Hands-On Neuroevolution with Python

The evaluate_solution_fitness function implementation

This function is an implementation of the maze-solver fitness function that we discussed earlier. This function receives the distance to the maze exit, the novelty score, and the list of coefficient pairs generated by the current generation of the objective function candidates. Next, it uses the received input parameters to calculate the fitness score as follows:

    normalized_novelty = novelty    if novelty >= 1.00:        normalized_novelty = math.log(novelty)    norm_distance = math.log(distance)    max_fitness = 0    best_coeffs = [-1, -1]    for coeff in obj_func_coeffs:        fitness = coeff[0] / norm_distance + coeff[1] * normalized_novelty        if fitness > max_fitness:            max_fitness = fitness best_coeffs[0] = coeff[0] ...