A few days back the BBC and several other news sources covered a very neat advance, in the form of a large set of simple robots that move along to organize themselves into interesting shapes:
The original paper was published in Science, and is well worth a read.
This research is very interesting, in that it brings to a very accessible, physical level a lot of work in the area of multi-agent systems. The biological relevance is obvious: most organisms are made up of cells where each cell is working with only a small part of its control mechanisms activated, and with very little input from its environment. And yet, these cells organize themselves into wonderful structures to provide all the functionality that we take for granted. And the biological analogy is not only with cells within organisms, but also autonomous organisms – such as ants – that work in a community to build complex colonies.
Another embodiment of this idea of complex behaviours arising from simple agents is in the domain of bacterial colonies that can function in complex ways to ensure the survival of the colony in sometimes difficult conditions, sometimes be ensuring that some descendants of the colony are maintained as spores for later recolonization. At other times, the same agents work as individuals, even competing with each other for resources. I find bacteria fascinating in this regards, and I was lucky to have several collaborations with biologists where we build `robots’ – not physical robots like the ones shown in the paper above, but in the form of software agents that can be simulated.
The most fruitful of these collaborations has been in developing a spatial model of myxobacteria; this has been great but the team is now spread out and, alas, we have not been able to take this work forward.
In September I will be giving a talk about simulating microbial communities during a workshop at the Newton Institute at Cambridge. I am really looking forward to this and maybe building active collaborations in this field again.