The closer the cats get to the goal, the bigger their fitness is. Basically, the cats try to get to the other side. The one that gets the furthest has its genes pass on to the next generation, while all the others die. The green cat is the most fit one of the last generation. They start with only 40 moves(of move left, move right, and jump), and they gain 20 each generation. edit 2024: disclaimer. It's important to note that this project is an approximation of what you might consider "AI". While this is a form of "Artificial Intelligence" it isn't anywhere near the complexity of a proper neural network. Funnily, at the time of the project creation, I was unsure on this , hence the (?) on the title.
In my testing: The cats jump over the first block in generation 10 The cats get over the second wall around the 15th generation The cats climb the big wall around the 30th - 50th generation They finally reach their goal at around 70. But this is random, because natural selection* is random.
