Feeds:
Posts
Comments

Posts Tagged ‘Self Organisation’

This post is following the previous post in which i mentioned about the rules governing the motion of a bird swarm.

The video below is a breathtaking, sublime, amazing recording of thousands of starlings in a flock before roosting. This is from the UK country-side. While it is another awesome demonstration of how the “imagination” of nature can be like, it also gives a perfect example of how the entire swarm is organized and how it moves. Modern researchers are trying to imitate such emergent behavior to use in fields like robotics, data mining, internet mathematics, optimization etc etc.

How do birds exactly do this?

Here is an excerpt from a talk that i had given at a technical event in October ’07.

In a swarm if we say that there are ‘N’ number of agents, then we can say that these autonomous agents are in a way co-operating to achieve a global objective. This global objective can be better foraging, constructing shelter, serving as a defence mechanism among others. This apparent collective intelligence emerges from very simple individual agents. The actions of these agents are governed by local rules and through the interactions of the N agents the swarm achieves a global objective. A kind of “self organization” emerges in these systems. We see that there is no central controller in such cases. Swarm intelligence gives a basis which makes it possible to explore collective (or distributed) problem solving without centralized control or without the provision of a global model. [1]

The individual (but autonomous) agent does not follow directives from a central authority or work according to some global plan. As a common example, a bird in a flock, only adjusts its movements to coordinate with the movements of its flock mates or more precisely the members that are its neighbors. It simply tries to stay close to its neighbors, but avoid collisions with them. Each bird does not take commands from any leader bird since there is no lead bird. Any bird can fly anywhere in the swarm, either in the middle or the front or the back of the swarm. Swarm behavior gives the birds some distinct advantages like protection from predators, and searching for food (effectively in a swarm each bird is exploiting the eyes of every other bird). Scientists are trying to find out how these birds, fish etc move in flocks, schools in a way that appears orchestrated. A school of fish and a flock of birds move as if all the “steps” were pre planned. For one moment they are moving towards the left and in another they are darting towards the right. Among these researchers in 1987, Reynolds created a boidor bird-oid (bird like) model. This is a distributed behavioral model, to simulate the motion of a flock of birds on a computer [2]. Each boid is implemented as an agent which moves according to its own understanding of the dynamic environment. A boid observes the following rules. First, is that a boid must move away from boids that are too close, so as to reduce the chance of collisions. Second, boid must fly in the general direction that the flock is moving. Third, a boid should minimize exposure to the flock’s exterior by moving toward the perceived center of the flock. Flake later [3] added a Fourth rule, a boid should move laterally away from any boid that blocks its view. This boid model seems reasonable if we consider it from another point of view, that of it acting according to attraction and repulsion between neighbors in a flock. The repulsion relationship results in the avoidance of collisions and attraction makes the flock keep shape, i.e., copying movements of neighbors can be seen as a kind of attraction.

This is what i was talking about in the previous post. A certain set of rules is followed that would give a certain shape to the swarm, if the set is altered so will be the shape and maybe functionality as well!

Sounds nice, now how can these simple rules be modeled on a computer? They can be done using NetLogo. Here is a sample model . To get a good idea about the intricacies adjust the population, the turn angles and the vision. For playing around with the model and making your own you are instructed to go through the writeup to this model on the above page.

References:

[1] E. Bonabeau, M. Dorigo, and G. Theraulaz, Swarm Intelligence: From Natural to Artifcial Systems. NY: Oxford Univ. Press, 1999.

[2] C. Reynolds, ‘Flocks, herds, and schools: A distributed behavioral model,” Comp. Graph, vol. 21, no. 4, pp. 25{34, 1987.

[3] G. Flake, The Computational Beauty of Nature. Cambridge, MA: MIT Press, 1999.

Read Full Post »