Genetic algorithm of points trying to reach a goal: trough successive generations, points are learning from the two "parents" to reach the goal. Parents are defined as the two most efficient points from the previous generation in regard of "moves to reach". Each point has a basic energy which is spent up/down the slopes of the map. They die when their energy is when they run out of energy or touch the edges of the image. Several mutation parameters are available.
Feel free to reuse and modify - MIT License.