The program I wrote is a basic prototype that simulates gradual artificial intelligence through the genetic formula.

By generation 100 the dots will usually have a nice straight path to the dot at the top (goal) and will have a fitness of about 300.

In there early stages of training, they will not be very intelligent.

After generation 1, the dots will follow the one that did the best throughout the last generation and "branch" off to find better paths and get a better fitness.

The fitness is calculated by how close it gets and how fast it gets there, kind of like when you are racing someone. The winner is the one who got to the goal, well, fastest.

The link to this page is halestorm512.itch.io/artificial-intelligence-with-genetic-algorithm (Sorry for the long URL it was so people understood the URL name).

Here's some sample thumbnails of the generations improving.

As you can see, the difference from generation 0 to generation 1 is that at generation 1 they are a lot closer to the dot. The green dot stands for the one that did best in the previous generation.

Different things happen every time you run it so give it a try and comment on my page your guys's responses!