Hands-On Neuroevolution with Python

The evaluate_solutions function implementation

The evaluate_solutions function receives the Robot object as an input parameter, which maintains the population of the maze-solver agent and the maze environment simulator. Also, it receives a list of the coefficient pairs generated during the evaluation of the population of objective function candidates.

We use the input parameters of the function to evaluate each genome in the population and to estimate its fitness function. Here, we discuss the essential implementation details:

1. First, we evaluate each individual in the population against the maze simulator and find the distance to the maze exit at the end of the trajectory:

    robot_genomes = NEAT.GetGenomeList(robot.population) for genome ...