Einstein Science Reporting for Kids
[ E-mail ]
21-Jan-2010

Contact: Science Press Package
scipak@aaas.org
202-326-6440
American Association for the Advancement of Science

What can we learn from a slime mold?



Physarum was placed where Tokyo would be, bounded by the Pacific Ocean (white border) and surrounded by food sources (white dots) placed where other major cities in the region would be. The slime mold grew out from the initial food source and colonized each of the other food sources. Eventually, it resolved into a network of tubes, much like the Tokyo rail system, interconnecting the food sources.
[Image courtesy of Science/AAAS]

Recent research suggests that human engineers could learn a lot from the lowly slime mold, known as Physarum polycephalum. It seems that the gelatinous, fungus-like mold might actually lead the way to more reliable computer and mobile communication networks in the future.

This finding comes after a team of Japanese and British researchers observed that the slime mold connected itself to scattered food sources in a design that was nearly identical to Tokyo's rail system.



Network formation in Physarum polycephalum.
[Image courtesy of Science/AAAS]

Researcher Atsushi Tero and colleagues placed oat flakes on a wet surface in locations that corresponded to the cities surrounding Tokyo, and allowed the Physarum polycephalum mold to grow outwards from the center. They watched the slime mold organize itself, spread out, and form a network around the food sources that was very comparable in efficiency, reliability, and cost to the real-world train network that connects Tokyo to its surrounding cities.

The researchers then captured the core strategies used by the slime mold to connect its food sources in such an efficient manner, and incorporated them into a mathematical formula. Since the slime mold has been through countless rounds of evolutionary selection, this formula based on its feeding habits could provide a route to highly efficient network designs for humans as well.

Tero and colleagues say that their model is a good starting point for improving the efficiency and decreasing the costs of self-organized networksólike phone or computer networksóthat don't have any centralized control.

###