A short explanation of what this computer model is all about
(adapted from "Oscillations and Chaos in Ant Societies", by R.V. Sole, O. Miramontes, and B.C. Goodwin, J. Theor. Biol. 161, pp.343-357, 1993).
The experimental basis of the model: "There is experimental evidence that ant behaviour can be switched as a consequence of worker interaction. The activity level of nearest ants can switch the state of an individual from active to inactive."
How the ant colony and the interaction is displayed in the computer model: "The square space (on the left of this webpage) represents the ants nest. Black dots and grey dots represent, respectively, active and inactive ants."
How the ants interactions are modelled: "If the ant is inactive or all nearest points are occupied, no movement takes place."
What one can observe by experimenting with the computer model: "Some kind of collective emergent behaviour will be observed as the result of local coupling. In neural organizations, retrieval of associative memory (and maybe consciousness) can be thought of as emergent properties. In ant societies, such new properties are, for example, nest building, trail formation or foraging behaviour."
How experimental data on ant behaviour is confluent with the computer model: "The implicit richness of chaotic behaviour can provide a given (simple) individual the necessary diversity of behaviour to explore, or avoid predators. In fact, recent studies have shown how chaotic elements can perform efficient parallel synchronous computation. For our system under study, the existence of social interactions produces an additional degree of complexity. Nevertheless, we will see that non-linear chaotic individuals lead to predictable behaviour under spatial interaction."
In an interview with David Suzuki, Brian Goodwin explains what experimental evidence and the computer model demonstrate about the relationship of order and chaos in ant behaviour: "The rhythmic behaviour was observed first by an experimentalist. He just saw this and he said, 'Well, that's interesting, how does it come about?' And the first assumption was, each ant is individually rhythmic, periodic. And they just synchronize, which is quite an easy process, we understand that. But then somebody else made a detailed investigation of the behaviour of individual ants and found that they're not rhythmic, they're chaotic. So then we just said, well, I wonder if we can get the rhythm, chaotic individuals interacting with this positive stimulation, will a rhythm come out of that? We didn't know and there's no theorem that would tell us that that would happen, but it did. And so we said, okay, there's our emergent property, an unexpected dynamic feature which actually serves the colony. Because that rhythmic behaviour is in fact more efficient at tending the queen and the brood than chaotic individual behaviour which would be rather sporadic. The young wouldn't get as good, even attention if they were chaotic."
The Java Applet on this web page is "Ant Society" by Akira Kageyama based on
"Oscillations and Chaos in Ant Societies", by R.V. Sole, O. Miramontes, and B.C. Goodwin, J. Theor. Biol. 161, pp. 343-357
Here are some details about the inner workings of this program:
- Periodic boundary condition.
- Interaction matrix J = {1,1,1,0}
- Interaction parameter g = 0.1 or 0.2
- Activity threshold theta = 1.e-9
- Activation probablity 0.01
- Single-ant-to-single-ant interaction with self-interaction
- Random walk in 4 (north, south, east, and west) directions
|
