Genetic Algorithms:

The Genetic Algorithm (GA) track has always been the largest scientific track at GECCO. We invite submissions to the GA track that present original work on all aspects of genetic algorithms, including, but not limited to:

• Theoretical aspects of GAs
• Design of new GAs (including representation and operators)
• Improvements over existing algorithms
• Comparisons with other methods / empirical performance analysis
• Hybrid approaches
• New application areas
• Handling uncertainty (dynamic and stochastic problems, robustness)
• Metamodelling
• Interactive GAs
• Parallelization and other efficiency-enhancement techniques

As a large and diverse track, the GA track will be an excellent opportunity to present and discuss your research/application with experts and participants of GECCO.

to top

Genetic Programming:

Genetic Programming (GP) is the automatic induction of computer programs and other variable-size structures representing executable programs or computable functions from a high-level statement of a < problem through evolutionary algorithms. This track invites submissions of original work in all areas and derivatives of GP. Areas of interest include: theory, algorithm design, and novel representations, operators and algorithms. Authors interested in submitting manuscripts are encouraged to look at previous years' papers.

to top

Both evolution strategies (ES) and evolutionary programming (EP) are nature-inspired optimization paradigms that generally operate on the " natural" problem representation (i.e., without a genotype-phenotype mapping). For example, when used in connection with real-valued problems, both ES and EP use real-valued representations of search points. Moreover, both may rely on sophisticated mechanisms for the adaptation of their strategy parameters. ES and EP owe much of their success to their universal applicability, ease of use, and robustness.

This track invites submissions that present original work on ES/EP that may include, but is not limited to, theoretical and empirical evaluations of ES/EP, improvements and modifications to the algorithms, and applications of ES/EP to benchmark problems and test function suites. Particularly encouraged are submissions with focus on

• adaptation mechanisms
• interesting ES/EP applications
• ES/EP theory
• ES/EP in uncertain and/or changing environments
• comparisons of ES/EP with other optimization methods
• hybrid strategies
• meta-strategies
• constrained and/or multimodal problems

to top

The Real World Applications (RWA) track invites submissions that present rigorous applications of Evolutionary Computation (EC) to real world problems. Of particular interest are:

(1) Papers that describe advances in the field of EC for implementation purposes.
(2) Papers that present rigorous comparisons across techniques in a real world application.
(3) Papers that present novel uses of EC in the real world.
(4) Papers that present new applications of EC to real world problems.

Domains of applications include all industries (e.g., automobile, biotech, chemistry, defense, oil and gas, telecommunications, etc.) and functional areas include all functions of relevance to real world problems (logistics, scheduling, timetabling, design, pattern recognition, data mining, process control, predictive modeling, etc.).

The RWA track differs from the Evolutionary Computation Practice (ECP) workshop in that

(1) RWA only accepts papers with the same high technical and scientific quality as that of the rest of the GECCO track papers. ECP is generally< suitable for researchers and managers from industry, who have less time to write a technical paper but still would like to present significant successes of the technology solving a real-world problem.

(2) Papers accepted in the RWA track will be published in the GECCO 2006 Proceedings. Therefore, if publication is important to you, we suggest you submit your papers to RWA.

to top

Evolutionary methods can be applied to all and we refer to any such work as Genetics-Based Machine Learning (GBML). The GECCO 2007 GBML track encompasses all work on machine learning problems using evolutionary algorithms, whether alone or hybridised with other machine learning methods.

Indeed, submission of work on hybrid systems, and work which otherwise develops connections with non-evolutionary ML methods, are encouraged.

The field of Learning Classifier Systems (LCS), introduced by John Holland in the 1970s, is one of the most active and best-developed forms of GBML and we welcome all work on LCSs.

Recently, GBML has been experiencing a strong renaissance thanks to three key factors: (1) advances in GA theory have not only deepened the understanding of evolutionary learning and optimization but have also enabled the successful analysis of GBML systems; (2) advances in machine learning theory and understanding have enabled further successful and robust combinations of machine learning with evolutionary computation techniques (3) successful applications of GBML systems to real-world problems such as data mining and control problems have confirmed the strength, robustness, and broad applicability of the GBML approach.

The GBML track encourages submissions including but not limited to the following areas.

1. Theoretical Advances in GBML

• Theoretical analysis of mechanisms and systems
• Identification of learning and scalability bounds
• Connections and combinations with machine learning theory
• Analysis and robustness in stochastic (or noisy) environments
• Complexity analysis in MDP and POMDP problems

2. Systems and Frameworks

• Incremental evolutionary rule learning, including but not
limited to:

+ Michigan style (SCS, NewBoole, EpiCS, ZCS, XCS, UCS...)
+ Pittsburgh style (GABIL, GIL, COGIN, REGAL, GA-Miner, GALE, MOLCS, GAssist...)
+ Anticipatory LCS (ACS, ACS2, XACS, YACS, MACS...)

• Genetic-based inductive learning
• Genetic fuzzy systems
• Learning using evolutionary estimation of distribution algorithms
• Evolution of Neural Networks
• Evolution of ensemble systems
• Other hybrids combining evolutionary techniques with other machine learning techniques

3. System Enhancements

• Competent operator design and implementation
• Encapsulation and niching techniques
• Hierarchical architectures
• Default hierarchies
• Knowledge representations, extraction and inference
• Data sampling
• (Sub-)Structure (building block) identification and linkage learning for GBML systems
• Integration of other machine learning techniques

4. Application Areas

• Data mining
• Bioinformatics and life sciences
• Robotics, engineering, hardware/software design, and control
• Cognitive systems and cognitive modelling
• Rapid application development frameworks for GBML
• Dynamic environments
• Time series and sequence learning
• Artificial Life
• Adaptive behaviour

Biological Applications:

The scope of this track is any research that applies evolutionary (and related) techniques to solving biological and biomedical problems. All " flavors" of evolutionary techniques consistent with GECCO are included in this scope, including genetic algorithms, genetic programming, estimation of distribution algorithms, evolution strategies, evolutionary programming, ant colony optimization, swarm intelligence, artificial life and hybrid systems with any of these components. Papers that integrate evolutionary methods into bioinformatics, biomedical informatics, biostatistics, computational biology and systems biology, for example, are particularly welcome. Papers that provide experimental validation and/or biological interpretation of computational results are also particularly welcome.

Some specific examples of biological and biomedical issues that papers could address include:

• Data mining in biological or biomedical databases
• Diagnostic or predictive testing in epidemiology and genetics
• Functional diversification through gene duplication and exon shuffling
• Gene expression and regulation, alternative splicing
• Genetic association studies
• Haplotype and linkage disequilibrium analysis
• Image analysis and pattern recognition
• Metabolomics
• Microarray analysis
• Network reconstruction for development, expression, catalysis etc.
• Pharmacokinetic and pharmacodynamic analysis
• Phylogenetic reconstruction and analysis
• Relationships between evolved systems and their environment (e.g. phylogeography)
• Relationships within evolved communities (cooperation, coevolution, symbiosis, etc.)
• Sensitivity of speciation to variations in evolutionary processes
• Sequence alignment and analysis
• Simulation of cells, viruses, organisms and whole ecologies
• Structure prediction for biological molecules (structural biology)
• Systems biology

to top

Much exciting research in recent years has shown that just as ideas from evolution can be appropriated to create strong algorithms, these algorithms can in turn be used to address issues in the biological realm.

For instance, artificial life research explores questions regarding the necessary and sufficient conditions for lifelike processes to emerge in software. Evolutionary robotics seeks to instantiate lifelike and intelligent behaviors in mechanical devices. Evolvable hardware strives to bring the fluid, robust and adaptable capabilities of biological processes to bear on the problem of creating and maintaining electronic devices.

The aim of this track is to provide a forum for exploring the bidirectional conversation between biology and computer science.

A sampling of more specific issues include

• Necessary and sufficient conditions for simulating lifelike processes
• The role of emergence in complex systems
• Maximizing adaptability and robustness in automatically designed systems
• General principles underlying adaptive behavior
• The role of biological realism in artificial life research

to top

AIS techniques take inspiration of different immunological mechanisms in vertebrates in order to solve computational problems. Various aspects of these mechanisms have been used to develop algorithms for distributed and adaptive control, machine learning, pattern recognition, fault detection, computer security, optimization, and distributed system design.

Submissions of original and previously unpublished work in the following areas of ACO/SI/AIS research are encouraged:

• applications of these algorithms to real-world problems
• applications of these algorithms to scientific test cases
• new theoretical results
• new computational models and techniques
• new hybrids between these algorithms and other methods
• biological foundations
• models of the behavior of social insects

Track website: http://iridia.ulb.ac.be/gecco2007/

to top

Evolutionary Multiobjective Optimization (EMO):

EMO is one of the largest and most significant emergent fields of research and application in the area of genetic and evolutionary computation. Major research, starting around the beginning of nineties, the EMO field has now matured to have standard algorithms, public domain as well as commercial softwares, and many real-world case studies. A significant edge of EMO over classical multiobjective optimization is its ability to find a number of trade-off optimal solutions in a single simulation. The knowledge of many high-performing trade-off solutions not only provides a better decision-making power for choosing a particular solution for implementation, the flexible yet efficient search power of EMO and the information hidden in multiple solutions can also be utilized in other kinds of problem solving tasks effectively. This year's GECCO track on EMO will provide a platform for researchers and applicationists to come and discuss their latest studies in all aspects of EMO by submitting papers through GECCO paper submission guidelines.

The EMO track encourages papers related to, but not limited to, the following aspects of EMO:

1. Theoretical aspects of EMO
2. Test problems, dynamic EMO, scheduling EMO, robust EMO, reliability-based EMO, EMO for handling noise
3. New EMO algorithms and comparative studies
4. EMO performance measures and analysis
5. Show-casing EMO for better decision-making through applications
6. EMO for other kinds of problem-solving tasks -- through use of additional objectives
7. EMO for machine learning, data-mining etc.
8. Hybrid EMO and decision-making
9. Hybrid EMO and classical MCDM methodologies
10. EMO in other evolutionary computing strategies
11. Interactive EMO
12. Review of a specific EMO aspect

EMO tracks within GECCO are usually well-attended. This will be an excellent opportunity to present and discuss your research/application with experts and participants of GECCO and the EMO track.

to top

Coevolution can be an effective approach for problems where performance can be measured using tests, as well as for problems in which multiple components that make up a whole are to be co-adapted. In addition to these forms of optimization, the adaptive nature of the evaluation process in coevolution may in principle give rise to a self-propelled and open-ended evolutionary process.

It has been found early on that the dynamic evaluation of coevolution can lead to unreliability. In recent years however, the possible goals for coevolutionary algorithms have become better understood, and for several algorithms theoretical properties have been provided. These developments generate the exciting prospect that practical reliable algorithms for coevolution may now be within reach.

The Coevolution Track of the 2005 Genetic and Evolutionary Computation Conference provides a venue where researchers from all directions and approaches to coevolution can meet. Submissions on any aspect of coevolution are encouraged, including but not limited to the following:

• Applications
• Measuring progress
• Game-theoretic studies
• New coevolutionary algorithms
• The structure of coevolution problems
• Empirical studies of coevolutionary methods
• Behavioral dynamics of coevolutionary setups
• Theoretical guarantees for coevolutionary algorithms
• Empirical comparisons between coevolutionary and other methods

to top

Estimation of Distribution Algorithms:

Estimation of distribution algorithms (EDAs) replace traditional variation operators of genetic and evolutionary algorithms, such as mutation and crossover, by building a probabilistic model of promising solutions and sampling the built model to generate new candidate solutions. Using probabilistic models for exploration in genetic and evolutionary algorithms enables the use of advanced methods of machine learning and statistics for automated identification and exploitation of problem regularities for broad classes of problems. As a result, EDAs provide a robust and scalable solution to many important classes of optimization problems with only little problem specific knowledge.

Track website: http://medal.cs.umsl.edu/gecco-2007-eda/

to top

• develop and extend the emerging community working on Search-Based Software Engineering
• continue to inform researchers in Evolutionary Computation about problems in Software Engineering
• Increase awareness and uptake of Evolutionary computation technology within the Software Engineering community
• Provide definitions of representations, fitness/cost functions, operators and search strategies for Software Engineering problems.

Topics include (but are not limited to) the application of search- based algorithms to:

Requirements engineering
System and software design
Implementation
Network design and monitoring
Software security
System and software integration
Quality assurance and testing
Project management, control, prediction, administration and organization
Maintenance, change management, optimization and transformation
Development processes

As an indication, `search- based' techniques are taken to include (but are not limited to):

• Genetic Algorithms
• Genetic Programming
• Evolution Strategies
• Evolutionary Programming
• Simulated Annealing
• Tabu Search
• Ant Colony Optimization
• Particle Swarm Optimization

Papers should address a problem in the software engineering domain and should approach the solution to the problem using a heuristic search st rategy. Papers may also address the use of methods and techniques for i mproving the applicability and efficacy of search-based techniques when applied to software engineering problems. While experim ental results are important, papers that do not contain results, but rather present new approaches, concepts and/ or theory will also be considered.

to top

Formal Theory:

The aim of this track is to provide a forum for strong theory papers that up to now were scattered over different tracks. With the current track, we increase the visibility of these works also towards the traditional algorithm theory community. The definition of a theory paper is as follows.

A theory paper considers a particular randomized search heuristic on a particular problem or class of problems. It contains formally stated theorems which are proven rigorously. It is not enough to validate a model or a theory by experiments.

Track website: http://www.mpi-inf.mpg.de/~doerr/gecco.html

to top