Methdology

The Min Conflict Heuristic is one of the first heuristics used
for constraint satisfaction problems.
(Minton et Al, 1992).

It chooses a variable randomly from the set of variables with
conflicts (red edges). The value which results in the fewest
conflicts is chosen from the possible assignments for this variable.

This heuristic is almost identical to Greedy Descent with Random Walk
with a two stage heuristic choosing any red node 100% and best 
value 100%. The only difference is the MCH may return to the 
previous assignment.