I’ve written a fair bit – in my book, in a couple of articles here and here, and on this blog from time to time – about the extraordinary development of the new science of networks, an amalgam of research in biology, physics, systems engineering, mathematics, sociology, and other disciplines, all focused on unraveling the laws governing the growth and development of networks. So here comes a wonderful illustration of some of the connections that these researchers are uncovering, in this week’s issue of Science. A team of researchers at Hokkaido University in Japan investigated the network-creation behavior of the slime mold, an extraordinary createure that grows on forest floors in the form of an interconnected network of tube-like tendrils which transport nutrients around the organism. The researchers put out a bunch of oat flakes (a favorite slime mold food) on a platform in a pattern replicating the distribution of population centers in and around Tokyo, Japan, and then plunked the mold down in the middle and let it get to work. The idea was to see what kind of network the mold would build, connecting the various food sites together.
And lo and behold, the network of tendrils the mold — an organism, incidentally, without anything remotely resembling a “brain” — came up with bore a striking resemblance to Tokyo’s actual existing rail network. As a report in The Economist put it, the slime mold “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.” It’s quite astonishing. The Shortest Viable Network problem is mathematically quite daunting — engineers have any number of algorithms that attempt to produce the right pattern of links and connections, but it’s a tricky, difficult business. But somehow, the most primitive organism imaginable seems to “know” how to solve the problem. Fabulous! How do they do it? Good question!
wehted says:
The state of mind of an organism depends upon the externalities of its environmental situation
February 9, 2010, 8:15 pmMikhail Koulikov says:
This particular metaphor – linking slime moulds to network theory – has been used in scholarly writing at least as far back as 2003 – see Griffin, R. (2003), From slime mould to rhizome: An introduction to the groupuscular right. Patterns of Prejudice, 37(1), 27-50.
February 9, 2010, 8:40 pmSammy Finkelman says:
In both cases, you had case by case local decisions.
February 9, 2010, 9:57 pmCornellian says:
I saw that Economist article. Remarkable.
February 9, 2010, 10:03 pmanon says:
When life imitates BART
February 9, 2010, 10:56 pmNetwork Engineering and the Slime Mold: | Liberal Whoppers says:
[...] more: Network Engineering and the Slime Mold: [...]
February 10, 2010, 12:12 amzuch says:
While slime molds may have considerable intelligence, it turns out that even soap films have the smarts necessary to form a cube, a sphere, and any number of other geometric shapes.
I’ll have to read the Science article when it comes….
Cheers,
February 10, 2010, 1:08 amzuch says:
Prof. Post:
Do they? Are there pathological cases where the “network” settles on suboptimal geometries?
Analog computers, FWIW, might be better at such type problems than algorithmic approaches (particularly if exact answers are not needed). One such “computer” for such type problems is just some string and knots.
One difference between rail networks and slime molds is that we don’t plan to build all kinds of rail lines all over, and let the ones we don’t use (or need) fall into disrepair (although this may happen). Such approximation/optimisation is a different approach than are algorithmic computational solutions.
An interesting sidelight on this topic is that a slime mold is the largest living organism by far, with one in Washington covering a thousand acres.
Cheers,
February 10, 2010, 1:22 amSoronel Haetir says:
Too bad rail systems, like the slime molds themselves, tend to grow organically rather than plopping down a fully realized network all at once. The applications for slime mold analysis of rail proposals seems rather limited because of that.
February 10, 2010, 3:25 amFub says:
Word is that it’s still in a quiescent phase, trying to definitively settle whether P=NP. After that, world domination is inevitable. That’s why the park service has banned hikers carrying oatmeal.
February 10, 2010, 11:31 amMichael F. Martin says:
How do they do it? The same way we do. Blind variation and selective retention.
February 10, 2010, 1:18 pmNot My Leg says:
What’s most interesting to me is the retention of redundant systems. Most of this is (easily) explainable by a simple growth process. The slime mold grows to acquire nutrients and transport them. Less useful transport arms die, and more useful ones thrive, eventually leading to an efficient system. (I am assuming that’s how it works, but I really haven’t looked). My question is how it develops redundancies. They are obviously beneficial, but it’s not clear to me how they would form without either foresight or memory.
February 10, 2010, 2:38 pmAwesome « The Lure says:
[...] Awesome Slime mold mimics the Tokyo rail system. (h/t) [...]
February 10, 2010, 2:40 pmzuch says:
Thread winner. ;-)
Cheers,
February 10, 2010, 2:49 pmRandom Nuclear Strikes » The Brilliance of Slime Mold says:
[...] Slime mold effortlessly duplicates intricate Japanese rail network. The scary thought is that if you take an oatmeal-for-brains leftie and apply slime mold to the grey matter, you’d get a duplicate leftie brain. Maybe even a network of ‘em. [...]
February 10, 2010, 4:28 pmAlanDownunder says:
A good illustration of why real scientists don’t insist on micro-based macro.
February 10, 2010, 7:54 pm