Note: Where available, the PDF/Word icon below is provided to view the complete and fully formatted document
Disclaimer: The Parliamentary Library does not warrant or accept liability for the accuracy or usefulness of the transcripts. These are copied directly from the broadcaster's website.
Ant Algorithms -

NARRATION

It's never a good idea to get between an ant and its dinner. Their strategies to find the shortest
way to food are so efficient they're used to program computers. Ant behaviour inspires the software
that runs our phone networks, airport terminals and delivery systems, but no computer can cope with
an unpredictable world like an ant can.

Mark Horstman

This is a puzzle game called the Towers of Hanoi. The aim is to move this stack of discs to this
peg - but, you can only move one disk at a time and you can't stack a larger disc on top of a
smaller one. Now if you translate this puzzle to a maze, you can test it on ants. The rules may be
simple, but I gotta tell you, it's harder than it looks!

NARRATION

By mapping all the possible solutions onto a diamond-shaped maze, mathematical biologist, Chris
Reid gives his ants more than 32,000 ways to solve it.

Chris Reid

If we can look at how ants solve these problems, which are the kinds of problems they solve every
day, then perhaps we can use them in our own industry as well.

NARRATION

These are Argentine Ants, a feral species found in many kitchens. When one finds food, it drops a
chemical marker called a pheromone for others to follow. The more pheromone, the more ants are
recruited.

Chris Reid

This is an hour into the experiment so the ants have had time to set up a really strong solution
along the outside of the maze.

NARRATION

And then Chris alters their world. He blocks off the shortest path. What do the hungry ants do?

Chris Reid

They end up switching from this area of the maze which is now is strongly pheromoned but
inefficient, to this area of the maze which is a new area and a shorter solution.

Mark Horstman

And is that surprising?

Chris Reid

It is surprising they're able to do that.

NARRATION

It looks like the ants finding the food use different pheromones than the ants out exploring.

Chris Reid

What we've found is that the individual ants themselves have much better navigational abilities
than was previously thought. None of the current computer algorithms based on ants' movements
incorporate this into their system.

NARRATION

This insight into the navigation of real ants can be tested on virtual ants in a computer model.

Chris Reid

We can give them virtual pheromone which again is just lines of code. This pheromone can be laid
down behind the data packets as they travel through the network.

Mark Horstman

This experiment tells us for the first time that this species of ant is far more adaptable to
changes in its environment than previously thought. Now the question is, can we make computer
networks to mimic what the ants do?

Chris Reid

Yes, we can do that. And it could be just a matter of altering the algorithms that we already have
to include this additional information that we've found. What's more important here is that nature
has much more inspiration in store.

Topics: Nature, Technology

Reporter: Mark Horstman

Producer: Mark Horstman

Researcher: Nicky Ruscoe

Camera: Peter Sinclair

Sound: Steve Ravich

Editor: Chris Spurr

Story Contacts

Chris Reid

School of Mathermatical Biology

University of Sydney

Related Info

Optimisation in a natural system: Argentine ants solve the Towers of Hanoi

Chris Reid's PhD research