
Genetic Algorithms Digest    Friday, 23 February 1990    Volume 4 : Issue 3

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

Today's Topics:
	- GA Demonstration for PCs
	- Request for references on games
	- I need a C coded GA
	- A GA skeleton
	- TSP and other sequencing problems
	- Question about F5
	- Problem with fitness values of zero
	- Developmental biology

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

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

Double Auction Tournament - Sante Fe Institute  (v3n12)       Mar 1990
Workshop on GAs, Sim. Anneal., Neural Nets - Glasgow (v3n15)  May 9, 1990
7th Intl. Conference on Machine Learning (submissions 2/1/90) Jun 21-23, 1990
Workshop Foundations of GAs (v3n19)                           Jul 15-18, 1990
Conference on Simulation of Adaptive Behavior, Paris (v3n21)  Sep 24-28, 1990

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

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

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

From: Mark Hughes <mrh@camcon.co.uk>
Date: Wed, 17 Jan 90 13:40:32 GMT
Subject: GA Demonstration for PCs

I have produced a genetic algorithm demonstration program for PCs, and
offer it to anyone who thinks it may come in handy. It requires
an IBM compatible PC with EGA display and a mouse.

Below is an extract from the README file of the program. If you
would like a copy please mail me with your details stating whether
you want it emailed or sent on disk. There is no charge for either.

If sent a disk, you will receive a self extracting archive,
if emailed, you will receive several uuencoded sections that
can be concatenated and decoded to give the self extracting
archive.

If you want the software emailed, please also give your postal
address in case I can't reach you by email.

Mark Hughes
21 Arbury Road, Cambridge, CB4 2JB. UK.
mrh@camcon.co.uk

README EXTRACT STARTS----------------------------------

WHAT IS THIS?
-------------

This file is distributed with a self extracting archive containing a
program call the Genetic Algorithm Workbench and its documentation.

The program is an interactive tool/demonstration for experimentation with
genetic algorithms (see next).  It enables you to set problems (i.e. draw
a function f(x) using a mouse) and to try out different flavours of genetic
algorithm to see how well they can find the global maximum of the function.

HOW DO I USE THIS SOFTWARE?
---------------------------

You must have access to an IBM compatible personal computer running MS-DOS or
PC-DOS, a mouse and an EGA compatible display.  It will work in VGA mode, but
unless the VGA display is put into EGA emulation will give a squashed display.

If you have access to such a machine, you will need to extract the files from
the archive which is done by changing to a (preferably empty) directory on
a disk with about 400K of free space, inserting the distribution disk in
drive a: and typing

	a:extract

This should result in the following files being extracted into the current
directory:

	gaw.exe	- the executable program

	gaw.txt	- gaw manual (Unformatted ASCII text)

	gaw.ps	- gaw manual (Postscript format)
	fig1.ps	- figure 1 for insertion in the manual (Postscript format)

	gaw.fx	- gaw manual (Epson FX80 format)
	fig1.fx	- figure 1 (Epson FX80 format)

README EXTRACT ENDS------------------------------------

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

Date: Thu, 25 Jan 90 13:49:14 CST
From: "John M. Wildenthal" <N123DO%TAMVM1.BITNET@ricevm1.rice.edu>
Subject: Request for references on games

     I was wondering if anyone has seen or done anything on GA-classifiers
playing games other than Prisoner's Dilemma Games.  I was wondering if Tic-Tac-
Toe in particular had been attempted, but would be interested in any instance
of GA-classifiers being used in non-cooperative strategic games.

                                                  Sincerely Yours,
                                                  John M. Wildenthal

                                                  N123DO @ TAMVM1 (bitnet)
                                                  N123DO @ TAMVM1.EDU (arpanet)

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

Date: Tue, 30 Jan 1990 14:33-EST 
From: Jin-Oh.Kim@FAS.RI.CMU.EDU
Subject: I need a C coded GA

Hello.
I'm a Ph.D. sutdent in Carnegie-Mellon University.
As a subtopic of thesis, I need to apply GA to my complex nonlinear
optimization problem. I have a GENESIS, but it looks a little
complicated. If you have  GA written in C, which can be understood
easily, Could you send me  a copy of it?

Thanks.
Jin-Oh....

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

Date: Wed, 31 Jan 90 10:20:48 GMT
From: pat%computer-science.strathclyde.ac.uk@NSFnet-Relay.AC.UK
Subject: A GA skeleton

Below is a skeleton GA proposed by Dekang Lin. I would be
interested in any comments.

  (0) Create population of size N
  (1) Sort population in ascending order of fitness
  (2) generate random x, where  1 <= x <= N-1
  (3) generate random y, where  x < y <= N
  (4) C1 := xth member of population
  (5) C2 := yth member of population
  (6) C3 := first or second crossover(C1 C2 random(L))
  (7) Get fitness of C3
  (8) merge C3 into population
  (9) Remove first (weakest) member of population
  (10) apply secondary genetic operators probabilistically
  (11) Terminate or go to (2)

Patrick Prosser

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

Date: Wed, 31 Jan 90 15:28:30 GMT
From: pat%computer-science.strathclyde.ac.uk@NSFnet-Relay.AC.UK
Subject: TSP and other sequencing problems

Is the travelling salesman problem (CSP) significantly different from
other sequencing problems, such as single resource scheduling or bin
packing?

Why am I interested in this? Well early work done by Derek Smith gave
us his crossover operator for sequences used in bin packing (ICGA
1985).  About the same time we get Golberg and Lingle and
Grefenstette (ICGA 1985) working on operators for the TSP (such as PMX)

Is one of these operators better than the other, and can either be
applied to both domains (TSP, sequencing)?  Has anyone made a
comparison?

Patrick Prosser

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

Date: Tue, 06 Feb 90 12:24:21 EST
From: Ljubomir Perkovic <PERHC@CUNYVM.CUNY.EDU>
Subject: Question about F5

I need to know what the matrix A(i,j) appearing in De Jong's function F5
exactly is. Whoever knows it, please send it to: PERHC@CUNYVM.CUNY.EDU
Thank you.

Sincerely,
Ljubomir Perkovic.

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

From: John Douglas Turner <douglas@ms.uky.edu>
Date: Fri, 23 Feb 90 10:39:31 EST
Subject: Problem with fitness values of zero

I am working on a GA problem where we are searching for the MIN of a
function and some times we get zero as the fitness value (f(x)) but
using this as the roulette wheel fraction yelds zero % and no chance
of survival.

We have thought about subtracting the fitness value from the max possible
fitness (ie the worst case) and use this for the roulette wheel.

Is this a valid method, and what are some other solutions?

john.....

-- 
John Douglas Turner		douglas@ms.uky.edu   or  douglas@UKMA.BITNET 
University of Kentucky			{rutgers,uunet}!ukma!douglas
902 Patterson Office Tower  			(606)-257-6824 
Lexington, Ky.  40502		

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

From: rudnick@cse.ogi.edu (Mike Rudnick)
Date: Tue, 20 Feb 90 21:05:20 PST
Subject: Developmental biology

Is anyone doing work on the developmental (as in developmental
biology) aspects of NNs?  I'm (at least somewhat) aware of Edelman &
Reeke's work on neuronal group selection, and Wilson's papers on
L-systems, but no other work.

My particular interest is in using genetic search techniques for the
design of artificial neural nets.  I want to find a useful analog of
developmental biology to aid in both the design of compact genetic
codes and in converting those genetic codes to nets (more or less)
ready to be trained.

I'm eager to contact new people working in these areas, and would
appreciate any pointers or references that may be appropriate.

Thanks,
Mike Rudnick				Computer Science & Eng. Dept.
Domain:	rudnick@cse.ogi.edu		Oregon Graduate Institute (was OGC)
UUCP: {tektronix,verdix}!ogicse!rudnick	19600 N.W. von Neumann Dr.
(503) 690-1121 X7390 (or X7309)		Beaverton, OR. 97006-1999

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

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

