Newsgroups: comp.ai.genetic
Path: cantaloupe.srv.cs.cmu.edu!rochester!cornellcs!newsstand.cit.cornell.edu!news.kei.com!newsfeed.internetmci.com!in1.uu.net!halon!sybase!usenet
From: George Van Treeck <treeck>
Subject: Re: Object GA's
Content-Type: text/plain; charset=us-ascii
Message-ID: <DI27tL.3wz@sybase.com>
Sender: usenet@sybase.com
Content-Transfer-Encoding: 7bit
Organization: Sybase, Inc.
References: <488d8m$brr@crc-news.doc.ca> <30A8C902.381A@stn14.me.calpoly.edu>
Mime-Version: 1.0
Date: Wed, 15 Nov 1995 00:53:44 GMT
X-Mailer: Mozilla 1.1N (X11; I; SunOS 5.3 sun4m)
X-Url: news:30A8C902.381A@stn14.me.calpoly.edu
Lines: 15

"Mark S. Johnson" <mjohnson@stn14.me.calpoly.edu> wrote:
>Could you tell me what you find out about object GAs? I have the EXACT same
>interest. I use Smalltalk, and translating back and forth to bit strings just
>feels wrong....

Only mutation and cross-over functions operate at the bit level.  All you
have to do is pass your array of genes to them.  They can internally
cast a pointer to genes to type pointer to char in C/C++.  In Pascal,
Modula, Ada, etc., you can cast the pointer to a packed array of 0..1
(boolean).  All the rest of your software should only see the abstract
definitions of the genes -- in other words there should be no "translating"
back and forth!

-George

