Quote:
Originally posted by Colin@Jan 13 2003, 11:36 AM
Peaches, Lady law,
Take the case of flocking behavior. If you were making a movie and wanted to simulate what a flock of flying geese look like, what rules would you give each goose to make it look real? It turns out that you can use just a few simple rules and the computer-generated flying flock will look for the most part like a real one. (whether real flocks use such simple rules, no one knows).
For example.
1. Don't fly too close to other birds
2. Fly in the average direction of the other birds.
3. Fly towards the average position of the other birds.
You can combine these rules in such a way as to resolve conflicts between competing rules. ("Don't fly into another bird" would have high priority).
That's am old simple one that is decent by itself. One can make more elaborate goals and behavior on top of this. AI birds are called "boids".
Why might someone care?
1. To make a better movie. Why program the position of 100s of flocking/hearding creatures when one can just use a few simple rules and let the program run?
2. To understand how nature really works. Is there a flock leader? If not, how do
birds know where to go without a leader?
|
Yes, all that and more in fractals...
Mandelbrot established an "index" which each self-replicating phenome can be scaled by but still maintain its essential structure... like DNA
sometimes threads grow fractally...
Last edited by heqdvd at Jan 13 2003, 05:27 PM