Hands-On Neuroevolution with Python

NoveltyItem

This class is the main structure that holds information about the novelty score of each individual evaluated during the evolution. It has several fields that store relevant information, as we can see in the source code:

    def __init__(self, generation=-1, genomeId=-1, fitness=-1, novelty=-1):        self.generation = generation        self.genomeId = genomeId        self.fitness = fitness        self.novelty = novelty        self.in_archive = False        self.data = []

The generation field holds the ID of the generation at which this item was created. Basically, genomeId is the ID of the genome that was evaluated, and fitness is a goal-oriented fitness score of the evaluated genome (the proximity to the maze exit). Furthermore, novelty is the novelty score given to the ...