
Genetic Algorithms Digest   Monday, April 1 1991   Volume 5 : Issue 6

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

Today's Topics:
	- Typo in Handbook of GAs?
	- Info on TABU search wanted
	- Cognitive Science at Birmingham
	- Reports and Reprints Available (Goldberg)
	- Thesis available: Genetic Neural Networks on MIMD Computers

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

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

4th Intl. Conference on Genetic Algorithms (v4n17)           Jul 13-16, 1991
AAAI 91, National Conference on AI, Anaheim, CA              Jul 14-19, 1991
IJCAI 91, International Joint Conference on AI, Sydney, AU   Aug 25-30, 1991
AISB 91, Leeds, UK (v5n5)                                    Apr 16-19, 1991

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

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

Date: Sat, 16 Mar 91 11:52:21 -0500
From: folta@cs.UMD.EDU (Wayne Folta)
Subject: Typo in Handbook of GAs?

   I just gave Lawrence Davis' Handbook of GAs to a professor who is a
   skeptic about GAs. Unfortunately, in pages 9 and 10, he found a typo
   regarding the function f6. This set a really negative tone (for him) for
   the rest of the book.

				  (sin (sqrt (x^2 + y^2)))^2
   Page 9 claims that f6 is 0.5 - --------------------------
				  1.0 + 0.001(x^2 + y^2)^2

   Unfortunately, this value will never go above 0.5, since the numerator is
   squared. The professor assumed that the square of the numerator was a
   typo, but then other things were wrong.

				     (sin (sqrt (x^2 + y^2))^2 - 0.5
   I looked in OOGA, and f6 is 0.5 + -------------------------------
				     (1.0 + 0.001(x^2 + y^2))^2

   Since the maximal value in the graph on page 10 is 1, and this is achieved
   at x = 0, and since f6 is symmetric in x and y, I assume that the maximal
   value of f6 should be 1 at (0, 0). Unfortunately, the above function
   yields 0. If you change the + to a -, you do get 1 at (0, 0), which
   matches the graph on page 10 (I think--I don't have Mathmatica).

   My guess is that the graph on pg. 10 should be labelled F6 (not f6) and
   that the OOGA f6 is then correct. Is this true?

   (I know that this is a trivial point, unfortunately, if it is an error, it
   is the first hard-math item in the book and my skeptic professor zoomed in
   on it and got a feeling of slopiness. Of course, it is his own fault that
   he let a typo cloud his picture of GAs, but maybe that was due to his
   having the flu at the time.)

   Wayne

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

Date: Mon, 18 Mar 91 09:48:27 +0100
From: Carlos Lucasius <lucasius@sci.kun.nl>
Subject: Info on TABU search wanted

    Recently I've heard a lot about TABU search and I am eager to know more
    about it.  Could someone provide any pointers to this search technique?
    Thank you very much in advance!
    Carlos

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

Date: Thu, 21 Mar 91 15:18:09 GMT
From: PetersonDM@computer-science.birmingham.ac.uk
Subject: Cognitive Science at Birmingham


			     University of Birmingham 

		       Graduate Studies in COGNITIVE SCIENCE 

    The Cognitive Science Research Centre at the University of Birmingham
    comprises staff from the Departments/Schools of Psychology, Computer
    Science, Philosophy and Linguistics, and supports teaching and research in
    the inter-disciplinary investigation of mind and cognition. The Centre
    offers both MSc and PhD programmes.

			    MSc in Cognitive Science 

    The MSc programme is a 12 month conversion course, including a 4 month
    supervised project. The course places a particular stress on the relation
    between biological and computational architectures.
	    Compulsory courses: AI Programming, Overview of Cognitive
    Science, Knowledge Representation Inference and Expert Systems,
    General Linguistics, Human Information Processing, Structures for Data
    and Knowledge, Philosophical Questions in Cognitive Science,
    Human-Computer Interaction, Biological and Computational
    Architectures, The Computer and the Mind, Current Issues in Cognitive
    Science.
	    Option courses: Artificial and Natural Perceptual Systems,
    Speech and Natural Language, Parallel Distributed Processing.
	    It is expected that students will have a good degree in
    psychology, computing, philosophy or linguistics.
	    Funding is available through SERC and HTNT. 

			    PhD in Cognitive Science

    For 1991 there are 3 SERC studentships available for PhD level
    research into a range of topics including:

    o computational modelling of emotion
    o computational modelling of cognition
    o interface design
    o computational and psychophysical approaches to vision 

			      Computing Facilities

    Students have access to ample computing facilities, including networks
    of Apollo, Sun and Sparc workstations in the Schools of Computer
    Science and Psychology. 
				   Contact

    For further details, contact: Dr. Mike Harris CSRC, School of
    Psychology, University of Birmingham, PO Box 363, Edgbaston,
    Birmingham B15 2TT, UK.
    Phone:  (021) 414 4913
    Email:  HARRIMWG@ibm3090.bham.ac.uk

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

Date: Fri, 15 Mar 91 06:06:39 CST
From: "David E. Goldberg" <GOLDBERG@vmd.cso.uiuc.edu>
Subject: Reports and Reprints Available

    A number of reports and journal article reprints are available:

    Goldberg, D. E., and Rudnick, M. (1991). Genetic algorithms and the
    variance of fitness (IlliGAL Report No. 91001), describes the calculation
    of fitness variance from Walsh transforms.  The focus on variance yields
    new population sizing formulae, and suggests one way of turning the schema
    theorem into a rigorous, probabilistic bound on schema convergence.

    Horner, A., & Goldberg, D. E. (1991). Genetic algorithms and
    computer-assisted music composition (IlliGAL Report No. 91002), describes
    computational results using GAs to generate music.  Operators, objective
    functions, and resulting compositions are described, as are contemplated
    extensions to existing work.

    Goldberg, D. E. (1990). A note on Boltzmann tournament selection for
    genetic algorithms and population-oriented simulated annealing.  Complex
    Systems, 4, 445-460, describes a method of achieving Boltzmann
    distributions across a population of structures using a form of tournament
    selection.  Although exponentially scaled objective functions offer a
    means of controlling selection pressure, the analogy to simulated
    annealing is not complete unless a distribution of structures representing
    different function values can be held stably at fixed intermediate
    temperature values.  This paper shows how to do this using a local
    tournament selection scheme.  A method for achieving similar results using
    sharing functions is also described.  Ties to extant niching methods are
    drawn.

    Goldberg, D. E., Deb, K., and Korb, B. (1990). Messy genetic algorithms:
    Studies in mixed size and scale. Complex Systems, 4, 415-444, describes
    theory and experiments with messy GAs.  Straightforward theory suggests
    that mGAs are probabilistically polynomial for problems of bounded
    deception.  Experiments with thresholding and effective-length
    tie-breaking suggest that mGAs can handle problems of arbitrary building
    block scale and length.  Messy classifiers, messy floating-point codes,
    and messy permutations are suggested.

    Papers may be requested from Kelsey Milman, milman@gal3.ge.uiuc.edu.
    Please include your postal address with your request.

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

Date: Mon, 25 Mar 91 15:30:02 WET
From: Nick Radcliffe <njr@castle.edinburgh.ac.uk>
Subject: Thesis available: Genetic Neural Networks on MIMD Computers

    Now available:

	    Genetic Neural Networks on MIMD Computers
	      Ph.D. Thesis (University of Edinburgh)
		       Nick Radcliffe

    My thesis, which was actually finished last year, has finally been
    reformatted in such a way that it is reasonable to issue it as a
    (long) technical report.  The parts of the thesis likely to be of
    special interest are the sections on Forma Analysis (an extension of
    schema analysis), including suggested design principles for genetic
    algorithms, and a critical (selective) literature survey of work using
    genetic algorithms to train neural networks.   The work in the thesis
    partly overlaps with my papers

	    Equivalence Class Analysis of Genetic Algorithms
	    Complex Systems, to appear
    and
	    Random Respectful Recombination,
	    submitted to ICGA91

    but there is currently no separate copy of the literature survey
    available.

    To receive a copy of the thesis, mail me as njr@castle.ed.ac.uk or write
    to me at the address below.


    Nick Radcliffe
    Edinburgh Parallel Computing Centre
    University of Edinburgh
    King's Buildings
    Edinburgh
    EH9 3JZ
    Scotland

    Phone: +44 31 650 5020

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