
Genetic Algorithms Digest   Friday, September 27 1991   Volume 5 : Issue 31

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

Today's Topics:
	- Coefficients for DeJongs function number 5
	- CFP: COGANN, Combinations of GAs and NNs @ IJCNN-92
	- Long Genomes; rationale revealed!

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

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

 First European Conference on Artificial Life (v5n10)         Dec 11-13, 1991
 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
 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: Smith <brosen@cis.udel.edu>
Date: Mon, 23 Sep 91 12:29:24 EDT
Subject: Coefficients for DeJongs function number 5

   I am trying to locate the a sub i coefficents used in DeJongs famous
   function number 5 of is five testbed functions. the Gensis 1.2
   simulator (ucsd) contains the following code:

   >/************************************************  file f5.c  ****/
   >
   >
   >static int a[2][25] ={
   >    {
   >            -32, -16, 0, 16, 32, -32, -16, 0, 16, 32, -32, -16, 0, 16, 32,
   >            -32, -16, 0, 16, 32, -32, -16, 0, 16, 32        },
   >    {
   >            -32, -32, -32, -32, -32, -16, -16, -16, -16, -16, 
   >            16, 16, 16, 16, 16, 32, 32, 32, 32, 32  }
   >};
   >
   >static int K = 500;
   >
   >double f5(x)
   >register double *x;
   >{
   >        register int i, j;
   >        register int n;
   >        double ans;
   >        double fj; 
   >        double prod;
   >        double diff;
   >
   >        ans = 1.0/K;
   >        for (j=0; j<25; j++)
   >        {
   >                fj = j+1;
   >                for (i=0; i<2; i++)
   >                {
   >                        diff = x[i] - a[i][j];
   >                        for ( prod=1, n=0; n<6; n++)
   >                                prod *= diff;
   >                        fj += prod;
   >                }
   >                ans += 1.0 / fj;
   >        }
   >        return (1.0 / ans);
   >}

   The array "a" is not fully filled in the declaration statment.
   Indicies a[2][20-24] are not inititalized.
   Does anyone have a correct implementation of the the 5 functions,
   or can tell me what the a coefficients are?
	   Thanks
	   Bruce

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

From: Russ Eberhart <RCE1%APLVM.BITNET@CUNYVM.CUNY.EDU>
Date: Thu, 26 Sep 91 09:35:52 EDT
Subject: CFP: COGANN, Combinations of GAs and NNs @ IJCNN-92

   CALL FOR PAPERS

   In conjunction with IJCNN-92, a one day workshop is scheduled
   on combinations of genetic algorithms (GAs) and neural
   networks (NNs).
   These paradigms are both inspired by information
   processing schemes used by Nature, but they typically
   have vastly different time constants.
   Recently, researchers have begun experimenting with
   combining them into adaptive/learning systems with
   new capabilities. The time is right for a workshop to
   bring together researchers to share concepts and experience.

   Three copies of original papers (10 pages, 12 point type) should be
   submitted to:

   Dr. Darrell Whitley
   Program Chair: COGANN
   Department of Computer Science
   Colorado State University
   Fort Collins, CO 80524 USA
   whitley@cs.colostate.edu

   by February 1, 1992.
   Papers will be rejected if they exceed the page limit
   or if they fail to describe work combining both these
   technologies.

   Relevant topics include, but are no limited to:
   Using GAs to train NNs
   Using GAs to design NN topologies and parameters
   Using GAs to analyze NN performance
   Artificial life applications and the evolution of learning
   Comparative studies
   Applications (especially encouraged)

   Workshop date: 6 June 1992
	   place: IJCNN-92, Baltimore, MD
	 sponsor: IEEE NN Council
     cooperating: International Society for Genetic Algorithms

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

From: rbelew@UCSD.EDU (Rik Belew)
Date: Fri, 24 May 91 17:48:06 PDT
Subject: Long Genomes; rationale revealed!

[many forwards removed]

(Ed's. Note: This was passed on to me a while ago by Rik, but
believing that GA researchers have no sense of humor, I had 
censored it.  Well, having been convinced that we do indeed 
have a sense of humor, and since submissions are slowing down
(the real reason), I have decided to print the following. -ACS ;-)


  For many years molecular biologists have been mystified by the fact that
  very little of an organism's DNA seems to serve any useful function.

  I have solved the mystery.

  The reason why only 30% of human DNA performs any useful function is that
  the rest of it is comments.

  Once we decode a typical human genome, we see that the contents begin as
  follows:

  ===
  /* HUMAN_DNA.H
   *
   * Human Genome
   * Version 2.1
   *
   * (C) God
   */

  /* Revision history:
   *
   * 0000-00-01 00:00  1.0  Adam.
   * 0000-00-02 10:00  1.1  Eve.
   * 0000-00-03 02:11  1.2  Added penis code to male version. A bit messy --
   *                        will require a rewrite later on to make it neater.
   * 0017-03-12 03:14  1.3  Added extra sex drive to male.h; took code from
   *                        elephant-dna.c
   * 0145-10-03 16:33  1.4  Removed tail.
   * 1115-00-31 17:20  1.5  Shortened forearms, expanded brain case.
   * 2091-08-20 13:56  1.6  Opposable thumbs added to hand() routine.
   * 2501-04-09 14:04  1.7  Minor cosmetic improvements -- skin colour made
   *                        darker to match my own image.
   * 2909-07-12 02:21  1.8  Dentition inadequate; added extra 'wisdom' teeth.
   *                        Must remember to make mouth bigger to compensate.
   * 4501-12-31 14:18  1.9  Increase average height.
   * 5533-02-12 17:09  2.0  Added gay option, triggered by high population
   *                        density, to try and slow the overpopulation problem
   * 6004-11-04 16:11  2.1  Made forefinger narrower to fit hole in centre of
   *                        CD.
   */

  /* Standard definitions
   */

  #define SEX male
  #define HEIGHT 1.84
  #define MASS 68

  /* Include inherited traits from parent DNA files.
   *
   * Files must be pre-processed with MENDEL program to provide proper
   * inheritance features.
   */

  #include "mother.h"
  #include "father.h"

  #infndef FATHER
  #warn("Father unknown -- guessing\n")
  #include "bastard.h"
  #endif

  /* Set up sex-specific functions and variables
   */
  #include <sex.h>

  /* Kludged code -- I'll re-design this lot and re-write it as a proper
   * library sometime soon.
   */
  struct genitals
     {
  #ifdef MALE
     Penis *jt;
  #endif
     /* G_spot *g;   Removed for debugging purposes */
  #ifdef FEMALE
     Vagina *p;
  #endif
     }

  /* Initialization bootstrap routine -- called before DNA duplication.
   * Allocates buffers and sets up protein file pointers
   */
  DNA *zygote_initialize(Sperm *, Ovum *);

  /* MAIN INITIALIZATION CODE
   *
   * Returns structures containing pre-processed phenotypes for the organism
   * to display at birth.
   *
   * Will be improved later to make output less ugly.
   */
  Characteristic *lookup_phenotype(Identifier *i);
  ===

  ...and so on.


  [ Note that God uses three-space tabs ]

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