FAQ
Cover
This is the archive Discourse for the Processing (ALPHA) software.
Please visit the new Processing forum for current information.

   Processing 1.0 _ALPHA_
   Topics & Contributions
   Simulation, Artificial Life (Moderator: REAS)
   Emergent Behaviour		 « Previous topic | Next topic »

Pages: 1 
   Author  Topic: Emergent Behaviour  (Read 843 times)
Dan

WWW
Emergent Behaviour
« on: Apr 23rd, 2003, 8:04pm »
I'm really interested in emergent behaviour from simple rule sets, but haven't had much time to play around recently.  Here are 4 things based on the same simple rules, just slightly tweaked parameters.
 
http://pupius.co.uk/stuff/experiments/EM0/
http://pupius.co.uk/stuff/experiments/EM1/
http://pupius.co.uk/stuff/experiments/EM2/
http://pupius.co.uk/stuff/experiments/EM3/
 
I started out trying to simulateflocking movement in sheep, but got interesting results and carried off on a tangent instead of my original plan.
 
REAS


WWW
Re: Emergent Behaviour
« Reply #1 on: Apr 28th, 2003, 12:32pm »
I think EM1 has alot of potential. What are your primary inspirations for this area of exploration?
 
Dan

WWW
Re: Emergent Behaviour
« Reply #2 on: Apr 28th, 2003, 3:54pm »
Well, I did a BSc in Artificial Intelligence and my disertation was an investigation into using the probabalistic search method of the desert ant, Catagylphis Fortis, in mobile robots.
 
It turned out that the behaviour could be broken down into a basic set of rules.  I then used a radial-basis function net with a hardwired critic to ensure that the rules were obeyed.  The robot knew certain actions such as "Turn left", "turn right" with which to complete the goals.
 
These examples are just little play things that took around an hour, using only particles with hard wired rules.  If I get time I'd like to experiment further, with more complex behaviour, introducing rules that change/mutate over time, making the particles more like "agents".
 
Pages: 1 

« Previous topic | Next topic »