
Genetic Algorithms Digest   Monday, March 11 1991   Volume 5 : Issue 4

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

Today's Topics:

	- tutorial on GAs at TRANSPUTER 91
	- Technical report available:  The Evolution of Learning
	- A question about a graycode algorithm (to John Grefenstette)...
	- ...and a reply from John about a greycode algorithm.
	- genetic fusion algorithm?
	- Parallel Implementations of Classifier Systems

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

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

4th Intl. Conference on Genetic Algorithms (v4n17)          Jul 14-17, 1991
Special Issue of MLJ on Reinforcement Learning - (v5n2)

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

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

Date: Mon, 28 Jan 91 13:46:57 MST
From: stiles@multi.ee.usu.edu (G. S. Stiles)
Subject: tutorial on GAs at TRANSPUTER 91

   [ Ed.'s Note:  We received an announcement for a conference called
     TRANSPUTING '91.  For the most part, the announcement did not seem
     apropos to GAs, and so I had trashed it.  Then I received a follow-
     up on the tutorials they were planning, and I found one tutorial
     that is of interest.  So here is information on the tutorial, and
     if you want more information, contact the person above, or one
     of the presenters of the tutorial. ]


                  TRANSPUTING '91 tutorials
                  =========================

		       Roger M.A. Peel
		     University of Surrey
		      10th January, 1991

    Artificial Intelligence
    =======================
    J. Stender & E. Hillebrand (Brainware GmbH, Germany),
    S. Forrest (University of New Mexico, U.S.A.),
    D. Macfarlane & I. East (University of Buckingham, U.K.)

    The morning session of the workshop will provide an introduction
    and overview of parallel genetic algorithms, and it will discuss
    fine-grained parallel genetic algorithms.  The session will be
    divided into three roughly equal parts:

    1. Introduction and overview of (sequential) genetic algorithms

    -    introduction (biology or computer science?)
    -    mechanical details of the algorithm
    -    why genetic algorithms work
    -    example applications

    2. Theory of parallel genetic algorithms

    -    implicit parallelism
    -    explicit parallelism of population subdivision
    -    interactions among subpopulations
    -    explicit fine-grained parallelism

    3. Fine-grained parallel genetic algorithms

    -    one processor per individual model
    -    replication strategies (conventional, location-based,
	 resource-based, pattern-based)
    -    implementation issues
    -    research questions
    -    example applications

    The afternoon session of the workshop will focus primarily on the
    practical aspects of implementing genetic algorithms on transputer
    networks.  Many of the issues raised, such as scalability, load
    balancing, efficiency and deadlock freedom, have more general
    application in parallel processing.

    The session will be divided into four parts of approximately equal
    length.

    -  General issues in scalable efficient implementations of
       fine-grainedparallel genetic algorithms on coarse grained
       transputer networks.

    -  Three parallel genetic algorithms, farming, migration and diffusion,
       and their implementation on transputer networks.

    -  Applications and studies of parallel genetic algorithms.  Evaluation
       and comparison of the different models and implementations.

    -  Demonstration of implementations and visualisation software.
       Question time.

    The workshop will draw on work done in Europe using transputers and
    parallel genetic algorithms and in particular the ongoing research
    at Buckingham Univerity in the UK and GMD in Germany.  Two major
    applications of parallel genetic algorithms will be described, the
    solution of TSP problems and the search for application specific
    structured neural networks.

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

Date: Tue, 19 Feb 91 19:46:12 EST
From: David Chalmers <dave@cogsci.indiana.edu>
Subject: Technical report available:  The Evolution of Learning

    The following paper is available electronically from the Center for
    Research on Concepts and Cognition at Indiana University.

       THE EVOLUTION OF LEARNING:  AN EXPERIMENT IN GENETIC CONNECTIONISM

				David J. Chalmers

		  Center for Research on Concepts and Cognition
			       Indiana University
				   CRCC-TR-47

    This paper explores how an evolutionary process can produce systems that
    learn.  A general framework for the evolution of learning is outlined,
    and is applied to the task of evolving mechanisms suitable for
    supervised learning in single-layer neural networks.  Dynamic properties
    of a network's information-processing capacity are encoded genetically,
    and these properties are subjected to selective pressure based on their
    success in producing adaptive behavior in diverse environments.  As a
    result of selection and genetic recombination, various successful
    learning mechanisms evolve, including the well-known delta rule.  The
    effect of environmental diversity on the evolution of learning is
    investigated, and the role of different kinds of emergent phenomena in
    genetic and connectionist systems is discussed.

    A version of this paper appears in _Proceedings of the 1990 Connectionist
    Models Summer School_ (Touretzky, Elman, Sejnowski and Hinton, eds.)


    This paper may be retrieved by anonymous ftp from cogsci.indiana.edu
    (129.79.238.6).  The file is chalmers.evolution.ps.Z, in the directory
    pub.  To retrieve, do the following:

    unix-1> ftp cogsci.indiana.edu           # (or ftp 129.79.238.6)
    Connected to cogsci.indiana.edu
    Name (cogsci.indiana.edu:): anonymous
    331 Guest login ok, sent ident as password.
    Password: [identification]
    230 Guest login ok, access restrictions apply.
    ftp> cd pub
    ftp> binary
    ftp> get chalmers.evolution.ps.Z
    ftp> quit
    unix-2> uncompress chalmers.evolution.ps.Z
    unix-3> lpr -P(your_local_postscript_printer) chalmers.evolution.ps

    If you do not have access to ftp, hardcopies may be obtained by sending
    e-mail to dave@cogsci.indiana.edu.

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

Date: Tue, 05 Feb 91 08:10:32 CST
From: AARON KONSTAM <AKONSTAM%TRINITY.BITNET@ricevm1.rice.edu>
Originally To: John Grefenstette <gref@AIC.NRL.Navy.Mil>
Subject: A question about a graycode algorithm

    In the GENESIS package there are functions that implement algorithms
    for conversion between graycode and binary numbers and visi versa.
    These algorithms work but I can find no reference to them anywhere
    nor an explanation of why they work. I have consulted with
    our math department to no avail. Could you provide a reference to
    the origin of these algorithms or a reference to the derivation of these
    algorithms.
    Thanks.

    AARON KONSTAM
    Trinity University
    715 Stadium Dr.
    SAN ANTONIO, TX 78212
    (512)-736-7484
    AKONSTAM@TRINITY.BITNET

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

Date: Wed, 20 Feb 91 07:25:14 EST
From: gref@AIC.NRL.Navy.Mil
Originally To: AKONSTAM%TRINITY.BITNET@ricevm1.rice.edu
Subject: Re:  A question about a graycode algorithm, John's response...


    The Gray code used in GENESIS is a reflective Gray code (there are many
    others).  Gray codes have the property that codings for adjacent
    integers differ by exactly one bit, and so they seem appropriate for
    GAs. I think they were first used in GAs by Holland's student Hollstein
    in about 1971 (see Goldberg's book).  I don't have a reference handy,
    but you should be able to find a discussion of Gray codes in a text book
    on information theory or data communication.  I frankly don't remember
    where the algorithm in GENESIS came from.  Either I made it up or it was
    suggested by one of my students, probably Dave Schaffer.

    I hope this helps.

    - John Grefenstette

[Ed's Note:  Can anyone else contribute to this??]

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

Date: Thu, 21 Feb 91 08:31:26 EST
From: ds1@philabs.Philips.Com  (Dave Schaffer)
Subject: genetic fusion algorithm?

    An article appeared in the Feb 18, 1991 issue
    of the Electronic Engineering Times (p37)
    It describes a "genetic fusion" algorithm
    developed by Kunihiko Kaneko of the University of
    Tokyo and Takashi Ikegami of Kobe University
    that has enabled them to optimize algorithms
    that "slash CPU processing time by a factor of
    30 over existing algorithms."

    The description sounds very GA-like. Anyone
    know about this? Can technical publications
    be cited?

    Dave Schaffer

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

Date: Mon, 4 Mar 91 14:10:47 EST
From: kirk@cat.syr.edu (Kirk Twardowski)
Subject: Parallel Implementations of Classifier Systems

       I am currently investigating associative architectures for the 
    implementation of classifier systems and would like to hear from 
    anyone involved in similar work ( any non-uniprocessor implementation ).
    I would also be interested in any information concerning the 
    computational requirements or a complexity analysis of classifier 
    systems.
       I will summarize and repost if there is an interest.

    Kirk Twardowski            kirk@cat.syr.edu
    2-212 Center for Science and Technology
    Syracuse University    
    Syracuse, NY 13244

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