Originally posted by: BastianMeier
Hello,
i am trying to become familiar with Branch&Cut in Cplex. I have read about Goals and tried to implement one.
By optimizing the trivial linear problem:
"Maximize
obj:
subject to
10 tau >= 1
Generals
tau
End
"
i found that (presolved switched off) the optimal value for tau of the relaxed problem in the root node, as read by my Goal, is 1.0.
Solving the problem without the "Generals tau"-part, which is the root problem, led to tau = 0.1.
Of course both solutions are right. I just wonder, if cplex automatically tries to find the optimal solution with as less fractional values for int-variables as possible or if i got something wrong by using Goals to read out the solutions at every node in the b&b-tree.
I couldn't find any description on how cplex is postprocessing the results of a node in the b&b-tree.
Thank you for satisfying my curiosity,
Bastian
#CPLEXOptimizers#DecisionOptimization