GCG

This classifier adds constraints to classes according to the differences in their names with a certain threshold corresponding to the Levenshtein distance.

Classification

Dantzig-Wolfe and Benders Decomposition

In both modes, the handling is as follows:

• For each constraint \(i\), the constraint name is read and the distance to each other constraint name is calculated (complexity \(\mathcal{O}(n^2)\))
  • A breadth first search is conducted until every constraint is assigned to a class
    • Iff the Levenshtein distance bewtween two constraints is higher than connectivity (by default 1), they aren't grouped into the same class.
• If the number of classes exceeds nmaxconss (by default 5000), the classifier will abort.

Example