home

about us

articles

software

get involved

 

Primitive Life

Primitive Life is a basic artificial intelligence simulation that allows little insect-like creatures to choose between mating and eating food. The models are based on a very basic neural-network-based design that analyzes the benefits a create feels from eat and mating. While an artificial learning environment would change these values, the applet allows you to modify them and see their effects.

Specifically, there are four values one can modify:

Food Growth

Food (green squares) grows using a cellular automata-based system where new food appears depending on where old food is. If there is too much food in an area, certain food (i.e. plants) does not have enough nutrients to grow, and so it dies. This is a good example of emergent behaviour: the food actually self-organizes into grids, which is not built into the system.

Creature Characteristics

It is possible to set the amount of utility / pleasure a creature gains from mating and eating by setting a score between -10 and 10. The lower the score, the less pleasure a creature feels. It is also possible to set the initial population, between 1 and 200. Note that each square on the can hold an infinite number of creatures.

The source code for Primitive Life is available, but is not commented. The simulation is available on this server as a Java applet.

Launch Java Applet

Source Code (Zip - 9 kb)