
Genetic Algorithms Digest    Friday, 15 June 1990    Volume 4 : Issue 9

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

Today's Topics:
	- using GA to generate hypothesis valid against historical data
	- constrained optimization
	- Classifier Systems Benchmark (testbed)
	- A GA Tutorial and a GA Short Course
	- ICGA-91 Announcement

******************************************************************************

CALENDAR OF GA-RELATED ACTIVITIES: (with GA-List issue reference)

7th Intl. Conference on Machine Learning (Austin)             Jun 21-23, 1990
Workshop Foundations of GAs (v3n19)                           Jul 15-18, 1990
Conference on Simulation of Adaptive Behavior, Paris (v3n21)  Sep 24-28, 1990
Workshop Parallel Prob Solving from Nature, W Germany (v4n5)  Oct 1-3,   1990
2nd Intl Conf on Tools for AI, Washington, DC (v4n6)          Nov 6-9,   1990
4th Intl. Conference on Genetic Algorithms (v4n9)             Jul 14-17, 1991

(Send announcements of other activities to GA-List@aic.nrl.navy.mil)

******************************************************************************

Date: Tue, 22 May 90 12:54:21 -0400
From: havener@Kodak.COM
Subject: using GA to generate hypothesis valid against historical data

  I am a GA neophyte and can not contribute much in the way of
  advances in the state of the art. My job is to learn this stuff and
  make it simple enough to apply today to problems in my company. I am
  very interested in [the] concept learner package [reference to
  v4n8 article by Bill Spears], and would appreciate being included
  in a mailing list of any information generated on this 
  project that you can share.

  I need to apply GA to learn from a vast database of historical
  readings. One tough part is to design a payoff function that will
  reward a GA for generating a hypothesis that can be demonstrated
  valid against the historical data.

  Example:
  Goal:  explain occasions when product attribute X > value Y
  GA generated hypothesis:  That happens when variable Z > a constant C.

  Can you recommend any reference material on how to design such a
  GA and especially the associated payoff function (what has been tried
  and what has failed), or any work associated with GA's applied to mine
  historical data?

  [Deleted text about sending paper - ACS]

  John P. Havener -  Advanced Chemical Engineer - TEX Core AI Group -
  Eastman Kodak, Texas Eastman Co. -  B1, Box 7444 - Longview TX   75607
  Net: havener@Kodak.com - Phone:(214)237-6368 - Fax: (214)237-5371 Automatic  

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

Date: Mon, 4 Jun 90 11:18:43 cdt
From: Steven A. Harp <harp@hi-csc.honeywell.com>
Subject: constrained optimization

  I have a constrained optimization problem with characteristics that I hope
  will ring some bells with the GA intelligentsia. 

  The problem domain is statistical model fitting. A typical instance of the
  problem would involve 6 variables and nine constraints. (I would ideally like
  to handle problems 4 times this size.)  Both the objective function and most
  of the constraints (inequalities all) are nonlinear. Neither the objective
  function nor the constraints are difficult to compute--in fact, they're
  differentiable.  Despite this fact, several standard techniques including
  flexible tolerance, barrier function, and augmented lagrangian have proved to
  be very unreliable in finding minima. The existence of a unique minimum is an
  open theoretical question. (The GA would be useful even if all it could do
  would be to shed some light on this question.)

  Constrained genetic optimization techniques I have read about employ penalty
  function methods. This is not such a good strategy with my problem since the
  objective function is undefined outside the feasible region (it tends to blow
  up). One could arbitrarily assign draconian penalties to all nonfeasible
  points, however, it has been noted (Richardson et al. GA89) that this 
  approach is rather wasteful of information.

  QUESTION: Are there any results, published or otherwise, that describe GA
  optimization strategies that only examine feasible points? 

  Any "feasible" suggestions would be welcome.

  Steven Alex Harp,  harp@hi-csc.honeywell.com

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

Date: Wed, 13 Jun 90 18:31 N
From: "MARCO DORIGO" <DORI@IPMEL1.POLIMI.IT>
Subject: Classifier Systems Benchmark (testbed)

    I would like to know if there is some standard set of problems to
    use as a benchmark for Classifier Systems.
    Thank you to anybody that will help me,
    Marco Dorigo

    Marco Dorigo
    Dipartimento di Elettronica
    Politecnico di Milano
    Via Ponzio 34/5
    20133 Milano
    Italia
    Tel.  +39-2-2399-3622
    e-mail:  dori%ipmel1.polimi.it@iboinfn.bitnet

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

Date:         Thu, 07 Jun 90 15:53:03 CDT
From: "Dave Goldberg (dgoldber@ua1vm.ua.edu)" <DGOLDBER@UA1VM.ua.edu>
Subject:      A GA Tutorial and a GA Short Course

    A tutorial entitled "Genetic Algorithms and Classifier Systems" will be
    presented on Wednesday afternoon, August 1, at the AAAI conference in
    Boston, MA by David E. Goldberg (Alabama) and John R. Koza (Stanford).
    The course will survey GA mechanics, power, applications,
    and advances together with similar information regarding classifier
    systems and other genetics-based machine learning systems.  For further
    information regarding this tutorial write to AAAI-90, Burgess Drive,
    Menlo Park, CA 94025, (415)328-3123.

    A five-day short course entitled "Genetic Algorithms in Search,
    Optimization, and Machine Learning" will be presented at Stanford
    University's Western Institute in Computer Science on August 6-10
    by David E. Goldberg (Alabama) and John R. Koza (Stanford).
    The course presents in-depth coverage of GA mechanics, theory and
    application in search, optimization, and machine learning.  Students
    will be encouraged to solve their own problems in hands-on computer
    workshops monitored by the course instructors.  For further information
    regarding this course contact Joleen Barnhill, Western Institute in
    Computer Science, PO Box 1238, Magalia, CA 95954, (916)873-0576.

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

Date: Thu, 14 Jun 90 14:42:39 EDT
From: dejong@AIC.NRL.Navy.Mil
Subject: ICGA-91 announcement

   Here's what you've all been waiting for.  I hope you'll all
   contribute in one form or another to making it a success.  And
   special thanks are in order to Rik Belew for hosting it.

   Ken

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

				Announcing


				 ICGA-91

		    The Fourth International Conference on
 			     Genetic Algorithms


			      July 14-17, 1991

		     University of California at San Diego




ICGA-91 Conference Committee:

	Conference Chair:	Kenneth A. De Jong, George Mason University
				J. David Schaffer, Philips Labs

	Vice Chair:		David E. Goldberg, University of Alabama

	Program Chair:		Richard K. Belew, UCSD
				Lashon Booker, Naval Research Lab

	Publicity Chair:	David E. Goldberg, University of Alabama

	Financial Chair:	Gil Syswerda, BBN

	Local Arrangements:	Richard K. Belew, UCSD



Tentative Submission Date:	February 1, 1991


For further information contact:

	Dr. Richard K. Belew
	Computer Science & Engr. Sept. (C-014)
	University of California at San Diego
	La Jolla, CA  92093
	rik@cs.ucsd.edu
	(619) 534-5288
  or
	Dr. Lashon Booker
	NCARAI - Code 5510
	Naval Research Laboratory
	Washington, D.C.  20375-5000
	booker@aic.nrl.navy.mil
	(202) 767-2382

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