This project describes a genetic algorithm used to make a picture. _ALGORITHM_ We are first given a template picture. We make multiple copies of pictures with randomly colored dots. Each of these pictures, called "offspring" is compared to the template, and given a calculated number. The lower the number is, the closer the offspring is to the template. Now we sort the pictures in order, from most similar to the template to least similar. The least similar half of all the pictures are then deleted, and replaced with copies of the other half of the pictures, though now each point will be slightly changed, or mutated, by a random number. This process is called a generation of offspring, and this process is repeated over and over again. After every generation the best offspring will be shown. This best offspring will change over time, each time looking more and more like the template. This is true because each generation will only preserve the best offspring, even though the mutations are random. It usually takes quite a while for the shown offspring image to start appearing anything like the template. This is normal, as the mutations are very minute, which simulate the evolution of living things. The offspring will start to resemble the template at around Generation #4000.
Thanks to the genetic algorithm videos on YouTube for inspiration. (Genetic algorithms can be used to make a rough identification of a criminal, by making computer-generated faces. They typically perform better than police sketches, as seen in https://www.youtube.com/watch?v=h81SuD2pltM )
