#52 suvorov too slow

[Jeremy Carroll]

I thinking pruning bad solns earlier may speed things up.
Tsukamoto has only 10% success rate

[Jeremy Carroll]

I have added some pruning by looking for unnecessary connections at vertices and blocking edges with two unnecessary connections early.

Also need to block:

• single vertex where all its edges have unnecessary connections
• two vertices joined by a necessary link but where all other edges are unnecessary
• could also precompute all sixes and block any extension of one of those
• could also for each edge, and vertex and pair of vertices in the original graph do a precompute search for first solution, if none, then remove edge, vertex, or vertex-pair from search space.