Telephone and Internet companies turn to nature's ordered simplicity
Telephone and Internet companies are finding that autoclassification can provide simple solutions to complex network routing problems created by InfoGlut. To their delight, autoclassification permits them to harness the beauty of nature and its simple rules.

This very idea was discussed in the March 2000 issue of Scientific American. In their article, "Swarm Smarts," Eric Bonabeau and Guy Theraulaz looked at the way ants solve the problem of finding the shortest path to food using what is essentially autoclassification.

We as humans use a sophisticated mathematics called linear programming to obtain solutions to similar problems. Unfortunately, this is enormously difficult to do. Those of you who've dabbled in mathematics may remember the traveling salesman problem. Bonabeau and Theraulaz describe it this way:

The problem calls for finding the shortest route that goes through a given number of cities exactly once. This test is appealing because it is easy to formulate and yet extremely difficult to solve… The solution requires a number of computational steps that grows faster than the number of cities raised to any finite power.

Ants, however, solve this problem all the time-and they do it instinctively. According to the Scientific American article, when ants go on their treks to find and bring food back to their nest, as seen in Figure A, they leave behind them a chemical dropping that other ants follow towards the food source.

FIGURE A


When foraging for food, ants leave behind a pheromone trail for others to follow. Roll over picture for a larger image.

When there's no scent to follow, they strike out in random order. While this scent trail also helps them find their way back to the nest, it has a much more powerful utility. Other ants follow the chemical trail, so they too can find food and bring it back to the nest.

The article goes on to explain that, not surprisingly, the ants who find the food at the closest distance end up getting back to the nest earlier and thus make more trips. This makes the scent of their chemical trail more powerful than the trails of ants that find food at a longer distance. Then, if the food source dries up at the shorter distance, the chemical trail to another source eventually becomes stronger, and the ants will follow that trail.

This is Darwinian Survival at its finest, and it is an extremely simple solution to a very complex problem for mathematics to solve.

Collective intelligence
Let me give you my own example of how ants are solving even more dynamically complex problems in a yet more powerful way. Each ant in a colony has only about five or so simple details it handles when sensing how to read and react to the challenges it faces.

However, when you put a swarm of ants together, as in Figure B, they interlock their senses and react in lock-step, acting en masse through common context.

FIGURE B


Ant swarms exhibit collective intelligence. Roll over picture for a larger image.

They create a powerful collective brain that can unravel the most complex problems in very simple ways. This technique is a great way to find responsive solutions to complex, dynamic problems-the same problems that InfoGlut delivers at our feet today. Therefore, we need to refocus and take advantage of the powerful rules of simplicity that nature has afforded us.