1402121.9693    28416    0.22%          x66557 D    906     20      8 799   254  1403452.3277   189  1405281.1161  1402121.9693    28418    0.22%          x39474 D    693    691      7 800   255  1403479.5912   203  1405281.1161  1402121.9693    28423    0.22%          x39474 D    908    906      9
    

    However, there is an exception for the node #795, whose NodeID is 691 while the Parent's NodeID is 12.

    Anyone can help with this?
    #CPLEXOptimizers
    #DecisionOptimization


  • 2.  Re: Weird node id found in cplex log

    Posted Tue January 17, 2012 09:24 AM

    Originally posted by: SystemAdmin


    The id of a node is just a number that CPLEX uses to uniquely identify a node. You should consider the node id as an opaque value. No special meaning is attached to the actual value of that id. In particular, CPLEX does not enforce the parity property you observed. It is just coincidence that this property seems to hold for a lot of nodes with CPLEX' current strategy of choosing ids.
    #CPLEXOptimizers
    #DecisionOptimization