In class, it was suggested that if one had an ensemble of admissible heuristics $,$ one can define a new heuristic that gives for any$1,2,...,$

node :$$