Genetic Algorithm Obstacle Course

A genetic algorithm designed to teach the robots how to reach the end goal of the obstacle course. A new generation cycle starts every 15 seconds. 

Each robot has 5 thrusters that have 5 stats each (Power, angle, activation time, interval, duration). All stats start completely random. Every generation the organisms get a fitness score based on multiple criteria: 

1. Death (by hitting a big rock) 

2. Number of collisions with smaller rocks.

3. Distance towards the end goal.

After which each robot gets a certain percentage chance of breeding based on their fitness score. The child robots have a chance to take a thruster stat of one of their parents. Besides that there is a percentage chance per stat to mutate slightly to avoid getting stuck.

That's basically it, with every generation cycle you will see robots getting closer towards the goal, until most of them hit the goal.