After having read quite a bit about self-organization, mostly concerning ants and termites I want to throw in some thoughts. An interesting presentation can be found here.

Quote Originally Posted by Wiki
According to Scott Camazine.. [et al.]:

“ In biological systems self-organization is a process in which pattern at the global level of a system emerges solely from numerous interactions among the lower-level components of the system. Moreover, the rules specifying interactions among the system's components are executed using only local information, without reference to the global pattern.[7]
So far so good. A slide of the linked PPP.


Basic ingredients:

  1. Multiple interactions
  2. Amplification of fluctuations and Randomness
  3. Positive feedback (e.g., recruitment and reinforcement)
  4. Negative feedback (e.g., limited number of available foragers)

To make self-organization work you need a lot of active and responsive components which interact through a complex web communications and feedbacks. For termites this includes a "random" walk in search of food, the amazing construction of huge mounts to the highly integrated defense of it against attackers*.


Another slide.

How is self-organization achieved?

Communication is necessary:

  1. Point-to-point: antennation, trophallaxis (food or liquid exchange), mandibular contact, direct visual contact, chemical contact, . . . unicast radio contact!

  2. Broadcast-like: the signal propagates to some limited extent throughout the environment and/or is made available for a rather short time (e.g., use of lateral line in fishes to detect water waves, generic visual detection, actual radio broadcast

  3. Indirect: two individuals interact indirectly when one of them modifies the environment and the other responds to the new environment at a later time. This is called stigmergy (e.g., pheromone laying/following, post-it, web)

We see that in this case pretty much every form of conceivable communication is used to self-organize the colony. There is lot trying things out (random walk, coordinated raid), reaction to local events (scout reports food sources to the next ant), to global events (the broadcasted "alarm" signal warning termites of an attack gets broadcasted through the whole colony by "relay termites"), adaption (switching from a no longer worthwhile food source, stopping an attack on too well defended termite mount). At the core all interesting stuff, most already known, but hard to implement as it touches a lot of subjects. Due to our human nature organizational things become both a further bit more complicated and easier.


Firn


*Very interesting stuff but I do not know if it fits in.