Re: Need to talk with somebody who has dug through Breeder library and/or GW library

[Kevin S. Lim]

Hi Paul,

> Here is the other question about GAs.  I understand well how this works
> if the agent attributes are all 0 and 1.  What do you do if an agent's
> attribute is from a continuum, such as [0,1] or the like.  Are you just
> supposed to "discretize" by breaking it up into little segments and
> treat that attribute as if it were 000100000 or 0000010000 or some such
> string?

In my work with the land use change in the Amazon (LUCITA) ABM, I use 270
bit strings.  My rule strings represent the monthly labour and capital
requirements for an agent to implement a given land use strategy.  Each
monthly labour requirement is translated from base 10 to base 2 and uses 10
bits for each month.  Capital requires 20 bits to encode the largest
possible value.  So in my case my agent detectors do not sense an
"environment" stimulus that evaluates to a true or false state (1 or 0), but
rather an agent internal "environment" parameter that can't be captured
adequately with 1 bit.  I suppose if you break up your continuum described
above into small enough segments, you could approximate the continuum, but
that would probably require *a lot* of segements!

In a nutshell, my "intelligent" household agents own a property lot.  They
must decide from a set of 8 land use strategies, which one to implement
based on their available labour and capital, and their past experiences with
the rule base.  Ideally effective rules are reinforced through simulated
time and inadequate rules excluded from future decision making.  If you want
more info, check out www.fes.uwaterloo.ca/u/klim for an overview of my
Master's thesis work.

> I could be persuaded to work with GW instead, if somebody outlined the
> reasons why I would prefer it to Breeder.

I'm using Breeder and the best reason I can think of is that the design is
easy to follow and is based on an CS example described in the Booker,
Goldberg, Holland (1989) text on GA and CS.  At least, it a good basic
starting point to explore and from my experiences meet my modelling needs.

Hope this helps -- probably not as it's most likely too cryptic! =)  I'm
open to more discussion through the mailing list or directly by e-mail.
Catch ya later.

Cheers,
- Kevin

Kevin S. Lim - B.Sc.(Env) - MES Candidate
Department of Geography - University of Waterloo
www.fes.uwaterloo.ca/u/klim  ICQ#52585289




                  ==================================
   Swarm-Modelling is for discussion of Simulation and Modelling techniques
   esp. using Swarm.  For list administration needs (esp. [un]subscribing),
   please send a message to <address@hidden> with "help" in the
   body of the message.
                  ==================================