Genetic programming is an area of research in computer science interested in finding an "optimal" solution in a large search space. Traditional genetic programming relies on a fitness function which gives an indication of how close a given genome is to optimal. Unfortunately, many interesting problems don't have any natural fitness function, such as images, objects, L-Systems, music, etc...
I wrote a general toolkit for investigating such problems. Judging from the reactions of people who have used the toolkit, it is an easily understood, useable, and addicting system for exploring visual genetic programming.
The original report on the toolkit, as it appeared in the Geometry Center Preprint GCG-74, is available in either Postscript or HTML.
The user interface:
(184K PNG)(53K JPG)
(632K PNG)(168K JPG)
(392K PNG)(120K JPG)
(92K PNG)(78K JPG)
(360K PNG)(74K JPG)
(272K PNG)(58K JPG)
696K MPG
1512K MPG
216K MPG
320K MPG
520K MPG
960K MPG
792K MPG
