An article in a recent issue of Scientific American (I believe) described a simple algorithim that duplicated to a surprising degree the algorithim that a cockroach uses as it searches for food. As I recall, the algorithim went roughly like this.
smell for food
move forward a random number of steps
repeat until we've found food:
smell for food
if we are closer to food then we were
walk forward a random number of steps else
spin in a random direction walk forward a random number of steps
Does anybody recall the particular issue of Scientific American this article appeared in, or have a URL to any other references on this. I am attempting to implement this algorithim in a robot built with the Lego Mindstorms kit and I want to make sure I have the algorithim right.
On a related note, has anybody ever heard of the term "retreating search curve" used to describe a search for a particular point on a two dimensional surface. If so, could you please direct me to any sources you may know of?
Thank you in advance...
Darin Sunley
rsunley@escape.ca