![]() |
Center Bryn Mawr College | ![]() |
October 30
Panama Geer
"Decentralization and Self-Organization:
Not Just for Ants"
Panama began by showing us a series of overheads illustrating Ant Facts, The Characteristics of Collective Intelligence, The Colony as the Organism, The Behavior of Real Ants , The Features of a Swarm System, A Centralized Mindset, The Concept of Decentralization Not New, Organizational Pyramids, Scientific Models, Theories of Self and Mind, A Single Stream of Consciousness, and Theories of Knowledge. To demonstrate the applicability of the model of an emergent system, she also ran a computer program from the Netlogo Model Library simulating the action of termites gathering wood chips into a single pile. The rules of the system were very simple, and decentralized: there was no foreman, no designated spot, no shared vision, not even a description of the task to be accomplished.
Then the questions began: Were the number of chips critical, for self-organization to work? Is scale important? Is the design approach more unwieldly, the more complicated the problems are? Or is the concept of emergence less useful, as the number of variables in the system increase? How efficient is an emergent system, compared to one that is organized from the top down? How context specific are these models? Does decentralization occur in a vacuum, independent of context, or is it affected by the features of the ecology/environment lending themselves to the activity? An approach used in models that exhibit emergence is to simply test explanations for how large scale organization evolves: What rules are sufficient to describe what will happen? What sort of complex global behavior emerge from individual choices? Do we know whether there is a class of problems that is better solved from the bottom up? One possible application of this idea might be the Centers @ Bryn Mawr. Does their success stem from charismatic leaders, or from the intersecting desires and interests of a number of different people? What are the morphological and phenomenological characteristics of a decentralized system? What is the matrix, what the value of such processes: is it organization? efficiency? Is randomness fundamental? It seems essential to a swarm system, but one could change the algorithm to make a distributed system which is completely deterministic. L-systems, for instance, are not random; they follow rules--although you may not be able to generate the rules which will give you the final structure.
Or is randomness (or irrationality? or unpredictability) key to emergent systems, because it allows something unplanned to come out? One implication of this sort of thinking is formal and practical: using emergent systems, you may be able to get a solution you can't get another way. This way of thinking, which runs against intuition (when faced with a problem, our first solutions involve hierarchies, planning, managers). also has a much more fundamental, even metaphysical dimension: every pattern may stem from randomness; even if planned, every pattern involves random elements.
But there are different permutations, different cultures, different behavioral or ecological types of emergence. On one level we can observe and understand and emulate phenomena as simple interactive systems. The structures of cities (for example, their intense segregation) is the result of simple, local interactions and decisions; the L.A. riots were an example of emergence. And attempts to create deliberate change which are unaware of such distributed dynamics are bound to fail. But to turn around and do problem solving with this kind of understanding is a task of a different order. It was suggested that anytime one is working towards a predefined goal, one has broken away from the model of a emergent system, even though it may still be a distributed one, with fundamental interactions occuring on the local level. It is human behavior to think about goals; but is it fair to differentiate between goals and rules? Are goals emergent?
We concluded with the question: do emergent systems exist on a continuum, from determinist phenomena to with some unpredictability (such as algorithms based on ant colony behavior)? It was suggested that every emergent system has the qualities of unpredictable, local, decentralized interactions, with no goal.
Our discussion next week will continue next week, when Michael Tratner
will lead us in a conversation about whether cultural studies methods can be used to
study the culture of science.
Home
| Calendar | About
| Getting Involved
| Groups | Initiatives | Bryn Mawr Home | Serendip Home
Director: Liz McCormack -
emccorma@brynmawr.edu
| Faculty Steering Committee
| Secretary: Lisa Kolonay
© 1994-
, by Center for Science in Society, Bryn Mawr College and Serendip
Last Modified:
Wednesday, 02-May-2018 11:57:05 CDT