Comment by cerved

Define large. We've written model which solves real business issues in 8K lines of MiniZinc and it wasn't slow.

The conventional wisdom is the larger you make an NP hard problem, the slower is going to get. Irregardless of algorithm.