
Genetic Algorithms Digest   Monday, March 30 1992   Volume 6 : Issue 12

 - Send submissions to GA-List@AIC.NRL.NAVY.MIL
 - Send administrative requests to GA-List-Request@AIC.NRL.NAVY.MIL
 - anonymous ftp archive: FTP.AIC.NRL.NAVY.MIL (see v6n5 for details)

Today's Topics:
	- What is TCGA?
	- TCGA description 
	- New TCGA Report
	- NN & ARTIFICIAL LIFE: Need Info
	- Numeric Representations and F1-F10
	- TR announcement: GAs: Deception, Convergence and Starting Conditions
	- Request for references to the counting 1's problem

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

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

 Canadian AI Conference, Vancouver,                           May 11-15, 1992
 COGANN, Combinations of GAs and NNs, @ IJCNN-92 (v5n31)      Jun 6,     1992
 ARTIFICIAL LIFE III, Santa Fe, NM                            Jun 15-19, 1992
 Evolution as a computational process, Monterey (v6n9)        Jun 22-24, 1992
 ML-92, Machine Learning Conference, Aberdeen (v6n8)          Jul  1-3,  1992
 10th National Conference on AI, San Jose,                    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
 SAB92, From Animals to Animats, Honolulu (v6n6)              Dec  7-11, 1992

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

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

From: Pierre Lecocq CRIF <M3730@eurokom.ie>
Date: Wed, 18 Mar 1992 17:45 CET
Subject: What is TCGA?

   Hello,

   Excuse me for my ignorance, but what is exactly the TCGA (The
   Clearinghouse for Genetic Algorithms)? I must obviously have missed its
   introduction (I'm on this list since last November or so).

   Thank you.
   Emanuel Falkenauer

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

From: rob@galab2.mh.ua.edu
Date: Mon, 30 Mar 92 11:00:54 CST
Subject: TCGA description

The Clearinghouse for Genetic Algorithms at The University of Alabama
(TCGA) is a distribution mechanism for GA related technical reports by
faculty and students at the UofA and associated agencies. TCGA was
established by Dave Goldberg in 1985, and I have been in charge of TCGA
since Fall 1991. TCGA currently distributes 32 reports. You can obtain a
catalog by contacting me at the addresses and numbers listed below. Most
reports are available free-of-charge. However, TCGA requests $9.00 ($12.00
overseas) for dissertations and theses, to offset the costs of copying,
binding and shipping these large documents. Most reports are available in
hardcopy form only.  TCGA distributes 3 programs (SGA-C, SGA-Cube, and
mGA) that are available via email (in UNIX shar files), via ftp at the
ga-list archive (ftp.aic.nrl.mil), and on various (requester-supplied)
media.

Robert Elliott Smith
    Department of Engineering of Mechanics
    Room 210 Hardaway Hall
    The University of Alabama
    Box 870278
    Tuscaloosa, Alabama 35487
<<email>> @ua1ix.ua.edu:rob@galab2.mh.ua.edu 
<<alternate email>>  resmith@ua1ix.ua.edu
<<phone>> (205) 348-1618
<<fax>> (205) 348-6419    

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

From: rob@galab2.mh.ua.edu
Date: Thu, 19 Mar 92 17:05:53 CST
Subject: New TCGA Report

   The Clearinghouse for Genetic Algorithms at The University of Alabama
   is pleased to announce the availability of the following technical report:



           Searching for Diverse, Cooperative Populations
                       with Genetic Algorithms

                                  by

                           Robert E. Smith
                      Dept. of Engin. Mechanics
                        University of Alabama

                          Stephanie Forrest
                      Dept. of Computer Science
                       University of New Mexico

                                 and

                          Alan  S. Perelson
                         Theoretical Division
                    Los Alamos National Laboratory


                               ABSTRACT 

   In typical applications, genetic algorithms (GAs) process populations of
   potential problem solutions to evolve a single population member that
   specifies an ``optimized'' solution. The majority of GA analysis has
   focused these optimization applications. In other applications (notably
   learning classifier systems and certain connectionist learning systems), a
   GA searches for a population of cooperative structures that jointly
   perform a computational task. This paper presents an analysis of this type
   of GA problem. The analysis considers a simplified genetics-based machine
   learning system: a model of an immune system. In this model, a GA must
   discover a set of pattern-matching antibodies that effectively match a set
   of antigen patterns. Analysis shows how a GA can automatically evolve and
   sustain a diverse, cooperative population. The cooperation emerges as a
   natural part of the antigen-antibody matching procedure. This emergent
   effect is shown to be similar to fitness sharing, an explicit technique
   for multi-modal GA optimization. Further analysis shows how the GA
   population can adapt to express various degrees of generalization.  The
   results show how GAs can automatically and simultaneously discover
   effective groups of cooperative computational structures.

   Copies of the report can be obtained by contacting:

   Robert Elliott Smith
       Department of Engineering of Mechanics
       Room 210 Hardaway Hall
       The University of Alabama
       Box 870278
       Tuscaloosa, Alabama 35487
   <<email>> @ua1ix.ua.edu:rob@galab2.mh.ua.edu 
   <<alternate email>>  resmith@ua1ix.ua.edu
   <<phone>> (205) 348-1618
   <<fax>> (205) 348-6419    

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

From: "ANDREA M. LONON" <AMLONO00@UKCC.uky.edu>
Date: Fri, 20 Mar 92 14:27:03 EST
Subject: NN & ARTIFICIAL LIFE: Need Info

   DOES ANYONE KNOW ABOUT ANYONE INTEGRATING ARTIFICIAL LIFE TECHNIQUES AND
   NEURAL NETWORK TECHNIQUES?  ANY WRITTEN INFORMATION WOULD BE PARTICULARLY
   HELPFUL, BUT ALL INFORMATION WILL BE APPRECIATED!!

					  ANDREA M. LONON

   AMLONO00@UKCC.UKY.EDU [INTERNET]
   AMLONO00@UKCC [BITNET]
		
------------------------------

From: oldingpa <oldingpa@cs.aston.ac.uk>
Date: Fri, 20 Mar 92 13:57:42 GMT
Subject: Numeric Representations and F1-F10

   Hello,

   Does anyone have any info/references etc pertaining to the use of numeric
   codes for parameter encoding i.e. binary, decimal, real. I'd be interested
   to hear about comparisons between them, both theoretically and
   experimentally.

   Also. I'd like an idiot proof introduction and guide to the test-bed
   functions F1-F10, in particular F5 which I can't really follow!

   Any help greatly appreciated

   Cheers
   Peat Olding

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

From: J.Kingdon@cs.ucl.ac.uk
Date: Wed, 25 Mar 92 19:40:25 +0000
Subject: TR announcement: GAs: Deception, Convergence and Starting Conditions

   The following technical report has been submitted to JETAI.

   Jason Kingdon
   Computer Science Department,
   University College London,
   Gower Street,
   London WC1E 6BT,
   England.


       Genetic Algorithms: Deception, Convergence and Starting Conditions.

			      Jason Kingdon

		     Department of Computer Science

		       University College London

		     Email: J.Kingdon@uk.ac.ucl.cs

				Abstract.

   Starting points, convergence and the class of problems genetic algorithms
   find hard to solve are investigated.  We redefine the notion of competing
   schemata to include schemata of arbitrary dimensions. The probability of
   convergence to such schemata is then given.  The probability of guaranteed
   non--convergence and an upper bound on the probability of finding point
   solutions is also derived.  It is shown that the probability of finding
   point solutions decreases to zero at a rate approximately equal to
   $\left({2\over 3}\right)^L$. Where $L$ is the string length, or chromosome
   dimension. In establishing such limits we prove that mutation is pivotal
   to an effective genetic search.

   Following the work of Liepins and Vose current definitions of deceptive
   problems are investigated.  This is done by way of examples. It is shown
   that the class of fully deceptive problems can be expanded to include
   arbitrary dimensions of deceptive problems within the traditionally
   defined fully deceptive problem. Examples of such problems are given and
   it is shown that the smallest class of encoding solutions to such hard
   problems is exponential. It is also shown that the current definition of
   orders of deception does not capture the true level of difficulty a
   genetic algorithm will encounter. The notion of genetic barriers is
   introduced as a tool for the analysis of such deceptiveness.

   Starting points for the algorithm are also discussed and it is shown that
   no optimal starting set exists for dimensions greater than 3 and sample
   size greater than 2.

   Lastly the the suitability of genetic algorithms for finding point optimum
   is discussed.  It is suggested that genetic algorithms real power lies in
   their ability to optimize over dynamic fitness conditions.

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

From: FISHER_JD@snd02.pcr.co.uk
Date: Wed, 25 MAR 92 09:54:31 GMT
Subject: Request for references to the counting 1s problem

   I am looking for references to any theoretical treatment of the "counting
   1s" problem.  What I mean by this is a fitness function that counts the
   number of 1s in an individual, and the fitness is how close that number is
   to a predetermined number.

   Thank you,

   Jeremy Fisher,
   Pfizer Central Research,
   Sandwich,
   Kent
   UK.

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




