Genetic Algorithms Digest    Wednesday, March 25, 1987    Volume 2 : Issue 1

 - Send submissions to GA-List@NRL-AIC.ARPA
 - Send administrative requests to GA-List-Request@NRL-AIC.ARPA

Today's Topics:
	- Genetic Rules for Learning Systems
	- 1987 GA Conference

------------------------------


To: GA-List@nrl-aic.arpa
Subject: Genetic Rules for Learning systems
Date: Wed, 04 Feb 87 15:48:35 +0000
From: Julian Onions <jpo%computer-science.nottingham.ac.uk@Cs.Ucl.AC.UK>


I have a few more dumb questions about genetic learning techniques.

I was watching a program (on BBC TV, England) the other night
which was about Evolution actually but included a large section
on the use of genetic algorithms. It showed several research teams
(including the Michigan group) using these techniques to solve a
variety of problems. The one I was interested in in particular was
use of genetic algorithms to govern learning rule based programs.

I have difficulty seeing how this is achieved in practice.
I would have thought that you would either need thousands
of parameters and have to set up hard wired rules to interpret
these or have the rules themselves as the parameters.

Neither of these approaches seem very effective to me, but then
I've only really thought about them, not tried doing anything.
Any advice on the subject would be very welcome.

Now a couple of trivias
Has the bibliography actuallty appeared or have I missed something?

With reference to the above, is these any practical limit to the size
of the genetic parameters (10, 1000, 100000000)?

Julian.

------------------------------


Date: Wed, 25 Mar 87 08:17:03 est
From: John Grefenstette <gref>
To: GA-List
Subject: 1987 GA Conference

This is a note of encouragement to please write up a
description of your work on genetic algorithms and submit
it to the conference.

Many people have contacted me with questions about the severity
of the April 1 deadline.  The committee will start reviewing
submissions about April 10. Papers that are much later than
that have less of a chance of acceptance.  On the other hand,
the final version need not be submitted at this time -- the
camera ready copy will be due by May 30.

Please send three hard copies to:

	John Grefenstette
	Navy Center for Applied Research in AI
	Naval Research Lab
	Washington, DC 20375

-- JJG

------------------------------


End of Genetic Algorithms Digest
********************

