Phil Henshaw wrote: > So I guess you're saying that if you were to make your models to be > consistent with nature, i.e. have agents that all develop their own > parameters as they go, then it couldn't be 'described' or 'reproduced'. > That sounds like a neat way to state the difficulty of using single > self-consistent ideas to represent a multiplicity of independently behaving > things. > The meaning of the word `parameters' here is a bit muddy. It could be some array of numbers that characterize the behavior of certain agents. Let's say, that certain set of agents use a lot more energy than others but without saying why (and that the other agents have a different array of numbers). Or it could be a set of arbitrary rules added to some agents that result in them using more energy. Or it could be a simple rule that in all agents, when evaluated over and over in a different shared circumstance results in lots of energy use. By an adaptive model, I mean the latter, and by parameters I'm thinking of the former two cases. If a modeler introduces more and more degrees of freedom, they can get any answer you want.
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org
