1. Workshop On Challenges in Real World Optimisation Using Evolutionary Computing
2. Workshop on Games and Emergent Behaviors in Distributed Computing Environments
3. Workshop on Nature Inspired Approaches to Networks and Telecommunications
4. Workshop on Intelligence Before Training Neural Nets (cancelled)
   Workshop on Evolving Neural Networks
5. Workshop on Foundations of Learning Classifier Systems
6. Workshop on Future Directions for Learning Classifier Systems
7. Workshop on Memetic Algorithms

1 Background

Optimisation algorithms are becoming increasingly popular for solving real-life problems. They are extensively used in those problems where the emphasis is on maximising or minimising a certain goal. Whilst the traditional techniques have been used with considerable success to tackle a wide variety of applications, everyone of these, without exception, can only optimise existing designs and is application specific. The need for developing a compact package of robust optimisers has led to the growth of evolutionary computation techniques.

The aim of this workshop is to explore the use of evolutionary computation techniques for solving real-life optimisation problems. These problems pose additional challenges for the optimisation techniques due to their following characteristics:

• The principal feature of most real-life problems is the presence of multiple measures of performance, or objectives, which should be optimised simultaneously.
• Most of these problems are difficult to solve due the presence of multiple interacting decision variables.
• In most of these problems, there is no prior knowledge regarding the shape of search space. There is also no prior information about the performance and location of the optimal and sub-optimal points in the search space.
• The complexity of these problems is also increased due to the qualitative issues, like manufacturability and designers’ special preferences, invariably associated with real-life problems.
• Furthermore, most of these problems are multi-modal and require some constraints to be satisfied.
• Finally, the model development for the solution of real-life optimisation problems is a very complex task.

These characteristics of real-life optimisation problems have provided an impetus to the growth of evolutionary-based optimisation algorithms. Research is currently underway throughout the world to explore the application of these techniques to a variety of real -life optimisation problems.

It is the purpose of this workshop to bring together researchers working in the area of industrial application of evolutionary-based computation techniques like genetic algorithms, evolutionary programming, genetic programming and evolutionary strategies. The workshop would provide a great opportunity for presenting and disseminating latest work in the fields of multi-objective, multi-modal and constraint optimisation. It would provide a forum for identifying and exploring the key issues that affect the industrial application of evolutionary-based computation techniques.

The topics of the workshop include, but are not limited to:

• Multi-objective Optimisation.
• Multi-modal Optimisation.
• Constraint Optimisation.
• Evolutionary Computing.
• Evolutionary Programming and Evolutionary Strategies.
• Hybrid Optimisation Techniques.
• Optimisation in Unknown Search Space.
• Optimisation of High Dimensional Problems.
• Variable Interaction in Multi-objective Optimisation Problems.
• Integrating Qualitative Knowledge in Optimisation.
• Real-life Applications of Evolutionary Computing.
• Inhibitors to Industrial Applications of Evolutionary-based Optimisation Algorithms.
• Training Requirements for Popularising Evolutionary Computing in Industry.

2 Workshop Format

The workshop format would be as follows:

• The workshop is proposed to be for half a day consisting of individual presentations and an open group discussion.
• Each individual presenter would be given 15 minutes to present his/her work and another 5 minutes for discussion.
• The last hour of the workshop would be utilised for discussing the issues related to the topic. This open group discussion would be an interactive session, involving the use of adhesive notes and overhead projector. It would give the participants an opportunity to comment and express their views on the topic.

It is expected that about 20 to 25 participants would be attending this workshop. A substantial proportion of these participants is expected to be from industry.

3 Further details

The workshop would invite papers of not more than six pages. An international programme committee would be set up by the workshop organisers for selecting the papers.

1 Description of the Topic and Focus:

This workshop will provide a unique forum for researchers working on theory and applications of game theory, evolutionary algorithms, multi-agent systems, artificial intelligence on distributed computing to meet and discuss the emerging field of rationality-based computing and agent-based distributed computing middleware.

The main topics of the workshop are:

Emergent behaviors in massively distributed and rationality-based computing environments and solutions and research on optimal computing resource allocations and utilizations in these environments, Massively distributed computing as self-organizing systems.

Additional topics are:

Applications of game theory, evolutionary algorithms, multi-agent systems, and artificial intelligence techniques on distributed real-time computing, Internet-wide rationality-based distributed computing, multi-agent based computing, Peer-to-Peer resource sharing in untrusted distributed network, Resource sharing and allocations in massively parallel distributed systems, rationality-based computing, rationality-based mobile ad-hoc networks, Load-balancing over the Internet, Game theoretical data distribution, migration and replication techniques under untrusted P2P systems, and Rationality-based computing middleware for distributed real-time systems.

2 Workshop Format

The workshop is a full-day workshop consisting a morning and an afternoon sessions. The workshop is designed to foster and encourage discussions on the above topics and on how the researchers in the PPSN community can contribute and involve more to the computer systems research.

Each paper presentation will be approximately 25 minutes including discussions. In addition to paper presentations of the accepted papers, we plan to have:

1. invited speakers to expose important problems from the distributed systems perspective,
2. panel discussions to discuss a few possible solutions,
3. small group discussions on potential solutions to the defined and refined problems.

We also plan to organize a journal special issue with selected papers from the workshop proceedings.

3 Workshop Details

Important Dates:

Paper format:

Papers solicited are short extended abstracts (6 pages maximum with the same format as the main conference paper format: Springer Verlag LNCS style, http://www.springer.de/comp/lncs/authors.html). All the papers should be in PDF, PS, or WORD format. Send papers to Dr. Jae C. Oh (jcoh@ecs.syr.edu) via email as an attachment.

1 Workshop Overview

The great expansion of telecommunications networks and networking applications have given rise to a large number of associated problems. Traditional methods are often not suitable to solve these new challenges. Newly emerging computational heuristics inspired by nature have provided very good results for tackling these problems. This Workshop would focus on the application of nature-inspired approaches to Telecommunication/Networking problems, and would be a forum for scientists and engineers from academia and industry to discuss their latest research results on this topic.

Topics of interest would include, but would not be limited to:

• General Network Design Problems
  • Physical Topology Design Problems
  • Survivability and Reliability
  • Quality of Service
  • Protection and Restoration
  • Network Management
  • Congestion Control
  • Simulation and Queuing Models
• Optical networks
  • Routing and wavelength assignment
  • Traffic grooming
  • Placement of wavelength converters
  • Placement of optical amplifiers
  • Protection and Restoration on Optical Layer
• Fixed Networks Design Problems
• Cellular and Wireless Networks
• Satellite Communications Networks
• Other Topics
  • Ad Hoc networks
  • Bluetooth/Personal Area Networks
  • IP/WDM
  • GMPLS and MPLS
  • Internet Applications

The papers accepted by our workshop will be recommended to be published in International Journal of Computational Intelligence and Applications.

2 Workshop Format

The workshop format will be as follows:
The workshop will be a half day of individual presentations and open group discussion. Each presenter will have 15 minutes to present their work and another 5 minutes for discussion. A general discussion linked to the topics raised will take place in the last hour of the workshop.

3 Workshop Details

Important dates

Submission of Papers

The authors are advised to follow the instructions of the main conference to prepare their papers. All the papers should be in PDF, PS or WORD format and be submitted to either Dr. Y. Xu or Dr. S. Salcedo-Sanz via email as an attachment.

Holland's Learning Classifier System (LCS) is a Machine Learning technique which combines a Genetic Algorithm with Reinforcement Learning. Traditionally within such systems, the simulated evolutionary process searches the space of a given knowledge representation scheme whilst utility is assigned under a trial-and-error learning approach. Learning Classifier Systems have experienced something of a renaissance in recent years after Wilson introduced the eXtended Classifier System (XCS) which makes a number of significant alterations to Holland's algorithm. XCS has been shown to be competitive in a number of real-world domains such as data mining, circuit board design, distributed road traffic junction control and time series forecasting. Current formal understanding of Genetic Algorithms and Reinforcement Learning is significant but understanding of how the two interact within Learning Classifier Systems is severely lacking.

The purpose of this half-day workshop is to bring together the growing number of people interested in Learning Classifier Systems and those familiar with current theoretical work in Genetic Algorithms and Reinforcement Learning with the aim of identifying ways to improve our formal understanding of LCS and appropriate formal methods by which to achieve such analysis.

We invite submissions regarding this endeavor - novel contributions and reviews of previous work are sought (Springer LNCS format). An edited volume will be produced after the workshop. Papers should be submitted via email directly to the organizers at the email addresse larry.bull@uwe.ac.uk.

Important Dates

1 Background

Learning Classifier Systems (LCS) are one of the most complex algorithms within Evolutionary Computing as they seek to combine the transparency of production rules with the problem solving capability of evolution. Both cooperative populations of rules and temporal chains of rules may be formed in order to solve a wide range of problems (from data mining to robot control). Research over the last 30 years has led to an understanding of the various methods employed in LCS. Recent research, in the last 10 years, has developed tractable LCS that may be applied to real-world problem.

This workshop is a sister to the workshop on "Foundations of Genetics-Based Machine Learning" organised by Dr Tim Kovacs and Dr Larry Bull. Research appropriate to the understanding of the Theory of LCS is best covered under the foundations workshop. These two workshops will be run sequentially in order that attendees can attend both workshops.

This workshop is intended to discuss the future directions for LCS research. It is not intended as an encapsulated academic conference [please refer to the series of International Workshops on Learning Classifier Systems for such an event]. It is hoped that coherent research strategies, collaborative research and an outline for an introductory/tutorial book will result from this workshop.

The workshop will provide a platform to present novel ideas, initial experimental results and extensions to proven concepts. This forum will allow discussion on the merits of approaches, including lessons learnt from past experience in developing LCS. It is intended that the attendees will represent decades of person years research in the field.

General topics for the workshop include:

1. Latest developments in LCS.
2. Research areas in LCS requiring further development.
3. Applications for LCS, both test problems and real-world problems.
4. Simplifying the LCS learning curve for new researchers in the LCS field.

Example topics under the general headings could include, but are not limited to:

• Use of confusion matrices to guide the learning within LCS
• Rule linkage mechanisms for temporal environments
• Appropriate fitness measures for stated tasks
• Control of mobile robots utilising temporal based LCS
• Data mining within medical databases
• Industrial problems suited to identified LCS characteristics
• "How to Build Learning Classifier Systems" resource for new researchers within the field.

2 Workshop Format

The workshop format would be as follows:

• The workshop is proposed to be for half a day consisting of individual presentations and open group discussions.
• Each individual presenter would be given 15 minutes to present his/her work
• Open group discussion will follow after each presentation, being interactive in nature and lasting between 5-15 minutes.

It is expected that about 20 to 25 participants would be attending this workshop. Industry participation is encouraged, especially with regard to future research topics and application areas.

3 Further details

The workshop would invite papers of not more than six pages. The recommended Springer Verlag LNCS style is advised, preferably in PDF format. An international programme committee would be set up by the workshop organisers for selecting the papers. Papers should be submitted via email directly to the organizers at the email address w.n.browne@reading.ac.uk.

The next international Workshop on Memetic Algorithms (WOMA-V), will be the fifth in a series of workshops dedicated exclusively to Memetic Algorithms and will take place in conjunction with PPSN 2004 in Birmingham UK on Saturday 18 September 2004. The WOMA series is a forum where the international community of researchers, practitioners and vendors, that work on aspects related to memetic algorithms, can engage in fruitful discussions, learning, research and where they can contribute to the advancement of this field. In previous occasions WOMA was co-located with GECCO 2000 in Las Vegas-USA, GECCO 2001 in San Francisco-USA, PPSN VII in Granada-Spain and with GECCO 2003 in Chicago-USA.

Motivation

Memetic algorithms (MAs) are evolutionary algorithms (EAs) that apply a separate local search process to refine individuals (e.g. improve their fitness by hill-climbing). These methods are inspired by models of adaptation in natural systems that combine evolutionary adaptation of populations of individuals with individual learning within a lifetime. Additionally, MAs are inspired by Richard Dawkin's concept of a meme, which represents a unit of cultural evolution that can exhibit local refinement. Thus a memetic model of adaptation exhibits the plasticity of individuals that a strictly genetic model fails to capture. Under different contexts and situations, MAs are also known as hybrid EAs, genetic local searchers, Baldwinian EAs, Lamarkian EAs, etc.

From an optimization point of view, MAs are hybrid EAs that combine global and local search by using an EA to perform exploration while the local search method performs exploitation. Combining global and local search is a strategy used by many successful optimization approaches, and MAs have in fact been recognized as a powerful algorithmic paradigm for evolutionary computing. In particular, the relative advantage of MAs over EAs is quite consistent on complex search spaces.

It is the goal of this new edition of the workshop to push forward our understanding of both the theory and the deployment of MA by engaging in a scholarly dialog with some of the world experts on this field. The format of WOMA this year will be based around a panel of invited experts on Evolutionary-Local Search hybrid algorithms and we are very pleased to confirm the participation in the panel of:

• Prof. D.Whitley (Univ. Of Colorado, USA)
• Prof. C. Coello-Coello (CINVESTAV-IPN, Mexico)
• Prof. P. Ross (Napier Univ., Scotland)
• Prof. C. Reeves (Coventry Univ., England)
• Dr. M. Lozano (Granada Univ., Spain)
• Dr. G. Raidl (Viena Univ. of Technology, Austria)
• Dr. Carlos Cotta (Malaga Univ., Spain)
• ......

Our intentions is to give each panel member between 10 and 15 minutes to informally address the audience on any topic he/she considers relevant and important with regards to these hybrid algorithms, the theory behind them and their application to challenging domains. The panel will be moderated by one of the workshop co-chairs. Once each panel member has had the opportunity to present his/her views we will open the floor in order that the audience can engage with the panel in open and scholarly discussions. It is our intention to have a rich and lively workshop made up of a variety of views and perspectives with regards to the state-of-the-art in research/applications on memetic evolutionary algorithms.

We invite you to bring along your ideas/issues/questions/insights on any of the following themes (but not limited to):

• Memetic algorithms applications, new challenging domains for MAs.
• Theoretical tools and methods likely to be useful for understanding the behaviour, and/or predicting the performacne of MAs, e.g., Kolmogorov, computational and PLS complexity issues, convergence proofs, landscape analysis or any other relevant analytical models and techniques.
• Theoretical/Experimental comparisons/integration with other soft techniques.
• Particular issues arising from the application of MAs to Multiobjective optimisation, discrete optimisation, continuous optimisation, mixed domains, non-stationary problems.
• Frameworks for describing and classifying Mas, Practical guidelines to combine local search and Eas, Scalability of Mas, New MA architectures.
• etc

Details

For details, inquiries, etc about the Fourth Memetic Algorithms Workshop don't hesitate to contact any of the organizers!

This call's web site is located at: http://www.cs.nott.ac.uk/~nxk/WOMAV/call4Participation.htm .

WOMA 5 Co-Chairs:

William E. Hart (Sandia National Labs)
Natalio Krasnogor (University of Nottingham)
Jim E. Smith (University of the West of England)