Re: Classifier systems

[Rick Riolo]

We at PSCS at UM almost have a classifier system 
ready for alpha testing.  Its being implemented by
Wojciech Kopczuk (with me trying to add helpful
advice but mostly just kitbitzing ;-).
I don't recall exactly what
features holland describes in Hidden Order, but this
will include all the basic features (detectors,effectors,
message list, bucket brigade, GA, other rule discovery
operators), plus some support for (amoung other things)
Stewart Wilson's "macro classifiers" and the capability
to create many instances of full classifier systems,
sending/receiving messages to/from each other as well 
as to/from some wider "enviroment", of course.  
We also hope it will be fairly easy for people to
modify many aspects of the basic architecture
(e.g., number and syntax/semantics of conditions).
So...stay tuned.
 - r

Rick Riolo                       address@hidden
Program for Study of Complex Systems (PSCS)
1061 Randall Lab     University of Michigan
Ann Arbor MI 48109-1120
http://pscs.physics.lsa.umich.edu/PEOPLE/rlr-home.html

On Thu, 13 Mar 1997, Tony Bigbee wrote:

> Date: Thu, 13 Mar 1997 21:31:07 -0500
> From: Tony Bigbee <address@hidden>
> To: address@hidden
> Subject: Classifier systems
> 
> Hi all,
> 
> Has anyone worked on, is working on, or will be working on
> implementation of a classifier system?  Specifically, I'm looking to
> discuss classifier systems as conceptualized by Holland ("Hidden Order")
> and the the bucket brigade algorithm.
> 
> I've searched through the list archives and didn't find anything
> relevant, but I could swear someone mentioned an interest in this area.
> 
> Tony
> 
>