Pathfinding made easier by chemical waves.

Computers can beat humans at many tasks, but they often do it by brute force. To find the quickest route through a maze, for example, a computer typically calculates every possible path and chooses the fastest. Now, researchers have found a shortcut to this computation: Just follow a chemical wave.

A chemical wave is a self-propagating reaction front that moves through certain kinds of media. Compared to physical waves, chemical waves have some odd properties: They move at a constant speed, skirt barriers without breaking up, and vanish at dead ends.

A group led by chemist Kenneth Showalter of West Virginia University in Morgantown took advantage of these properties to find the shortest route through a maze. To make a chemical wave, they soaked a small square of polymer membrane in an acidic solution of bromate ions, malonic acid, and an iron catalyst. They cut pieces from the membrane, leaving the corridors of a maze.

The researchers then touched a silver wire to one edge of their maze to set off a wave and took a video image of the wave front every 10 seconds. These snapshots, fed into a computer, gave them a composite image from which they derived a grid of vectors.

"You have a whole lot of information in this grid," Showalter says. A computer might have to do thousands of trial-and-error calculations to find the shortest path in a grid that size, he explains. But using the vector field from the chemical wave, "you can pick any point at random, and you automatically know how to get back to the starting point just by following the vector flow."

"You could call it a parallel approach," Showalter adds. "You do many things at once."

The team also simulated chemical waves in mazes on a computer. This approach to solving mazes could have practical applications, says Showalter, who describes the work with his colleagues Oliver Steinbock and Agota Toth in the Feb. 10 Science. For example, it could help a robot in a huge warehouse with hundreds of rows of shelves find the best path to ferry a package to the loading dock.

Showalter also speculates that pathfinding might play a role in the excitable media of living systems, such as the cerebral cortex.

The brain's efficiency seems to depend on the arrangement of neuronal pathways, he notes. "[Could] the fact that the cortex is an excitable medium [be] connected to optimization of these neuronal pathways?" Showalter asks. "You have to wonder."