Newsgroups: comp.ai.genetic
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!delmarva.com!newsfeed.internetmci.com!howland.reston.ans.net!torn!news.bc.net!unixg.ubc.ca!van-bc!vanbc.wimsey.com!news.cyberstore.ca!skypoint.com!umn.edu!news
From: "James A. Larson" <larso171@maroon.tc.umn.edu>
Subject: Re: GA's using Mathematica
To: fjfz@public.sta.net.cn
Message-ID: <27950.larso171@maroon.tc.umn.edu>
X-Minuet-Version: Minuet1.0_Beta_16
Sender: news@news.cis.umn.edu (Usenet News Administration)
Nntp-Posting-Host: dialup-3-192.gw.umn.edu
X-Popmail-Charset: English
Organization: University of Minnesota, Twin Cities
Date: Sun, 3 Sep 1995 09:26:14 GMT
Lines: 40

>: Hi,  does anyone know of genetic algorithm codes written in Mathematica
>: and where they may be found?

Ray,

There are a few that I'm aware of, but most of them are real weak -- basic 
SGA (Simple Genetic Algorithm like that presented in Goldberg's 1989 
book).  

The one exception might be Nachbar's Genetic Programming (GP) one, 
available on Mathsource, and with an article in the current issue of The 
Mathematica Journal.  While Genetic Programming is different than Genetic 
Algorithms (GA), in that GP modifies programmatic expressions rather than 
strings, they are otherwise very similar.  So GP code can be converted to 
GA code with relatively little effort.  

My information is several months to a year out date except what I pick up 
by hearsay (such as the very recent Nachbar article/code).  You might also 
ask on the Mathematica newsgroup comp.soft-sys.math.mathematica.  

If you are interested in a sophisticated one, I have one 90% completed that 
has multiple populations, gobs of statistics, convergence monitors,  and 
a lot of other bells and whistles that I feel I needed to solve a difficult 
combinatorial optimization problem.  I may finish it soon, though, because 
it has some features that I plan to add to Sugal (discussed below), and 
Mathematica is an easier development / test environment for new algorithms.

I put my Mathematica code aside for a month ago to look at SugalV2.0, a C 
program that comes in various flavors: X-Windows, Windows, and portable 
non-GUI.  The Windows version has a very nice interface, many options, 
built for extensibility, much much more.  I feel it is a more promising 
direction for me than to finish my Mathematica code, particularly since C 
runs 100 - 200X faster than Mathematica code for heavily numeric procedural 
kind of code.  That kind of speed difference is ultra-critical for solving 
all but small text-book kinds of problems.

Regards,

Jim Larson

