This repository was my first experimentation using neural networks evolved via a genetic algorithm. It has been through various revisions but the core concept is the same.
It consists of a simulation of creatures that must obtain food to increase their energy levels. After a set period of time (in ticks), the creatures are bred together to create a new population and the simulation restarts.
The new creatures are created by breeding and mutating two creatures from the previous generation. Parents are selected at random, with a greater chance of creatures with higher energy being selected (known as roulette wheel selection).