
Genetic Algorithms Digest   Tuesday, November 12 1991   Volume 5 : Issue 34

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

Today's Topics:
	- Re: dialogue on uniform crossover
	- revised technical report on DPE
	- GA software survey results

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

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
 FOGA-92, Foundations of Genetic Algorithms, Colorado (v5n32) Jul 26-29, 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: lje@philabs.Philips.Com  (Larry Eshelman)
Date: Thu, 31 Oct 91 09:54:09 EST
Subject: Re: dialogue on uniform crossover

   In the Christer Ericson/Ivan Ordonez-Reinoso dialogue on uniform crossover
   (UC) Ivan makes the tentative claim that "no experimental results exist for
   UC applied on harder functions."  This is not true.  In the Eshelman "CHC"
   paper in the Rawlins' FOGA book and in the Eshelman and Schaffer "Incest
   Prevention" paper in ICGA-91 results are published in which uniform crossover
   does better than two-point (2X) on hard functions.

   In the Schaffer and Eshelman "Spurious Correlation" paper in FOGA, we argue
   that the key to UX's success or failure is the selection mechanism used.
   UX is much more disruptive than 2X, but this is not a serious problem to
   the extent that the selection mechanism preserves the best individuals found:
   e.g., (1) using a biased replacement strategy (i.e., replace the worst
   member of the population) in a steady-state GA, or (2) using a population
   elitist selection strategy in a generational GA (i.e., merge the children
   and parents and pick the M best individuals where M is the population size).
   (Even using the individual elitist strategy with a generational GA (like
   GENESIS) helps.)

   We have found that UX almost always does as well as, and usually better
   than, 2X when combined with a population elitist selection strategy.  The
   main class of exceptions are deceptive problems where the deceptive bits
   are tightly grouped.

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

From: schraudo@cs.UCSD.EDU (Nici Schraudolph)
Date: Wed, 9 Oct 91 16:52:17 PDT
Subject: revised technical report on DPE

A revised version of the following technical report is now available.
Please do not reply to me to request copies but follow the instructions
given below.  Please do not forward this message to other lists.

- Nici Schraudolph.

        Dynamic Parameter Encoding for Genetic Algorithms
        =================================================

           Nicol N. Schraudolph       Richard K. Belew

The common use of static binary place-value codes for real-valued para-
meters of the phenotype in Holland's genetic algorithm (GA) forces either
the sacrifice of representational precision for efficiency of search or
vice versa.  Dynamic Parameter Encoding (DPE) is a mechanism that avoids
this dilemma by using convergence statistics derived from the GA popula-
tion to adaptively control the mapping from fixed-length binary genes to
real values.  DPE is shown to be empirically effective and amenable to
analysis; we explore the problem of premature convergence in GAs through
two convergence models.

                          =============

This report is available in compressed PostScript format via anonymous
ftp from cs.ucsd.edu (132.239.51.3), file pub/GAucsd/dpe-tr.ps.Z.  To
obtain a hardcopy, request technical report "LAUR 90-2795 (revised)"
via e-mail from office%bromine@LANL.GOV, or via plain mail from

	Technical Report Requests
	CNLS, MS-B258
	Los Alamos National Laboratory
	Los Alamos, NM 87545
	USA

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

From: schraudo@cs.UCSD.EDU (Nici Schraudolph)
Date: Wed, 16 Oct 91 21:20:57 PDT
Subject: GA software survey results

  Genetic Algorithm Software Survey - Overview
  ============================================

  As part of the workshop on Software Support and Test Functions at the 4th
  International Conference on Genetic Algorithms (ICGA) I began to solicit
  and collect information about available GA software packages.  Having re-
  ceived a surprising number of responses, I feel that the following is a
  fairly comprehensive picture of what's out there at this time.  Additions
  and corrections are of course welcome and may make it into future updates
  of this file.

  Special thanks to the many informants, including - but not limited to - the
  authors/contacts listed below.
						 San Diego, 10/16/91
  - Nici Schraudolph (nici@cs.ucsd.edu).

  *** PLEASE NOTE: ***

  For many of these software packages, specific ordering instructions are given
  in the descriptions below.  Please read and follow them before unnecessarily
  bothering the listed author or contact.  Also note that I haven't tested any
  of these programs (with the exception of the one I administer), so I can't
  make any recommendations or comments regarding their quality.

  ********************
   Name     Type  OS Lang Price  Author or Contact
   ====     ====  == ==== =====  =================

  Evolution  GE, DOS      free  voigt@iir-berlin.adw.dbp.de (Hans-Michael Voigt)
  Machine    ES

  GA         GE  DOS (C++) free  mrh@camcon.co.uk (Mark Hughes)
  Workbench

  OOGA       OO      Lisp  $60/  70461.1552@compuserve.com (Lawrence Davis)
  GENESIS    GE  DOS   C   both  gref@aic.nrl.navy.mil (John Grefenstette)

  GA         OO  Mac  C++   (1)  America Online: Wilson SD (Steve Wilson)
  Framework*

  XYpe*      SS  Mac  (C)  $725  (Ed Swartz)

  Splicer    GE  Mac,
		  X11  C    (2)  bayer@galileo.jsc.nasa.gov (Steve Bayer)

  ESCaPaDE   ES  Unix  C   free  iwan@gorbi.informatik.uni-dortmund.de
				 (Frank Hoffmeister)

  GENEsYs*   GE  Unix  C   free  baeck@gorbi.informatik.uni-dortmund.de
				 (Thomas Baeck)

  GAucsd     GE  Unix  C   free  nici@cs.ucsd.edu (Nici Schraudolph)

  Genitor    SS  Unix  C   free  whitley@cs.colostate.edu (Darrell Whitley)

  GAGA       GE  Unix  C   free  jon@cs.ucl.ac.uk (Jon Crowcroft)

  GAME*      OO  X11  C++       J.RibeiroFilho@cs.ucl.ac.uk (Jose Ribeiro Filho)

  Genetic-2L GE  Unix (C)  free  zbyszek@unccvax.uncc.edu (Zbigniew Michalewicz)
  Genetic-2N  "   "    "    "             "                      "
  Genocop     "   "    "    "             "                      "
  Gafoc       "   "    "    "             "                      "

  Legend:
  ======

  Name - asterisk: software not yet complete (see description for details)

  Type - GE: generational GA, SS: steady-state GA
	 ES: evol. strategy,  OO: object-oriented

  OS - Operating System; X11 implies Unix

  Lang - Programming Language; in parentheses: source code not included

  Price - (1): available on license basis
	  (2): free to government contractors, others pay nominal fee

  Author or Contact - given as Internet e-mail address if possible

  Genetic Algorithm Software Survey - Descriptions
  ================================================


  Evolution Machine:
  ==================

  The "Evolution Machine" presents a collection of evolutionary algorithms
  (Genetic Algorithms and Evolution Strategies) in a common framework.  It
  runs on PCs with MS-DOS and includes extensive menu techniques.  A more
  detailed description of the "Evolution Machine" is given by the manual,
  available in PostScript form via anonymous ftp from ftp.wtza-berlin.de 
  (141.16.244.4), file em-man.ps.Z.

  In this manual an introduction is given, the handling is fully described
  and the included algorithms are compared with regard to their performance.
  Interested parties can order the code of the "Evolution Machine"  free of
  charge by request from one of the authors:

	     Hans-Michael Voigt             Joachim Born
  Internet:  voigt@iir-berlin.adw.dbp.de    born@iir-berlin.adw.dbp.de
  Phone:     +372-674-5958                  +372-674-2484
  Address:
	     Institute for Informatics and Computing Techniques
	     Rudower Chaussee 5
	     1199 Berlin
	     Germany

  GA Workbench:
  ============

  A mouse-driven interactive GA demonstration program aimed at people wishing
  to show GAs in action on simple function optimizations and to help newcomers
  understand how GAs operate.  Features: problem functions drawn on screen
  using mouse, run-time plots of GA population distribution, peak and average
  fitness.  Useful population statistics displayed numerically, GA configura-
  tion (population size, generation gap etc.) performed interactively with
  mouse.  Requirements: MS-DOS PC, mouse, EGA/VGA display.

  Available on 5.25'' disk by request from:

	  Mark Hughes
	  Cambridge Consultants Ltd.
	  The Science Park
	  Milton Road
	  Cambridge  CB4 4DW
	  United Kingdom

  OOGA, GENESIS:
  ==============

  OOGA (Object-Oriented GA) is a genetic algorithm designed for industrial use.
  It includes examples accompanying the tutorial in the companion "Handbook
  of Genetic Algorithms".  OOGA is designed such that each of the techniques
  employed by a GA is an object that may be modified, displayed or replaced in
  object-oriented fashion.  OOGA is especially well-suited for individuals
  wishing to modify the basic GA techniques or tailor them to new domains.

  The buyer of OOGA also receives GENESIS, a generational GA system written
  by John Grefenstette.  As the first widely available GA program GENESIS has
  been very influential in stimulating the use of GAs, and several other GA
  packages are based on it.  This release sports an improved user interface.

  OOGA and GENESIS are available together on 3.5'' or 5.25'' disk for $60
  ($52.50 inside North America) by order from:

	  T.S.P.
	  P.O. Box 991
	  Melrose, MA 02176
	  U.S.A.

  GA Framework:
  =============

  This object-oriented framework for doing GAs includes classes of objects
  that are easily subclassable to add any special features.  There is also
  a library of objects designed to integrate neural networks with GAs.
  Originally written in Object Pascal, GA Framework is currently being
  rewritten in C++; Steve Wilson is also looking for collaborators for
  later improvements.

  GA Framework is not commercially available yet, but interested users
  should contact:

	  Steve Wilson
	  Emergent Behavior
	  635 Wellsbury Way
	  Palo Alto, CA 94306
	  U.S.A.

  XYpe:
  ====

  XYpe (The GA Engine) is a commercial GA application and development package
  for the Apple Macintosh.  Its standard user interface allows you to design
  chromosomes, set attributes of the genetic engine and graphically display
  its progress.  The development package provides a set of Think C libraries
  and include files for the design of new GA applications.  XYpe supports
  adaptive operator weights and mixtures of alpha, binary, gray, ordering
  and real number codings.

  The price of $725 (in Massachusetts add 5% sales tax) plus $15 shipping
  and handling includes technical support and three documentation manuals.
  XYpe requires a Macintosh SE or newer with 2MB RAM running OS V6.0.4 or
  greater, and Think C if using the development package.

  Currently the GA engine is working; the user interface will be completed
  on demand.  Interested parties should contact:

	  Ed Swartz
	  Virtual Image, Inc.
	  75 Sandy Pond Road #11
	  Ayer, MA 01432
	  U.S.A.

	  (508) 772-4225


  Splicer:
  =======

  Splicer is a genetic algorithm tool that can be used to solve search and
  optimization problems, created by the Software Technology Branch (STB) of
  the Information Systems Directorate at NASA/Johnson Space Center with
  support from the MITRE Corporation.  Splicer was written in C on an Apple
  Macintosh, then ported to Unix workstations running X11; it has a modular
  architecture with well-defined interfaces between a GA kernel, represen-
  tation libraries, fitness modules, and user interface libraries.

  The representation libraries contain functions for defining, creating,
  and decoding genetic strings, as well as multiple crossover and mutation
  operators.  Libraries supporting binary strings and permutations are
  provided, others can be created by the user.

  Fitness modules are typically written by the user, although some sample
  applications are provided.  The modules may contain a fitness function,
  initial values for various control parameters, and a function which
  graphically displays the best solutions.

  Splicer provides event-driven graphic user interface libraries for the
  Macintosh and the X11 window system (using the HP widget set); a menu-
  driven ASCII interface is also available though not fully supported.
  The extensive documentation includes a reference manual and a user's
  manual; an architecture manual and the advanced programmer's manual
  are currently being written.

  An electronic bulletin board (300/1200/2400 baud, 8N1) with information
  regarding Splicer can be reached at (713) 280-3896 or (713) 280-3892.
  Splicer is available free to NASA and its contractors for use on government
  projects by calling the STB Help Desk weekdays 9am-4pm CST at (713) 280-2233.
  Government contractors should have their contract monitor call the STB Help
  Desk; others may purchase Splicer at a nominal fee from:

	  COSMIC
	  382 E. Broad St.
	  Athens, GA 30602
	  U.S.A.

	  (404) 542-3265

  ESCaPaDE:
  ========

  ESCaPaDE is a sophisticated software environment to run experiments
  with Evolutionary Algorithms, such as e.g. an Evolution Strategy.
  Future versions of the software will provide a well-defined interface
  to any kind of Evolutionary Algorithm, for instance Genetic Algorithms.
  The main support for experimental work is provided by two internal
  tables:
	  (1) a table of objective functions and
	  (2) a table of so-called data monitors,

  which allow easy implementation of functions for monitoring all types
  of information inside the Evolutionary Algorithm under experiment.

  ESCaPaDE 1.2 comes with the KORR implementation of the Evolution
  Strategy by H.-P. Schwefel which offers simple and correlated mutations.
  KORR is provided as a FORTRAN 77 subroutine, and its cross-compiled
  C version is used internally by ESCaPaDE.

  ESCaPaDE 1.2 will be available by e-mail request in order to track the
  spread of the software as this is its first public release.  An extended
  version of the package was used for several investigations so far and
  has proven to be very reliable.  The software and its documentation is
  fully copyrighted although it may be freely used for scientific work;
  it requires 5-6 MB of disk space.

  To obtain ESCaPaDE via mail server, send a message consisting of the
  line "get ESCaPaDE" to iwan@lumpi.informatik.uni-dortmund.de; at this
  address you may also request ESCaPaDE on quarter-inch tape.

  GENEsYs:
  =======

  GENEsYs is a GENESIS-based GA implementation which includes extensions
  and new features for experimental purposes, such as selection schemes
  like linear ranking, Boltzmann, (mu, lambda)-selection, and general
  extinctive selection variants, crossover operators like n-point and
  uniform crossover as well as discrete and intermediate recombination.
  Self-adaptation of mutation rates is also possible.

  A set of objective functions is provided, including De Jong's functions,
  complicated continuous functions, a TSP-problem, binary functions, and a
  fractal function.  There are also additional data-monitoring facilities
  such as recording average, variance and skew of object variables and mu-
  tation rates, or creating bitmap-dumps of the population.

  GENEsYs is expected to become available in June 1992.

  GAucsd:
  ======

  GAucsd is a GENESIS-based GA package incorporating numerous bug fixes
  and user interface improvements.  Major additions include a wrapper
  that simplifies the writing of evaluation functions, a facility to
  distribute experiments over networks of machines, and Dynamic Parameter
  Encoding, a technique that improves GA performance in continuous search
  spaces by adaptively refining the genomic representation of real-valued
  parameters.

  GAucsd was written in C for Unix systems, but the central GA engine is
  easily ported to other platforms.  The entire package can be ported to
  systems where implementations of the Unix utilities "make", "awk" and
  "sh" are available.

  GAucsd can be obtained via anonymous ftp from cs.ucsd.edu (132.239.51.3),
  file pub/GAucsd/GAucsd12.sh.Z, or via mail server - send an empty message
  with the subject line containing "send GAucsd source" to nici@cs.ucsd.edu.
  Requests to be added to a mailing list for dissemination of GAucsd bug
  reports, patches and updates should be directed to the same address.

  Genitor:
  =======

  Genitor is a modular GA package containing examples for floating-point,
  integer, and binary representations.  Its features include many sequencing
  operators as well as subpopulation modelling.

  GAGA:
  ====

  GAGA (GA for General Application) is a self-contained, re-entrant procedure
  which is suitable for the minimisation of many "difficult" cost functions.
  Originally written in Pascal by Ian Poole, it was rewritten in C by Jon
  Crowcroft.  GAGA can be obtained by request from the author; given suffi-
  cient interest it will be made available via anonymous ftp.

  GAME:
  ====

  GAME (GA Manipulation Environment) aims to demonstrate GA applications
  and build a suitable programming environment.  Currently in the early
  development stage, the programming environment will comprise a graphic
  interface (using X-Windows), a library of parameterized algorithms and
  applications, a specialized high level language based on C++, and com-
  pilers to various workstations and parallel machines.

  Genetic, Genocop, Gafoc:
  =======================

  These are four bare-bones GA programs (no documentation, no comments, no
  interfaces) written by Zbigniew Michalewicz for his own research.  Genetic
  optimizes a transportation problem, either linear (Genetic-2L) or nonlinear
  (Genetic-2N).  Genocop is a system to optimize any function with any set of
  linear constraints, while Gafoc solves discrete optimal control problems.

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