Approximation and Learning in Evolutionary Computation

Dr. Yaochu Jin, Prof. Sushil J. Louis and Prof. Khaled M. Rasheed

Duration: Half Day

In real-world applications, it is often necessary to build approximate
models for fitness evaluation. One essential difficulty in applying
evolutionary algorithms to the optimization of complex systems is the
high time complexity of each fitness evaluation. In some applications,
no explicit mathematical functions are available for fitness evaluation.
Approximate fitness models have proved useful in dealing with noisy and
multi-modal fitness functions. This workshop aims to get together
researchers coming from different research areas to identify the
state-of-the-art in approximate fitness models, discuss the main
problems and identify future work in this area. Issues include but
are not limited to:

* Off-line and on-line learning for approximate model construction
* Off-line and on-line learning for performance improvement
* Evolution control and model management in evolutionary computation
* Multi-level evolutionary optimization
* Learning in multi-objective evolutionary optimization
* Fitness estimation in noisy environment
* Comparison of different modeling methods, such as neural networks,
response surface and least squares methods, and probabilistic models
for evolutionary computation
* Comparison of sampling techniques for on-line and off-line learning

Dr. Yaochu Jin
Future Technology Research
Honda R&D Europe
Carl-Legien-Str. 30
63073 Offenbach/Main
Germany
Email: yaochu_jin@de.hrdeu.com
Tel: +49-69-89011735
Fax: +49-69-89011749

Prof. Sushil J. Louis
University of Nevada, Reno
Reno, NV 89557
U.S.A.
Email: sushil@cs.unr.edu
Tel: (775)784-4315
Fax: (775)784-1877

Prof. Khaled M. Rasheed
Computer Science Department
The University of Georgia
Athens, GA 30602
U.S.A.
Email: khaled@cs.uga.edu
Tel: (706)542-3444
Fax: (706)542-2966

Further details

Biological Applications of Evolutionary Computation

Wolfgang Banzhaf and James A. Foster

Duration: Half Day

The field of Genetic and Evolutionary Computation has greatly benefited by borrowing ideas from Biology. In recent years, the ability of GEC to solve biological problems, and thereby to "repay the debt", has become apparent. It is also becoming apparent that the computer itself can be used as a model organism with which to study evolutionary processes in nature. The workshop on Biological Applications of Genetic and Evolutionary Computation is intended to serve as a catalyst for using GEC to answer biological questions.

Wolfgang Banzhaf
Department of Computer Science
Informatik 11
University of Dortmund
Joseph-von-Fraunhofer-Str. 20
44227 Dortmund,
GERMANY
email: banzhaf@cs.uni-dortmund.de
tel: +49-(0)231-9700-953
fax: +49-(0)231-9700-959
www: http://ls11-www.cs.uni-dortmund.de/people/banzhaf

James A. Foster
Department of Computer Science
University of Idaho
Moscow, ID 83844-1010
USA
email: foster@cs.uidaho.edu
tel: 208.885.7062
fax: 208 885-9052 (fax)
www: http://www.cs.uidaho.edu/~foster

Further details

Evolutionary COmputation and Multi-Agent Systems (ECOMAS 2002)

Robert E. Smith, Claudio Bonacina, Cefn Hoile and Paul Marrow

Duration: Full Day

Multi-agent systems (MAS) are collections of interacting autonomous entities. The behaviour of the MAS is a result of the repeated asynchronous action and interaction of the agents. Desirable self-organisation is observed in many biological, social and physical systems. However, fostering these conditions in artificial systems proves to be difficult and offers the potential for undesirable behaviours to emerge. Thus, it is vital to be able to understand and shape emergent behaviours in agent based systems. EC provides a paradigm for addressing this need. Moreover, EC techniques are inherently based on a distributed paradigm (natural evolution), making them particularly well suited for adaptation in agents. The goal of this workshop is to maintain a dialog among researchers and practitioners who are examining EC in MAS. It represents an important opportunity for those active or interested in this area to hear about current work, and discuss future directions and priorities.

Robert E. Smith
Director, Intelligent Computing Systems Centre
The University of The West of England
Coldharbour Lane
Frenchay
Bristol BS16 1QY
UK
Email: robert.smith@uwe.ac.uk
Tel: +44 117 942 1495
Fax: +44 870 1617266

Claudio Bonacina
Intelligent Computing System Centre
The University of The West of England
Coldharbour Lane
Frenchay
Bristol BS16 1QY
UK
Email: c2-bonacina@uwe.ac.uk
Phone: +44 117 344 3178
Fax: +44 117 344 3182

Cefn Hoile
Intelligent Systems Laboratory,
BTexact Technologies
Antares 2 PP 5,
Adastral Park,
Ipswich IP5 3RE,
UK
Email: cefn.hoile@bt.com
Phone: +44-1473-642405
Fax: +44-1473-647410

Paul Marrow
Intelligent Systems Laboratory,
BTexact Technologies
Antares 2 PP 5,
Adastral Park,
Ipswich IP5 3RE,
UK
Email: paul.marrow@bt.com
Phone: +44-1473-645166
Fax: +44-1473-647410

Further details Discussion Forum homepage

Grammatical Evolution Workshop (GEWS 2002)

Dr. Michael O'Neill and Dr. Conor Ryan

Duration: Half Day

Grammatical Evolution (GE) is an automatic programming system that can evolve programs in an arbitrary language from a binary string. GE adopts a genotype-phenotype mapping process taking as input a grammar that describes the syntax of the evolved program. In addition to the grammar, the search algorithm (the standard has been a variable-length genetic algorithm) is also a 'plug-in' component of the system. The workshop will address all aspects of GE including foundations, extensions, analysis and applications.

Dr. Michael O'Neill
Dept. of Computer Science & Information Systems
University of Limerick
Ireland
Email: michael.oneill@ul.ie
Tel: +353-61-202730
Fax: +353-61-202734

Dr. Conor Ryan
Dept. of Computer Science & Information Systems
University of Limerick
Ireland
Email: conor.ryan@ul.ie
Tel: +353-61-202730
Fax: +353-61-202734

Further details

Intelligent Interface and Interactive Agents Through Evolutionary Computation (IIIATEC)

Oswaldo Velez-Langs, Angelica de Antonio and Ricardo Imbert Paredes

Duration: Half Day

Intelligent Interface Agents (IIA) are those programs that present the property of interaction with a user and/or an environment, using adequate technology, and showing intelligent behavior in the interaction. It is important to understand phenomena such as self-emergence of behavior and adaptation to the user and/or the environment in the analysis and design of applications of IIAs. Similarly it is important to understand that IIAs can be applied to study these phenomena in real systems. In Evolutionary Computation (EC) we find a paradigm that shares with Agents the characteristics: distribution, autonomy, emergence and orientation to goals. The hybridization of these two computational paradigms of natural inspiration can help to extend our knowledge of both IIAs and of EC and establishes a foundation for evolution in new directions.

Oswaldo Velez-Langs
Laboratorio Decoroso crespo
Facultad de Informatica (UPM)
Campus de Montegancedo
28660 - Boadilla del Monte (Madrid)
España
Email: ovelez@zipi.fi.upm.es
Tel: (+34) 913366941
Fax: (+34) 913366917
Facultad de Ingenierias
Corporacion Universitaria del Sinu
Carrera 1w Calle 38
Monteria
Colombia
Tel: (+57) 4 7840340
Fax (+57) 4 7840677

Angelica de Antonio
Facultad de Informatica (UPM)
Campus de Montegancedo
28660 - Boadilla del Monte (Madrid)
España
Email: angelica@fi.upm.es
Tel: (+34) 913366925
Fax: (+34) 913366917

Ricardo Imbert Paredes
Facultad de Informatica (UPM)
Campus de Montegancedo
28660 - Boadilla del Monte (Madrid)
España
Email: rimbert@fi.upm.es
Tel: (+34) 913366923
Fax: (+34) 913366917

[Further details]

ISGEC Workshop on Standards

Dr Peter J Bentley

Duration: Half Day

Computer science is dominated by the need to publish, publish, publish, but sometimes this can happen at the expense of research. All too often poor papers, clumsy presentations, bad reviews or even bad science can clutter a conference, causing distractions from the more carefully prepared work. All good scientists and students studying to become scientists should be concerned with matters of quality. The only problem is defining what is good and what is bad! This workshop will focus on promoting high quality standards of research in our field. It will tackle issues such as "how to review papers", "how to present your work", "how to write clearly" and "good scientific practice".

Dr Peter J Bentley
Department of Computer Science
University College London
Gower Street
London WC1E 6BT
UK
Email: P.Bentley@cs.ucl.ac.uk
Tel: +44 020 7679 1329
Fax: +44 020 7387 1397

Further details

Learning and Adaptation in Evolutionary Computation

Sibylle Mueller, Nicol Schraudolph and Petros Koumoutsakos

Duration: Half Day

Applying adaptation schemes and/or machine learning techniques such as self-organizing nets or reinforcement learning to evolutionary computation methods is a challenging task which may help to:

* reduce the heuristics usually implied in evolutionary algorithms, hence acting as a meta-optimization strategy,
* offer new perspectives for optimization problems with expensive objective functions,
*accelerate evolutionary algorithms, e.g. by induction-based control
*reducing the time complexity of evolution strategies.

This workshop will bring together researchers from the fields of machine learning and evolutionary optimization to discuss how learning and adaptation, at either individual or population level, can improve the efficiency of evolutionary algorithms.

Sibylle Mueller
Nicol Schraudolph
Petros Koumoutsakos
Institute of Computational Sciences,
Swiss Federal Institute of Technology Zuerich,
ETH Zentrum, WET D3,
CH-8092 Zuerich,
Switzerland
Email: {muellers,petros}@inf.ethz.ch
Tel: +41 1 632 6827
Fax: +41 1 632 1703

Further details

Representations for Genetic and Evolutionary Algorithms

Franz Rothlauf and Illinois Genetic Laboratory

Duration: Half Day

For successful and efficient use of genetic and evolutionary algorithms (GEAs), it is not enough to simply use efficient genetic operators, but one must also find a proper representation for the problem. The representation must at least be able to encode all possible solutions of an optimization problem, and genetic operators such as crossover and mutation should be applicable to it. However, even though it is well known that the choice of proper representations is crucial for the success of GEAs, there is still no common knowledge regarding the properties high-quality representations should have and how representations exactly affect the performance of GEAs. The purpose of the workshop is to discuss the influence of representations on the performance of genetic and evolutionary algorithms. It will focus on theoretical properties of representations as well as on empirical investigations into performance characteristics.

Franz Rothlauf
Department of Information Systems
University of Bayreuth,
Germany
Email: rothlauf@uni-bayreuth.de

and:
Illinois Genetic Laboratory
Department of General Engineering
University of Illinois at Urbana-Champaign,
USA

Further details

Understanding Coevolution:
Theory and Analysis of Coevolutionary Algorithms

R. Paul Wiegand and Kenneth A. De Jong

Duration: Half Day

Coevolutionary algorithms promise several advantages over traditional evolutionary algorithms in terms of their adaptability and potential open-endedness. However, they also challenge us with new and difficult issues. For example, their very adaptability means that fitness assessments in the algorithm are in some sense subjective, and thus the existence of Red Queen dynamics can make it difficult to know whether real progress is being made in any objective sense and pathologies like mediocre stable states can cast doubt on whether optimization is being done at all. More generally, dynamics in these systems can be complicated and surprising. Theory and analysis of coevolutionary algorithms is far less advanced than that of traditional evolutionary algorithms, but the time has come to focus our collective attention on analysis issues more formally. The premise of this workshop is to foster and encourage open discussion about the issues surrounding the direction that analysis of coevolutionary algorithms might take in the future, as well as introducing existing theory and empirically analytical work to those who are looking for a place to start understanding coevolution.

R. Paul Wiegand
Krasnow Institute, MSN 2A1
George Mason University
Rock Fish Creek Lane
Fairfax, VA 22030,
USA
Tel: 703.993.4380
Email: paul@tesseract.org

Kenneth A. De Jong
Computer Science Dept., MSN 4A5
George Mason University
4400 University Drive
Fairfax, VA 22030
USA
Tel: 703.993.1553
Email: kdejong@cs.gmu.edu

Further details