Re: [Swarm-Modelling] Crichton's "Prey"

[Joshua O'Madadhain]

On 3 Jun 2005, at 12:12, Steve Railsback wrote:


> North, Michael wrote:
>
>
> > Steve:
> > I'd argue that collections of independent agents, interacting locally
> > with a changing set of neighbors, can learn without evolution,  
> > although
> > this depends on what you identify as the "agents."  If you define  
> > agents
> > as people, then human social systems learn this way.  However, if you
> > define agent as memes or ideas then the learning requires birth and
> > death.
>
> So- maybe if the agents themselves are pretty smart? That was  
> ambiguous in the novel- the agents were not described clearly but  
> seemed to be evolved microchips with wings and eyes.

Using terms like "smart" may be confusing in this context.  As  
Michael implied, there are basically two options, depending on  
whether the individual agents are mutable or immutable.  If each  
agent is immutable--that is, its capabilities and characteristics are  
fixed--then it can't learn, and learning can only take place by  
replacing unfit agents with better ones.  If each agent is mutable,  
then it can change its state based on its inputs (what it "sees",  
etc.)--that is, it can learn.

Joshua

 address@hidden Per Obscurius...www.ics.uci.edu/ 
~jmadden
  Joshua O'Madadhain: Information Scientist, Musician, Philosopher- 
At-Tall
 It's that moment of dawning comprehension that I live for--Bill  
Watterson
My opinions are too rational and insightful to be those of any  
organization.