Originally posted by: srinit34

hi

I have a branch callback which simply calls abort.

When I call solve a second time (ie after the abort), I want  optimization to continue from one of the two child nodes of the root, and not from the root itself.

Can this be done?

Thanks !

edit : I also tried using make branch with arguments (0, null ) and ( 1 , null ) just before the abort, but cplex still seems to restart from the root node. Try with MIP b1c1s1 from MIpLIB.

If you abort from the branch callback then branching decisions made in this callback invocation are discarded. Can you try aborting from a node selection callback instead?

Originally posted by: srinit34

thanks Daniel. Yes, I already figured that node callback is the better place to abort when the leaf count exceeds 1.

But I am still confused about 2 things with b1c1s1:

1) upon resumption, cuts seems to be generated again.

2) the node log seems to show 1 leaf, not 2 as I expected.

It also seems that on the second invocation of solve, I need an empty branch callback. If I just remove all callbacks and have traditional search, that is not enough to resume search?

       Nodes                                         Cuts/
   Node  Left     Objective  IInf  Best Integer    Best Bound    ItCnt     Gap         Variable B NodeID Parent  Depth

      0     1    20167.0784   234                   Cuts: 168    12465