We test the exact algorithm and the heuristic on real world, artificially created and random instances. Although the MILP is expected to optimally solve only the small test instances, it can be viewed ...