Credit: Malcolm Ricketts
With a rudimentary brain and no memory, individual ants aren’t particularly clever, but in groups they've shown the ability to solve complex problems.
Now as researchers gain a deeper understanding of the processes used by these tiny insects, programming engineers are taking notice, using them as inspiration for computer algorithms.
Chris Reid, a PhD student at the University of Sydney, shows me a photo of a diamond-shaped maze full of Y-shaped branching paths. It’s a spatial map of all possible solutions to an ancient toy puzzle, the Towers of Hanoi (see box at the end of the article).
A faint dotted trail of Argentine ants (Linepithema humile) traces the shortest path from one end of the maze to food at the other end. By finding the shortest path, the ants had solved the toy puzzle in the fewest possible moves.
I’m impressed, but Reid seems rather unsurprised by the ants’ feat. “Ants are particularly good problem solvers because of the pheromone system they use for navigation,” he explains.
As ants initially explore a new area they leave trails of pheromones, volatile chemicals that mark paths for other ants to follow. At first the ants create many trails through the maze, but over time the ants refine their routes via a pruning process.
“Less optimal trails are pruned away as the pheromone evaporates quickly on the longer, less-used paths,” says Reid, who published his findings in the Journal of Experimental Biology. Those paths that are maintained are the shortest path connections, where the pheromone trail is constantly reinforced by high levels of ant traffic.
The same pheromone system allows Argentine ants to construct minimum-distance networks between nest sites. In the wild, these trail networks allow colonies to quickly recruit large numbers of ants to travel to food sources and defend nests from predators.
“Argentine ants end up forming massive supercolonies containing hundreds or thousands of nests connected to each other by trails. There’s a colony in Europe that spans over 6,000 km!” explains Tanya Latty from the University of Sydney, who published her study in the Journal of the Royal Society this year.
“An individual ant has no idea where it’s going," she says. "They don’t know that they are constructing an optimal route or network. These networks are just an emergent property of all these individuals following what are probably really simple behaviour rules.”
“The ants need to solve the same problems that a transport engineer might face. “How can I connect all of these separate points using the minimum length of trail or road and minimum resources, but still achieve a network that is robust and that isn’t going to crash and become disconnected?” It’s interesting to see how natural systems solve the same problem, given that there is no centralised control,” says Latty.
Tower of Hanoi Problem
In your article, you state:
"Legend has it that Brahmin priests are working on a 64-disc version of the puzzle. When they have solved the puzzle, the legend says that the world will end."
The number of moves to solve the puzzle is 2^n - 1. If the priests were very quick and able to move one disc every second, it would take 2^64 - 1 seconds. This is approximately 18.4 quintillion seconds, which is over 585 billion years.
That certainly puts my mind at ease about the end of the world in the near future. :-)