Worth noting that the search algorithm, simulated annealing, is a metaheuristic algorithm - meaning, it's modelled after an existing natural phenomenon (in this case, metallurgical annealing). Other metaheuristic algorithms familiar to HN readers include genetic algorithms and ant colony optimization algorithms. Given the nature of these algorithms it isn't shocking to see them in the wild, but it is interesting.

It might even be more accurate to say the model of metallurgical annealing used in simulated annealing has parallels with a model of jellyfish search behaviour, rather than jellyfish "using" simulated annealing to find food.

The nervous system of jellyfish, the oldest multi-organ animals (0.5 billion years) [1], is primitive. The neurons are arranged into a nerve net with hints of limited central actions. It doesn't surprise me that something akin to annealing is taking place since that form of search is very simple and could be imagined to take place with very simple distributed actions, having just one global, time-varying parameter T, known as the temperature [2].

Wouldn't almost any other effective search technique be more complex to implement in a very simple nerve net? If jellyfish were found to utilize say genetic optimization based search then I would really be surprised.

For more information (than you probably want to know) about the nervous system of jellyfish see [3].

[1] http://societyofbiologyblog.org/the-peculiarities-of-the-jel...

[2] http://en.wikipedia.org/wiki/Simulated_annealing

[3] http://jeb.biologists.org/content/214/8/1215.full

I'm not getting what this "novel strategy" is. The jellyfish ride currents back and forth in search of food... so? Does anyone have an article with actual info about the strategy?

How to think backwards about nature and science. Jellyfish do not "use" formal mathematical models. They do stuff. Our mathematical models may be good approximations of what they do.