Information Technology Reference
In-Depth Information
even though an action or movement may be performed by a different ant from the one
initially involved in it. The execution of these rules, in this case, is conducted by the
flow (cue) of an ant's pheromone.
Multiple stable interactions
arise when other ants interact
with this cue and create new stimuli for further interactions to occur.
Ants leave their nest in order to find food. Once found, they load up and return to the
nest leaving a trail of pheromone. For simplicity, assume that ants 'would raise an alarm
for other ants to follow'; while this is a multiple interactive process, it could also be a
parallel-process
as other ants from the nest may find other food sources and also raise
an alarm. However, as Holland and Melhuish (1999, p. 4) assert, '… if there are many
locations with such cues, the subtask
2
will be performed faster at the location that has
greatest numbers of agents present' due to the higher interaction rate and stability of
the process.
Adaptation
Prigogine (1996, p. 711) argues that '… self-organising systems allow adaptation to the
prevailing environment'. Comfort (1994, p. 3) explains that:
… self-organisation recognises that individual choices, communicated across
organisational frameworks, affect the operation of the wider system [and] in
this respect, voluntary selection allows individuals operating within organisa-
tional systems to cluster around points of energy that they find more attractive,
creating a 'peak' of energy distribution over repeated interactions and aligning
other members to that point in a 'basin' of attraction.
This explains why ants perform faster at the location that has greatest numbers of ants
present: they cluster around the point of more attractive energy, adapting to environ-
mental conditions (Kauffman, 1993; 1995). This actually becomes a powerful mechanism
when coupled with one of the formal characteristics of self-organisation: positive feed-
back.
In summary, self-organisation is the result of utilising the system's capacity for patterns
and structure formation; processing communication and multiple interactions by choices
or cues; and the mutual adjustment in behaviour based on a shared goal among actors
of a given system and environmental conditions. Based on these signatures, we can now
start to identify the main four characteristics of that system.
Positive and negative feedback
While most self-organising systems use positive feedback, for such systems both negative
and positive feedback are indispensable. Camazine, et al. (2001) point out that negative
feedback often takes the form of regulation, competition, reduction or saturation. Con-
tinuing with the social insect analogy: in the ants' nest negative feedback dominates
when there is competition among food sources, the food source is fully consumed, too
many ants are feeding from a food source, there are not enough food sources in a partic-
ular area, lack of space or any other similar event that overtakes the positive feedback
processes of the ants' nest. Consequently, the ants are forced to hunt for other food
sources and commence the feeding cycle again. A different example used in the biology
literature is the case of pillar formation in termite nests (e.g. Franks and Deneubourg,
1997; Camazine et al., 2001). In this event, negative feedback takes over when there is
no more material in the area close to the formation of these types of pillars. It has also
been observed that there seems to be a certain type of competition among termites
2
This is the sequence of movements making up a task; in this case: search, food and nest.
Search WWH ::
Custom Search