Novelty Search for Deep Reinforcement Learning Policy Network Weights by Action Sequence Edit Metric Distance