It seems that the choice of decentralized control also has some biological reasoning

behind it. In his famous book called Turtles, Termites and Traffic Jams, Mitchel Resnick

states

 A flock of birds sweeps across the sky. Like a well-choreographed dance troupe,
 the birds veer to the left in unison. Then, suddenly, they all dart to the right and
 swoop down toward the ground. Each movement seems perfectly coordinated. The
 flock as a whole is as graceful-maybe more graceful-than any of the birds within it.
 Most people assume that birds play a game of follow-the-leader: the bird at the
 front of the flock leads, and the others follow. But that's not so...There is no special
 "leader bird." Rather, the flock is an example of what some people call "self
 organization." Each bird in the flock follows a set of simple rules, reacting to the
 movements of the birds nearby it. Orderly flock patterns arise from these simple,
 local interactions. None of the birds has a sense of the overall flock pattern... and...
 Bird flocks are not the only things that work that way. Ant colonies, highway
 traffic, market economies, immune systems in all of these systems, patterns are
 determined not by some centralized authority but by local interactions among
 decentralized components. [Res97, p. 3]

 Thus, the importance of decentralized control and its connection to self-

organization can be understood.

1.3.2 Decentralized Self-Organization

 Here, a simple and effective way of bringing decentralized self-organization in

multiple agents through an Information Particle Interaction Model (IPIM) is proposed.

IPIM is based on the theory of physically interpreting data samples in entropy estimation

as information particles (IPCs) and the interactions between data samples as information

forces (IF) [PriOO]. This theory can be applied to the self-organization of multiple agents

by considering each robot as an IPC interacting with each other in a potential field

[Tho03]. Since this requires each robot to have only a simple transmitter and receiver, it

is shown here that this technique greatly reduces the circuit complexity and hence the

cost of the whole application.