Termites

So far we have seen pattern formation (self-organisation) in several physical, chemical and computer-simulated systems exhibiting several features which seem to be similar to those observed in living systems. Let us now approach the problem from the reverse direction: We will now try to model actual patterns seen in some living systems. The term "autonomous agent" is often used to describe a sub-unit of a large system that acts independently, except for following some simple rules of interaction with the other agents and/or the environment. An example of an autonomous agent might be an ant, an individual stock buyer, or a gene. When multiple agents interact, one often observes emergent behaviour or dynamic order that is not a property of an individual agent or the rules. We will begin with an example in which self-organisation is manifested even when the multiple agents do not directly interact with each other. Consider a termite like agent following the following three simple rules (Resnick):

1. It moves via a random walk until it collides with a wood chip.
2. If the termite is not carrying a wood chip, it picks up the one it bumped into and continues its random walk.
3. If on the other hand the termite is carrying a wood chip when it bumps into another, it drops the chip and continues its random walk.

Note that although the movement of the termites is random, the rules for picking or dropping the wood-chips is completely deterministic. The figures below show the result of a cellular automaton simulation in which starting with an intial random configuration of wood-chips, a small number of termites arranges them into clusters. Though acting independently and locally, with no global plan, nor a leader, nor any external direction, the termites' effort has given rise to order! The emergence of order or self-organisation in this case can be made more realistic by changing the rules slightly - see Ref.[1].