Genetic Algorithms
more info:

Experimental Research in EC
more info:

Mengjie Zhang

Biosketches: Mengjie Zhang received a BE and an ME in 1989 and 1992 from China and a PhD in Computer Science from RMIT University, Melbourne, Australia in 2000. Since 2000, he has been working as a Lecturer then a Senior Lecturer now a Reader/Associate Professor at Victoria University of Wellington, New Zealand. His research is focused on Genetic Programming and Evolutionary Computing, Evolutionary Computer Vision, and Data Mining and Machine learning. He has published over 100 academic papers in refereed international journals and conferences in these areas.

He is a member of ACM, IEEE, IEEE Computer Society, IEEE Computational Intelligence Society, IEEE Systems, Man and Cybernetics Society, and the ACM SIGEVO group. He is also serving as a committee member of the IEEE New Zealand Central Section.

Further Information: http://homepages.mcs.vuw.ac.nz/~mengjie/

Stefano Cagnoni's main research interests are in the fields of Computer vision, Robotics, Evolutionary Computation and Neural Networks. He is editor-in-chief of the "Journal of Artificial Evolution and Applications", chair of EvoIASP, the European Workshop on applications of Evolutionary Computation to Image Analysis and Signal Processing, and member of the Programme Committee of several International Conferences and Workshops. He has been reviewer for over 15 journals.

Further Information: http://www.ce.unipr.it/people/cagnoni/

Gustavo Olague

Gustavo Olague is a research scientist at the CICESE Research Center working within the Computer Science Department of the Applied Physics Division. He hold a Bachelor's degree (Honors) in Electronics Engineering and a Master's degree in Computer Science from the Instituto Tecnológico de Chihuahua, Mexico. He received the Ph.D. (Diplôme de Doctorat en Imagerie, Vision et Robotique) in Computer Graphics, Vision and Robotics from Institut National Polytechnique de Grenoble, France, working at the INRIA Rhône Alpes within the MOVI Research team. He is presently a Professor of Computer Science at Centro de Investigación Científica y de Educación Superior de Ensenada, B.C.

Olague's research focuses on the principles of computational intelligence applied to close-range photogrammetry and computer vision. He is a member of the EvoNET, RSPSoc, ASPRS, SIGEVO, IEEE, IEEE Computer Society, and is listed in Who's Who in Science and Engineering. Dr. Olague is recipient of the "2003 First Honorable Mention for the Talbert Abrams Award", offered by the American Society for Photogrammetry and Remote Sensing. He has been awarded the Bronze Medal at the Human Competitive Awards for his work on synthesis of interest points in 2006.

Further information: http://cienciascomp.cicese.mx/evovision/

Experimental Optimization by Evolutionary Algorithms:

Description of Tutorial

This tutorial addresses applications of evolutionary algorithms to optimization tasks where the function evaluation cannot be done through a computer simulation, but requires the execution of an experiment in the real world (i.e., cosmetics, detergents, wind tunnel experiments, taste experiments, to mention a few). Those applications can, as we show in the tutorial, be handled very successfully by means of evolutionary algorithms.

To start with, the tutorial provides an overview of the landmark experimental optimization carried out by means of evolutionary algorithms, since their foundation in the late 60's. The main characteristics of such experimental work, in comparison to optimization of simulated systems, are discussed and practical guidelines for real-world experiments with EAs are outlined.

Special emphasis will be put on the first experimental work carried out at the Technical University Berlin, as well as on experimental quantum control, an emerging field in physics, with a presentation of state-of-the-art experiments.

In addition, some arguments are given for the usefulness of evolutionary algorithms in comparison to classical approaches (Design of Experiments), and we will discuss the role of noise in experimental optimization.

Ofer Shir

Biosketch: Ofer Shir, PhD, is currently a Postdoctoral Research Associate at Princeton University, conducting research at the Rabitz Group, within the Department of Chemistry. Ofer received his B.Sc. in Physics and Computer Science from the Hebrew University of Jerusalem in June 2003. His graduate studies were in Leiden University, The Netherlands, where he studied Niching in Evolution Strategies under Prof. Thomas Bäck. He also played a role of a Teaching Assistant in the Computer Science Department of Leiden University during the years 2005-2007.

His research activities included a joint project with Amolf-Amsterdam, where he collaborated with the XUV group of Prof. Marc Vrakking. His PhD in Computer Science from Leiden University was conferred in June 2008. His current topics of interest include Natural Computing, Evolutionary Algorithms, Experimental Optimization, Quantum Control and Quantum Information.

Synthetic Biology II: Modeling & Optimization:

Description of Tutorial

The ultimate goal of systems biology is the development of executable in silico models of cells and organisms. Systems biology attempts to provide an integrative methodology, which while able to cope with -on the one hand- the data deluge that is being generated through high throughput experimental technologies -and on the other hand- emerging technologies that produce scarce often noisy data, would allow to capture within human-understandable models and simulations novel biological knowledge.

In its more modest instantiations, Systems Biology seeks to *clarify* current biological understandings by formalizing what the constitutive elements of a biological system are and how they interact with each other and also it seeks to aid in the *testing* of current understandings against experimental data. In its most ambitious incarnations, however, it aims at *predicting* the behavior of biological systems beyond current understanding and available data thus shedding light onto possible new experimental routes that could lead to better theoretical insights.

Synthetic biology, on the other hand, aims to implement, in vitro/vivo, organisms whose behavior is engineered. The field of synthetic biology holds a great promise for the design, construction and development of artificial (i.e. man-made) biological (sub)systems thus offering viable new routes to genetically modified organisms, smart drugs as well as model systems to examine artificial genomes and proteomes. The informed manipulation of such biological (sub)systems could have an enormous positive impact on our societies, with its effects being felt across a range of activities such as the provision of healthcare, environmental protection and remediation, etc. The basic premise of synthetic biology is that methods commonly used to design and construct non-biological systems, such as those employed in the computational sciences and the engineering disciplines, could also be used to model and program novel synthetic biosystems. Synthetic biology thus lies at the interface of a variety of disciplines ranging from biology through chemistry, physics, computer science, mathematics and engineering.

In this tutorial I will provide an entry level understanding to Systems and Synthetic Biology, it goals, methods and limitations. Furthermore I will describe the many potential applications of evolutionary computation to these two fields. Indeed, I believe that the EC community has a beautiful new application domain in which its methods could be both valued and challenged.

More information is available at www.cs.nott.ac.uk/~nxk

Natalio Krasnogor

Biosketch: Natalio Krasnogor is an Associate Professor in the School of Computer Science at the University of Nottingham. He is a member of the Automated Scheduling, Optimisation and Planning Research Group (ASAP) and also leads the Interdisciplinary Optimisation Laboratory (IOL). Krasnogor's research activities lie at the interface of Computer Science and the Natural Sciences, e.g. Biology, Physics, Chemistry. In particular, he develops innovative and competitive search methodologies and intelligent decision support systems for transdisciplinary optimisation, modelling of complex systems and very-large datasets processing.

Synthetic Biology I: Fundamentals & Devices:

Description of Tutorial

The new field of Synthetic Biology has opened the doors to engineers to design, model, build and test cells with particular behaviours, including computational behaviours. The attraction of using cells as devices stem from the size, robustness, reproducibility of cells, as well as the many built-in input and output interfaces that cells have or can be made to have.

This talk is divided to two parts. First comes (Varin) with an introduction to the biological fundamentals necessary for the full appreciation of the second part (given by Kharma). Topics touched upon in the first part include:

* Gene structure and function;
* Gene regulation;
* Recombinant DNA technologies;
* Construction of synthetic gene networks.

The second part provides examples of how biologists and engineers have successful modified the genomes of various cell types to synthesize cellular computing devices and systems. These examples include (but are not limited to):

* Basic and more complex logic circuits;
* Clocks and sequential circuits;
* Extension of functionality via inter-cellular signaling;
* Means of configuring cells to give them different functionalities.

The importance of synthetic biology to evolutionary computation is that it provides engineers with an opportunity to study and be inspired by the intricacies of real life, while forging ahead with their endeavour to emulate it, simulate it, as well as ultimately create it.

Cartesian Genetic Programming:

Description of Tutorial

Cartesian Genetic Programming is a form of genetic programming. It is increasing in popularity. It was developed by Julian Miller with Peter Thomson in 1997. In its classic form it uses a very simple integer based genetic representation of a program in the form of a directed graph. In a number of studies, it has been shown to be comparatively efficient to other GP techniques.
Since then, the classical form of CGP has been enhanced in various ways by to include automatically defined functions. Most recently it has been developed by Julian Miller, Wolfgang Banzhaf and Simon Harding to include self-modification operators. This again has increased its efficiency.

The tutorial will cover the basic technique, advanced developments and applications to a variety of problem domains.

Julian Miller

Biosketches: Julian Miller is a lecturer in the Department of Electronics at the University of York. His main research interests are genetic programming (GP), and computational development. He has published over 130 refereed papers on evolutionary computation, genetic programming, evolvable hardware, and computational development. He has been chair or co-chair of twelve conferences or workshops in genetic programming, computational development, evolvable hardware and evolutionary techniques.

For further info on Cartesian GP see:

http://www.cartesiangp.co.uk/

Bio-inspired Telecommunications:

Description of Tutorial

Biological systems show a number of properties, such as self-organization, adaptivity, scalability, robustness, autonomy, locality of interactions, distribution, which are highly desirable to deal with the growing complexity of current and next generation networks. Therefore, in recent years a growing number of effective solutions for problems related to networked systems have been proposed by taking inspiration from the observation of natural systems and processes such as natural selection, insect societies, immune systems, cultural systems, collective behaviors of groups of animals/cells, etc.

The aim of the tutorial is to provide the cutting edge research on Bio/Nature inspired approaches to network-related problems. The tutorial will also focus on protocol engineering in order to introduce different frameworks that have been developed to realize such Bio/Nature inspired protocols inside the network stack of the Linux kernel. The designers of routing protocols can use it to verify their Linux model by comparing important performance values obtained from Linux model with the simulation model. Moreover, we also focus on "Formal modeling of Bio-inspired algorithms". The tutorial will also highlight Bio/Natue inspired solutions to MANETs and Sensor Networks. Finally, we will conclude the tutorial with an emerging domain of "Bio-inspired Security Solutions for Enterprize Security".

We believe that the tutorial will be instrumental in highlighting the potential of Bio/Nature inspired protocols in real world networks. The tutorial is intended for telecommunication managers, protocol developers, network engineers, network software developers and optimization researchers and graduate students who want to work in non-linear real time dynamic problems. For further details of the tutorial refer to the web site

http://www.nexginrc.org/muddassar/tutorial-gecco09.html

The following references would significantly help in understanding the tutorial:

1.http://www.springer.com/computer/artificial/book/978-3-540-85953-6 2.http://www.springer.com/computer/artificial/book/978-3-540-74088-9 [Chapter 4]

Muddassar Farooq

Biosketch: Muddassar Farooq received his B.E. degree in Avionics Engineering from National University of Sciences and Technology (NUST), Pakistan, in 1996. He completed his M.S. in Computer Science and Engineering from University of New South Wales (UNSW), Australia, in 1999. He completed his D.Sc in Informatics from Technical University of Dortmund, Germany, in 2006. In 2007, he joined the National University of Computer & Emerging Sciences (NUCES), Islamabad, Pakistan, as an associate professor. He is also the Director of Next Generation Intelligent Networks Research Center (nexGIN RC) at NUCES. He is the author of the book “Bee-inspired Protocol Engineering: from Nature to Networks” that is published by Springer in 2008.

He has also has coauthored two book chapters in different books on swarm intelligence. He is on the editorial board of Springer’s Journal of Swarm Intelligence. He is also the workshop chair of European Workshop on Nature-inspired Techniques for Telecommunication and Networked Systems (EvoCOMNET) held with EuroGP. He also serves on the PC of well-known EC conferences like GECCO, CEC, ANTS. He is the guest editor of a special issue of Journal of System Architecture (JSA) on Nature-inspired algorithms and applications. His research interests include agent based routing protocols for fixed and mobile ad hoc networks (MANETs), nature inspired applied systems, natural computing and engineering and nature inspired computer and network security systems i.e. artificial immune systems.

Theory of Randomized Search Heuristics in Combinatorial Optimization :

Description of Tutorial

The rigorous mathematical analysis of randomized search heuristics (RSHs) with respect to their expected runtime is a growing research area where many results have been obtained in recent years. This class of heuristics contains well-known approaches such as Randomized Local Search (RLS), the Metropolis Algorithm (MA), Simulated Annealing (SA), and Evolutionary Algorithms (EAs) as well as more recent approaches such as Ant Colony Optimization (ACO). Such heuristics are often applied to problems whose structure is not known or if there are not enough resources such as time, money, or knowledge to obtain good specific algorithms. It is widely acknowledged that a solid mathematical foundation for such heuristics is needed.

Most designers of RSHs, however, rather focused on mimicking processes in nature (such as evolution) rather than making the heuristics amenable to a mathematical analysis. This is different to the classical design of (randomized) algorithms which are developed with their theoretical analysis of runtime (and proof of correctness) in mind. Despite these obstacles, research from the last about 15 years has shown how to apply the methods for the probabilistic analysis of randomized algorithms to RSHs. Mostly, the expected runtime of RSHs on selected problems is analzyed. Thereby, we understand why and when RSHs are efficient optimizers and, conversely, when they cannot be efficient.

The tutorial will give an overview on the analysis of RSHs for solving combinatorial optimization problems. Starting from the first toy examples such as the OneMax function, we approach more realistic problems and arrive at analysis of the runtime and approximation quality of RSHs even for NP-hard problems. Our studies treat not only simple RLS algorithms and SA but also more complex population-based EAs and even ACO algorithms. The combinatorial optimization problems that we discuss include the maximum matching problem, the partition problem and, in particular, the minimum spanning tree problem as an example where Simulated Annealing beats the Metropolis algorithm in combinatorial optimization. Important concepts of the analyses will be described as well.

Carsten Witt

Biosketch: Carsten Witt studied Computer Science at the University of Dortmund, Germany, where he received his diploma and Ph.D. in 2000 and 2004, respectively. Since spring 2009, he is an assistant professor at the Technical University of Denmark in Copenhagen. Carsten's main research interests are the theoretical aspects of randomized search heuristics, in particular evolutionary algorithms and ant colony optimization. More information: http://ls2-www.cs.uni-dortmund.de/~witt

Fitness Landscapes and Graphs: Multimodularity, Ruggedness and Neutrality:

Description of Tutorial

The performances of evolutionary algorithms (genetics algorithms, genetic programming, etc.) or local search algotihms (Simulated annealing, tabu search, etc.) depends on the properties of seach space structure. One concept to analyse the search space is the fitness landscapes in which the problem to optimize and the search algorithm are taken into account. The fitness landscape is a graph where the nodes are the potential solutions. The study of the fitness landscape consists in analysing this graph or a relevant partition of this graph according to the dynamic or search difficulty.

This tutorial will give an overview, after an historical review of concept of fitness landscape, of the different ways to define fitness landscape in the field of evolutionary computation. Following, the two mains geometries (multimodal and neutral landscapes) corresponding to two different partitions of the graph, meets in optimization problems and the dynamics of metaheuristics on these will be given. The relationship between problems difficulty and fitness landscapes measures (autocorrelation, FDC, neutral degree, etc.) or the properties of the local optima networks, studied in recent work, will be deeply analysed. Finally, the tutorial will conclude with a brief survey of open questions and the recent researchs on fitness landscapes.

Sébastien Verel

Biosketch: Sébastien Verel Born in Lisieux (France) on October 29th, 1976. PhD in Computer Science at the University of Nice Sophia-Antipolis (France), ``Study and exploitation of neutral networks in fitness landscapes for hard optimization'', achieved in December 12th, 2005. Current position: Permanent Researcher since 2006 by computer science departement, university of Nice Sophia-Antipolis (France)

• Theory of evolutionary computation:
  study and measure of hardness of fitness landscapes specialy neutral fitness landscapes.
  • Proposition of fitness/fitness correlation to study dynamic and hardness (fitness cloud),
  • Proposition and analysis of the local optima networks,
  • Analysis of complex neutral fitness landscapes in GA and GP (majority problem, firing squad problem, etc.),
  • Proposition of metaheuristic for neutral fitness landscapes (scuba search).
• Complex Systems:
  where some "global" properties of the system comes from a large number of "local" interactions.
  • Study of the majority problem in cellular automata,
  • Design of new GA for majority problem.
• Cellular Genetic Algorithm:
  evolutionary algorithms where the population is structured by a grid or a graph.
  • Proposition of new selection selection method in Cellular GA,
  • Analysis of diversity and convergence time in Cellular GA.
• Cognitive sciences:
  design of models in cognitive area with the help of evolutionary computation
  • Proposition of Eyes tracking model
  • Proposition of a model on a combinaison of two cognitive tasks: memorization and information retrieval
  
  

Fitness Landscapes and Problem Hardness in Genetic Programming

Biosketch:

Main scientific contributions:
• Problem Hardness in Genetic Programming (GP). I have proposed some new hardness indicators for GP and other measures to characterize fitness landscapes.
•Parallel and Distributed GP. I have studied how the island GP model enables to maintain a higher diversity in the whole GP system, thus producing better quality solutions for many applications.
• Other Techniques to Improve GP Performance. The techniques that I have studied are: (1) reducing the number of test cases to evaluate fitness by means of statistics, (2) using populations which dynamically adapt their size according to some events happening during the evolution, (3) coevolving GP terminal symbols by means of Genetic Algorithms (GAs).
• Drug Discovery by Genetic Programming: I have realized a GP framework to mine large datasets of molecule aimed at discovering the model explaining the relationship between molecular descriptors and some important drug features, like toxicity, bioavailability or docking energy.
• Traffic Control Applications using Machine Learning. Using image sequences taken by a fixed camera, I have realized a system to recognize, classify and track the vehicles traveling along a street.
Techniques used are Neural Networks, GAs and IPAN Tracker. The system has been embedded with a television camera system built by the Comerson Srl and it is presently in use in some city crossroads and motorways.
• Genetic Algorithms for the Chemical Formulation Problem. This problem is very important to build new and performing chemical compounds and drugs. I have formalized it as an optimization problem and solved it in efficient way by means of Evolutionary Algorithms.

Web page:
http://personal.disco.unimib.it/Vanneschi