RANDOM BIG NETS: Rare unsuccessful noncoalition creation

[lcubrilo]

Bug for feature #8 #8

Although rare, they can easily happen. When making sure to generate a unclusterable network (by corrupting some coalitions' edges): have to make sure those edges aren't a bridge of the coalition.

In that case, we have not corrupted a cluster into a noncoalition, but merely split it into two coalitions.

Practically irrelevant, since the final count of noncoalitions DOES match reality (even if it isn't equal to the starting/planned number of noncoalitions).
However number of coalitions (and therefore also all clusters/components) has thus risen by one, which is unaccounted in every way. Again, law of large numbers means:
1. Reduces the amount of bridge nodes.
2. It's practically irrelevant for the final percentage