Ant Brains and Swarm Theory

Who remembers John Conway's Game of Life from 1970?  (I think my first "graphic" programming effort was this thing around 1977 - probably written in PL/1 - really dates me!).  Thoughts around these types of computer simulation (often mimicing something in the biological world) triggered much of the efforts around Artificial Intelligence (AI). 

AI came in fits and starts, and some of it has proven useful when applied to real-world problems.  AI has not have that stellar a success rate in the management of IT systems (who remembers Neugents from Computer Associates?). 

What got me thinking about this is an article in the July 2007 National Geographic by Peter Miller titled "Swarm Theory."  In the article, Deborah Gordon, a biologist at Stanford University says "Ants aren't smart, ant colonies are."  One implication is that while building AI software systems to do things that come naturally to human 3-year olds is a struggle, it is possible to build systems that are at least as smart as an ant.  Collections of these "software ants" can then do useful things together.

Simple creatures following simple rules, each acting on local information, drives "swarm intelligence."  (There's a nod to James Surowiecki's The Wisdom of Crowds here.)  In the article, Miller lists a number of amazing applications of this technique - both in the animal/insect world and in the technology world.  One example is telephone companies using "virtual pheromone deposits" in switches to indicate best paths for call setup and transmission. I can envision software using these methods in grid and datacenter computing, performance balancing, etc. Check out the article and especially the great photos.

(written 07/31/07)