I’ve decided not to add this project to my portfolio, because it’s more complex and needs some words of explanation on this blog. It’s main purpose is to simulate different algorithms used in AI programming. It does so, by applying bayesian nets to agents ( cube-like robots ). Each agent as it is seen on simulation, eats little blue cubes. They have 3 states of hunger – 3 = not hungry, 2 = hungry, 1 = v.hungry. Each robot can shoot, but shooting makes them more hungry. Robot shoots towards his oponent only if the oponent is in robot’s range of sight, and if the robot isn’t very hungry.
When food and enemy isn’t visible to the robot, robot simply walks around doing nothing. Each robot has its bayesian net which decides what to do – shoot enemy, eat , run from enemy and shoot, and so on. You can tech your own robot, by clicking “add human” and eating or shooting by yourself. Then, when you click “learn”, AI takes control of your robot and tries to imitate your moves.
Button called “decrobo” places new robot driven by decision network. It takes time to learn for the first time.
