
Genetic Algorithms Digest   Monday, January 13 1992   Volume 6 : Issue 1

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

Today's Topics:
	- Looking for some information
	- Ordering information for GA proceedings and books
	- Artificial life III - Sant Fe - 1992 (lack of info)
	- "Evolutionary Programming"
	- Re: GAs and Very Fast Simulated Re-annealing
	- hardware implemenation of genetic algorithms

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

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

 Canadian AI Conference, Vancouver, (CFP 1/7)                 May 11-15, 1992
 COGANN, Combinations of GAs and NNs, @ IJCNN-92 (v5n31)      Jun 6,     1992
 10th National Conference on AI, San Jose, (CFP 1/15)         Jul 12-17, 1992
 FOGA-92, Foundations of Genetic Algorithms, Colorado (v5n32) Jul 26-29, 1992
 COG SCI 92, Cognitive Science Conference, Indiana, (v5n39)   Jul 29-1,  1992
 ECAI 92, 10th European Conference on AI (v5n13)              Aug  3-7,  1992
 Parallel Problem Solving from Nature, Brussels, (v5n29)      Sep 28-30, 1992

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

**********************************************************************
----------------------------------------------------------------------

From: vemuri@babbage.ece.uc.edu (Ram Vemuri)
Date: Thu, 19 Dec 91 14:25:36 EST
Subject: Looking for some information

   I am looking to purchase the proceedings of the Fourth ICGA, and also the
   new book edited by Greg Rawlins.  If anyone has any information about the
   price and how to purchase the above, please let me know.  Thank you.

	  - Ram Vemuri

[Editors Note:  See the next message. --Alan]

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

From: Alan C. Schultz
Date: Mon Jan 13 11:35:19 EST 1992
Subject: Ordering information for GA proceedings and books

	   This message is in response to the many requests we receive for
   ordering information on GA-related books and proceedings.  If I have left
   any out, please feel free to let me know about them.  As usual, responses
   should go to ga-list@aic.nrl.navy.mil.

	   The First and Second ICGA processings are available
   from Lawrence Erlbaum Associates, Inc., 365 Broadway, Hillsdale, New
   Jersey, 07642, (800) 926-6579

	   The Third and Fourth ICGA proceedings as well as the
   Foundations of Genetic Algorithms (Edited by Rawlins) are available from
   Morgan Kaufmann Publishers, Inc., P.O. Box 50490, Palo Alto, CA,
   94303-9953, (415) 578-9911.

	   The Handbook of Genetic Algorithms (Edited by Lawrence (Dave)
   Davis) is available from Van Nostrand Reinhold, 115 Fifth Avenue, New
   York, NY, 10003, (800) 926-2665.

	   Genetic Algorithms and Simulated Annealing (Also edited by
   Lawrence Davis) is available from Morgan Kaufmann Publishers (address
   and phone number above). 

	   Genetic Algorithms in Search, Optimization and Machine Learning,
   (David Goldberg), is available from Addison-Wesley Publishing, (617)
   944-3700.

	   Parallelism and Programming in Classifier Systems, (Stephanie
   Forrest), ia available from Morgan Kaufman Publishers (address and
   phone number above).

	   Proceedings from the First Workshop on Parallel Problem Solving
   from Nature, (Edited by Schwefel and Manner), is published by
   Springer-Verlag, 175 Fifth Avenue, New York, NY, 10010, (212) 460-1500.

	   John Holland's seminal work, Adaptation in Natural and Artificial
   Systems, is now available from MIT Press, 55 Hayward Street, Cambridge, MA
   02142, (617) 625-8569.

	   Induction: Processes of Inference, Learning , and Discovery,
   (Holland, Holyoak, Nisbett and Thagard), is available from MIT Press
   (address above).

   -- Alan C. Schultz

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

From: dorigo@ipmel2.elet.polimi.it (Marco Dorigo)
Date: Tue, 24 Dec 1991 09:41:00 -0100
Subject: Artificial life III - Sant Fe - 1992

   Why Artificial life III - Sant Fe - 1992 is not in the CALENDAR OF
   GA-RELATED ACTIVITIES?
   Anyway I would like to receive a call for paper as soon as possible.
   Thanks to anybody that can send it to me,
   Merry christmas and happy new year to everybody,

   Marco Dorigo 
   Dipartimento di Elettronica
   Politecnico di Milano
   Via Ponzio 34/5
   20133 Milano
   Italia
   Tel.  +39-2-2399-3622
   Fax. +39-2-2399-3411
   e-mail:  dorigo@ipmel2.elet.polimi.it

[Editors Note: The reason it is not in the calendar is that no one has
sent any information to ga-list or to myself!  Can someone Cc:
ga-list@aic.nrl.navy.mil when they reply to the above?]

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

From: Melanie Mitchell <melaniem@lyra.eecs.umich.edu>
Date:  Tue, 31 Dec 91 10:37:30 -0500
Subject: "Evolutionary Programming"

  I picked this up from a news group.  I'd never heard of this society before.
  Strange that there's no mention here of GAs!  

  Melanie Mitchell

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

     Date: 18 Dec 1991 23:09:35 GMT
     Organization: Naval Ocean Systems Center, San Diego
     Lines: 86
     Sender: Ward C. Page
     Message-ID: <kkvl9fINNke8@roundup.crhc.uiuc.edu>
     NNTP-Posting-Host: uicadd.csl.uiuc.edu

				ANNOUNCING

	     First Annual Conference on Evolutionary Programming

			   February 21-22, 1992

			  La Jolla Marriott Hotel
			 4240 La Jolla Village Dr.
			      La Jolla, CA 92037


	Evolutionary programming is a stochastic optimization technique that 
	can be used to address various optimization problems.  Papers
	regarding the theory and application of evolutionary programming to 
	complex problem solving will be presented.  Topics include, but are 
	not limited to, automatic control, neural network training and design, 
	system identification, adaptive	representation, path planning, 
	hierarchic evolutionary programming and the relationship between 
	evolutionary programming and other otimization methods.  Panel 
	discussions are being planned.  For additional information, write to:


			       David B. Fogel
			     ORINCON Corporation
			    9363 Towne Centre Dr.
			     San Diego, CA 92121

			     CONFERENCE CHAIRS

	  General Chairman:  David B. Fogel,  ORINCON Corporation
	  Technical Program Chairman: Wirt Atmar, AICS Research, Inc.

				SPONSORED BY:
		       Evolutionary Programming Society

			    IN COOPERATION WITH:

		     The San Diego Section of the IEEE
			    ORINCON Corporation
			 Naval Ocean Systems Center
			     AICS Research, Inc.


	  REGISTRATION INFORMATION

  Standard Registration                         Rcvd Before    On or After
  (includes all sessions and conf. proceedings) Feb. 1, 1992   Feb. 1, 1992*
  -------------------------------------------------------------------------

  Standard registration fee:                          $140           $175

  Employee or member of IEEE, ORINCON Corp.
  Naval Ocean Systems Center or AICS Research, Inc.   $105           $130

  Member of Evolutionary Programming Society**        $100           $110

  Student Registration
  (includes all sessions, does not include conf. proceedings)
  ------------------------------------------------------------------------

  Standard student registration fee:                  $10/day       $10/day

  Student Member of Evolutionary Prog. Soc. or IEEE    $5/day        $5/day

  *  Registration will also be accepted at the conference.
  ** Membership in the Evolutionary Programming Society is $40/year for
     regular members, $10/year for student members.

  Please circle your registration category and send paymet in the
  form of a check or money order ($USA only) to:

  Evolutionary Programming Society
  9363 Towne Centre Dr.
  San Diego, CA 92121
  Attn: Bill Porto, Treasurer

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

From: schraudo@cs.UCSD.EDU (Nici Schraudolph)
Date: Fri, 3 Jan 92 19:16:12 PST
Subject: Re: GAs and Very Fast Simulated Re-annealing

   Since the recent debate regarding the Rosen/Ingber paper brought up the
   GAucsd simulator's default parameter settings, I feel that as the author
   of this default mechanism I should make a few remarks regarding its
   nature.

   A well-known problem with GAs is their abundance of free parameters, and
   the lack of heuristics to determine good values for them.  Thus to this
   time there are no "boxed" GAs that can simply be thrown at any problem;
   finding good points in GA parameter space is still more an art than a
   science.  My interpretation of the Rosen/Ingber paper is therefore not so
   much a statement about VFSRA than a rather obvious one about GAs, namely
   that they can produce mediocre results when used indiscriminately.

   The GAucsd setup program naturally reflects this state of affairs:
   although the default parameter settings it suggests reflect a number of
   heuristics proposed by various GA researchers, they should not simply be
   accepted at face value, for the following reasons:

   1) Each heuristic has a limited realm of applicability: empirically
   motivated ones are limited by the function suite they are based on,
   analytically moti- vated ones by the particular set of assumptions made
   for the analysis.  The paucity of commonly used GA function suites and the
   relative infancy of GA theory is reason enough to take the suggested
   defaults with a grain of salt.

   2) Moreover, the heuristics employed by GAucsd for individual parameters
   come from a variety of sources and methodologies; they do not form a
   coherent frame- work.  The suggested defaults may therefore well be
   inconsistent with each other, not to mention the particular application at
   hand.

   3) Finally, the choice of heuristics employed is rather ad-hoc and
   reflects my personal preference at the time more than anything else.  The
   default mechanism was written primarily to set peripheral parameters such
   as the File Suffix or Report Interval; its use for critical parameters
   like the Mutation Rate is quite experimental.

   In practice I find that GAucsd defaults may span the entire range from
   good to silly and can not replace thorough knowledge of the role and
   interactions of each parameter.  As Eshelman has observed, the default
   Population Size is indeed based on Goldberg's analysis which unfortunately
   suggests huge values for longer genomes, making this one of the less
   practical heuristics.

   I welcome any suggestions for improving the defaults provided by the setup
   program; this part of GAucsd is in constant flux.  So if you know of a
   nice piece of analytical or empirical work that puts contraints on GA
   parameters, feel free to let me know.

   - Nici Schraudolph.

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

From: samal@glacier.unl.edu (Ashok Samal)
Date: Tue, 7 Jan 92 11:24:11 CST
Subject: hardware implemenation of genetic algorithms

   Could someone please send me any references to hardware implementations
   of genetic algorithms and their applications?

   Thanks in advance.

   Ashok Samal
   Dept. of Computer Science and Engineering
   University of Nebraska-Lincoln
   Lincoln, NE 68588-0115
   (402) 472-2217

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