Copy Link
Add to Bookmark
Report
Alife Digest Number 020
ALIFE LIST: Artificial Life Research List Number 20 Monday, May 14th 1990
ARTIFICIAL LIFE RESEARCH ELECTRONIC MAILING LIST
Maintained by the Indiana University Artificial Life Research Group
Contents:
ERL and Lamarck
ES Bibliography
request for alife software
Alife Proceedings....
----------------------------------------------------------------------
Date: Wed, 9 May 90 16:11:22 -0400
From: fritz_dg%pogo.dnet@gte.com
Subject: ERL and Lamarck
Quite obviously, driving AL with mechanisms elucidated in the Modern Synthesis
would require an awful lot of turns of the genetic wheel. Getting
somewhere on a human timescale would require lower cost genetic mechanisms
that are up to this task.
Ackley and Littmman's ERL (Digest #10)
seems to be a nice example of directly linking germ-plasm to soma-plasm--
the wiring diagram I achieve in my cortex today can be accurately passed to
my offspring tommorrow, down to the last synaptic weight. Assuming the
environment doesn't make any hyperspace leaps in character, it should
therefore be trackable by a quite small population with a relatively tight
but not rigid set of replicators. Yet in the long run, a darwinist
organism will always do better.
In ERL-SIM there was little room for sustained variability in the E
population, and the founder effect quickly took over. The ERL population, in
contrast, was able to maintain variability thru malleability. Another reason
that the E strategy didn't do so well is that the agents appeared to be haploid
organisms (?) This tends to be an _r_ (vs. _K_) characteristic, and in
little containers (like CRT screens), these guys tend to bloom and die off.
Continuity in such situations is achieved by many containers, always refreshed,
and a means of time and space travel (spores, raindrops).
Random change in inherited networks in E populations should not necessarily
be immediately expressed, just held (as it decays). In diploids, alleles that
don't match the current environment often hide behind those that do.
Gene duplication might also be considered as a holding mechanism, as
this also provides enhancement-producing raw material. Both are natural means
of making one's genetic assets richer.
There is a dilemma embedded in the genetic process (achievement vs. cost), and
so the calls for lamarckian evolution. Confirmed neodarwinists (I, at least)
pale (not blanche :-)) at the thought. But maybe it's necessary for AL.
With lower cost a couple of things come to mind. Are the benefits of
the "hopeful monster" lost, and how is increasing complexity achieved with
a shallower replicator pool?
I think Lamarck would have loved the term Evolutionary Reinforcement
Learning.
Dave Fritz fritz_dg%ncsd@gte.com
------------------------------
Date: Wed, 09 May 90 16:29:05 SET
From: UIN005%DDOHRZ11.BITNET@CUNYVM.CUNY.EDU
Date: 09 May 1990, 16:28:53 SET
From: UIN005 at DDOHRZ11
Subject: ES Bibliography
To: alife-mailing-list
Since I have been asked for giving hints to references concerning
Evolution Strategies (ESs) by a couple of interested people, here is
that kind of information:
Selected references concerning Evolution Strategies (ESs)::
Rechenberg, I. : Cybernetic Solution Path of an Experimantal Problem,
Royal Aircraft Establishment, Library Translation
No. 1122, August 1965
(this was the first report about the work of a small research group
Bienert / Rechenberg / Schwefel at the Technical University of Berlin)
Rechenberg, I. : Evolutionsstrategie - Optimierung technischer
Systeme nach Prinzipien der biologischen Evolution
Frommann-Holzboog, Stuttgart, 1973
(Rechenberg's dissertation from 1971)
Schwefel, H.-P.: Numerische Optimierung von Computer-Modellen
Birkhaeuser, Basel, 1977
(Schwefel's dissertation from 1974)
Schwefel, H.-P.: Numerical Optimization of Computer Models
Wiley, Chichester, 1981
(translation of the former work with an extra chapter
on correlated mutations, i.e. polygeny and pleiotropy)
Rechenberg, I. : Bionik, Evolution und Optimierung,
Naturwissenschaftliche Rundschau 26
Rechenberg, I. : Evolutionsstrategien, pp. 83-114 in:
Simulationsmethoden in der Medizin und Biologie
(B. Schneider und U. Ranft, Hrsg.), Springer, Berlin,
1978
Schwefel, H.-P.: Optimierung von Simulationsmodellen mit der Evolutions-
strategie, pp. 115-129 in:
Simulationsmethoden in der Medizin und Biologie
(B. Schneider and U. Ranft, eds.), Springer, Berlin,
1978
Schwefel, H.-P.: Direct Search for Optimal Parameters within
Simulation Models, pp. 91-102 in:
Proc. 12th Annual Simulation Symposium, Tampa/
Florida, 14-16 March 1979
Rechenberg, I. : Problemloesungen mit Evolutionsstrategien,
pp. 499-502 in:
Proceedings in Operations Research 9
(J. Schwarze et al., eds.), Physica, Wuerzburg, 1980
Schwefel, H.-P.: Ein Leistungsvergleich ableitungsfreier Methoden der
nichtlinearen Parameteroptimierung, pp. 503-504 in:
Proceedings in Operations Research 9
(J. Schwarze et al., eds.), Physica, Wuerzburg, 1980
Schwefel, H.-P., F. Drepper, R. Heckler : Combining Estimation,
Simulation and Optimization in Computer-Aided
Energy Planning, Angewandte Systemanalyse / Applied
Systems Analysis 2
Rechenberg, I. : The Evolution Strategy. A Mathematical Model of
Darwinian Evolution, pp. 122-132 in:
Synergetics - From Microscopic to Macroscopic Order
(E. Frehland, ed.), Springer, Berlin, 1984
Schwefel, H.-P.: Evolution Strategies: A Family of Non-Linear
Optimization Techniques Based on Imitating Some
Principles of Organic Evolution,
Annals of Operations Research 1
Schwefel, H.-P.: Collective Phenomena in Evolutionary Systems,
31st Annual Meeting of the International Society
for General Systems Research, Budapest, 1-5 June 1987
preprints vol. II, pp. 1025-1032
(also available as report #233 University of
Dortmund, Dept. of Computer Science)
Schwefel, H.-P.: Collective Intelligence in Evolving Systems,
pp.95-100 in:
Ecodynamics (W.F. Wolff et al., eds.), Springer,
Berlin, 1988
Schwefel, H.-P.: Optimum Seeking by Imitating Natural Intelligence,
vol. II, pp. 52-55 in:
Proc. 12th IMACS World Congress on Scientific
Computation, Paris, 18-22 July 1988
Schwefel, H.-P.: Evolutionary Learning Optimum-Seeking on Parallel
Computer Architectures, pp. 217-225 in:
Systems Analysis and Simulation 1988 (A. Sydow et al.,
eds.), Akademie-Verlag, Berlin (GDR), 1988
Rechenberg, I. : Artificial Evolution and Artificial Intelligence,
pp. 82-103 in:
Machine Learning: Principles and Techniques
(R. Forsyth, ed.), Chapman&Hall, London, 1989
Rechenberg, I. : Evolution Strategy: Nature's Way of Optimization,
pp. 106-126 in:
Optimization: Methods and Applications, Possibilities
and Limitations (H.W. Bergmann, ed.), Springer, Berlin,
1989 (Lecture Notes in Engineering 47)
Campos, I., H.-P. Schwefel : KBOPT: A Knowledge Based Optimization
System, pp. 211-221 in:
Computer Aided Optimum Design of Structures
(C.A. Brebbia et al., eds.), Springer, Berlin, 1989
Schwefel, H.-P.: Simulation evolutionaerer Lernprozesse, pp. 17-30 in:
Erwin-Riesch-Workshop on Systems Analysis of Bio-
medical Processes (D.P.F. Moeller, ed.), Vieweg,
Braunschweig,1989
Schwefel, H.-P.: Natuerliche Intelligenz in evolutionaeren Systemen,
pp. 151-164 in:
Evolution und Evolutionsstrategien in Biologie,
Technik und Gesellschaft (J. Albertz, ed.),
Schriftenreihe der Freien Akademie, vol. 9,
Wiesbaden, 1989
FORTRAN listings are available (Public Domain) from:
Schwefel, H.-P.: Subroutines EVOL, GRUP, KORR - Listings and User's
Guides, Internal Report KFA-STE-IB-2/80, Nuclear
Research Centre KFA, Programme Group of Systems
Research and Technological Development, Juelich,
April 1980
Schwefel, H.-P.: Optimum Seeking Methods - User's Guides,
Internal Report KFA-STE-IB-7/81, Nuclear
Research Centre KFA, Programme Group of Systems
Research and Technological Development, Juelich,
October 1981
N.B. There are a lot of spin-off articles describing the application
of ESs in a variety of fields. If you are interested, please ask for
that list: Prof. Dr.-Ing. H.-P. Schwefel, University of Dortmund,
Dept. of Computer Science, Chair of Systems Analysis (Informatik XI),
P.O. Box 50 05 00, D-4600 Dortmund 50, F.R. Germany,
e-mail: uin005@ddohrz11 (EARN/bitnet)
------------------------------
Date: Thu, 10 May 90 15:27:33 MDT
From: cgl%pullet@LANL.GOV (Chris Langton)
Subject: request for alife software
As an appendix to the proceedings of the 1990 Artificial Life
meeting, I would like to write up a short review of the different
software packages available for artificial life experiments.
If you have a software package which is relatively "user-friendly"
and which you think would be useful to other researchers (or which
will be so by the end of the year) for conducting alife
research, please send a brief description of it and the
method for obtaining it to me at the address below. If you
think figures (B&W, with color a remote possibility) would enhance
the description of your software, feel free to include a few.
Cheers!
Chris Langton
Complex Systems Group
MS B213, Theoretical Division Phone: 505-667-9471
Los Alamos National Laboratory Email: cgl@LANL.GOV
Los Alamos, New Mexico, USA
87545
------------------------------
Date: Thu, 10 May 90 15:27:46 MDT
From: cgl%pullet@LANL.GOV (Chris Langton)
Subject: Alife Proceedings....
Just a reminder, contributions to the proceedings of the second
Artificial Life Workshop were due April 30th. Unless arrangements
are made with one of the editors in advance, we will probably not
be able to accept papers after mid-May.
So, if you have been thinking about getting a paper in, please
get it to us as soon as possible.
Remember, we are also soliciting short video-sequences for an
accompanying video-tape. The deadline for submissions of
these sequences is the end of May. If you need more time,
please let me know. Also, if you need help producing
good quality sequences, get in touch with me, as I may
be able to help.
Cheers!
Chris Langton
Complex Systems Group
MS B213, Theoretical Division Phone: 505-667-9471
Los Alamos National Laboratory Email: cgl@LANL.GOV
Los Alamos, New Mexico, USA
87545
------------------------------
End of ALife Digest
********************************
