Railways and slime moulds /Network-engineering problems ca

Show me the way to Shinjuku
FROM adhesives that mimic the feet of geckos to swimsuits modelled on shark skin, biologically inspired design has taken off in recent times. Copying nature's ideas allows people to harness the power of evolution to come up with clever products. Now a group of researchers has taken this idea a step further by using an entire living organism-a slime mould-to solve a complex problem. In this case, the challenge was to design an efficient rail network for the city of Tokyo and its outlying towns.

Slime moulds are unusual critters-neither animal, nor plant nor fungus. If they resemble anything, it is a colonial amoeba. Physarum polycephalum, the species in question, consists of a membrane-bound bag of protoplasm and, unusually, multiple nuclei. It can be found migrating across the floor of dark, damp, northern-temperate woodlands in search of food such as bacteria. It can grow into networks with a diameter of 25cm.

When P. polycephalum is foraging, it puts out protrusions of protoplasm, creates nodes and branches, and grows in the form of an interconnected network of tubes. As it explores the forest floor, it must constantly trade off the cost, efficiency and resilience of its expanding network.

Since the purpose of this activity is to link food sources together and to transport nutrients around the creature, Atsushi Tero at Hokkaido University in Japan and his colleagues wondered if slime-mould transport networks bore any resemblance to human ones. As they report in Science, they built a template with 36 oat flakes (a favoured food source) placed to represent the locations of cities in the region around Tokyo. They put P. polycephalum on Tokyo itself, and watched it go.

They found that many of the links the slime mould made bore a striking resemblance to Tokyo's existing rail network. For P. polycephalum had not simply created the shortest possible network that could connect all the cities, but had also included redundant connections that allow the creature (and the real rail network) to have resilience to the accidental breakage of any part of it. P. polycephalum's network, in other words, had similar costs, efficiencies and resiliencies to the human version.

How the creature does this is unknown, but Mark Fricker of Oxford University, who is one of Dr Tero's colleagues, speculates that the forces generated by protoplasm pulsating back-and-forth through the multinuclear cell are interpreted and used to determine which routes to reinforce, and which connections to trim.

Tokyo's is not the first transport network to be modelled in this way. A study published in December by Andrew Adamatzky and Jeff Jones of the University of the West of England used oat flakes to represent Britain's principal cities. Slime moulds modelled the motorway network of the island quite accurately, with the exception of the M6/M74 into Scotland (the creatures chose to go through Newcastle rather than past Carlisle).

Of course, neither Dr Tero nor Dr Adamatzky is suggesting that rail and road networks should be designed by slime moulds. What they are proposing is that good and complex solutions can emerge from simple rules, and that this principle might be applied elsewhere. The next thing is to discover and use these rules to enable other networks to self-organise in an "intelligent" fashion without human intervention-for example, to link up a swarm of robots exploring a dangerous environment, so that they can talk to each other and relay information back to base. The denizens of Carlisle, meanwhile, may wonder what objection slime moulds have towards their fine city.