Genetic Algorithms Digest Tuesday, February 8, 1994 Volume 8 : Issue 4 - Send submissions to GA-List@AIC.NRL.NAVY.MIL - Send administrative requests to GA-List-Request@AIC.NRL.NAVY.MIL - anonymous ftp archive: FTP.AIC.NRL.NAVY.MIL (Info in /pub/galist/FTP) Today's Topics: - Thesis ftp available on Genetic Neural Networks - GAs and Knowledge Generalizability - Information request - Distributed GA - GAucsd port to Borland C/C++? - WCCI Tutorial Abstracts ---------------------------------------------------------------------- **************************************************************************** CALENDAR OF GA-RELATED ACTIVITIES: (with GA-List issue reference) EP94 3rd Ann Conf on Evolutionary Programming, San Diego (v7n7) Feb 24-25, 94 IEE94 Colloquium on Molecular Bioinformatics, London, UK (v7n21) Feb 28, 94 SPIE, Neural & Stoch. Methods in Image & Sig Proc, Orlando(v7n18) Apr 5-8, 94 FLAIRS-94 Workshop on Artif Life and AI, Pensacola Beach, FL(v7n23) May 4, 94 The IEEE Conference on Evolutionary Computation, Orlando(v7n26) Jun 26-30, 94 FOGA94 Foundations of GAs Wkshop, Estes Park, Colorado(v7n26)Jul 30-Aug 3, 94 SAB94 3rd Intl Conf on Sim of Adaptive Behavior, Brighton(v7n11) Aug 8-12, 94 ECAI-94, 11th European Conference on AI, Amsterdam (v7n23) Aug 8-12, 94 IEEE/Nagoya Univ WW Wkshp on Fuzzy Logic & NNs/GAs, Japan(v7n33) Aug 9-10, 94 ISRAM94 Special Session on Robotics & GAs, Maui, Hawaii (v7n22) Aug 14-17, 94 COMPLEX94 2nd Australian National Conference, Australia (v7n34) Sep 26-28, 94 PPSN-94 Parallel Problem Solving from Nature, Israel (v7n32) Oct 9-14, 94 (Send announcements of other activities to GA-List@aic.nrl.navy.mil) **************************************************************************** ------------------------------ From: Frederic Gruau Date: 23 Jan 94 10:45:46-0100 Subject: Thesis ftp available on Genetic Neural Networks The laboratory for Parallel Computer Science (LIP) is pleased to inform that the following PhD Thesis is available by anonymous FTP. The title is: NEURAL NETWORK SYNTHESIS USING CELLULAR ENCODING AND THE GENETIC ALGORITHM by Frederic GRUAU. -----ABSTRACT------ Artificial neural networks used to be considered only as a machine that learns using small modifications of internal parameters. Now this is changing. Such learning method do not allow to generate big neural networks for solving real world problems. This thesis defends the following three points: (1) The key word to go out of that dead-end is "modularity". (2) The tool that can generate modular neural networks is cellular encoding. (3) The optimization algorithm adapted to the search of cellular codes is the genetic algorithm. The first point is now a common idea. A modular neural network means a neural network that is made of several sub-networks, arranged in a hierarchical way. For example, the same sub-network can be repeated. This thesis encompasses two parts. The first part demonstrates the second point. Cellular encoding is presented as a machine language for neural networks, with a theoretical basis (it is a parallel graph grammar that checks a number of properties) and a compiler of high level language. The second part of the thesis shows the third point. Application of genetic algorithm to the synthesis of neural networks using cellular encoding is a new technology. This technology can solve problems that were still unsolved with neural networks. It can automatically and dynamically decompose a problem into a hierarchy of sub-problems, and generate a neural network solution to the problem. The structure of this network is a hierarchy of sub-networks that reflects the structure of the problem. The technology allows to experience new scientific domains like the interaction between learning and evolution, or the set up of learning algorithms that suit the GA. The anonymous ftp is: 140.77.1.11 (lip.ens-lyon.fr) the directory is: pub/Rapports/PhD The file PhD94-01-E.ps.Z is an english version The file PhD94-01-F.ps.Z is a french versin ------------------------------ From: PAKATH@UKCC.uky.edu Date: Wed, 29 Dec 93 11:19:37 EST Subject: GAs and Knowledge Generalizability Dear Colleagues: Could any of you guide me to available literature, both theoretical and applied, on knowledge generalization by GA-based systems? That is, I am interested in work that studies the ability of GA-based systems to fruitfully apply knowledge (e.g., population members, GA operator probability levels) gathered in one or more prior problem solving episodes to other "similar" episodes, subsequently, where the term "fruitfully" refers to increased problem solving effectiveness and/or efficiency? In the event that this generally viewed by the GA research communit y as a topic not worthy of study, could anyone indicate why? I thank you for your interest and trouble and shall post all "interesting" responses that I receive. [Ed's Note: John Grefenstette and I have applied case-based methods in initializing the population of GAs that are used to guide search in changing environments. We have obtained very good results by using good members of populations that learned in previous similar cases (as defined by environmental parameters) to seed the current population. A paper on this topic appeared in ICGA-93. Zhou applied case-based methods to classifier systems, and Whitley, Mathias and Fitzhorn have a paper on Delta Coding, where they iteratively use previous best solutions as a basis for altering the representation used during a restart. References: Ramsey, C.L. and Grefenstette, J.J. (1993). Case-Based Initialization of Genetic Algorithms. Proc of the 5th Intl Conf on Genetic Algorithms. Whitley, D., Mathias, K. and Fitzhorn, P. (1991). Delta Coding: An Iterative Search Strategy for Genetic Algorithms. Proc of the 4th Intl Conf on Genetic Algorithms. Zhou, H.H. (1990). CSM: A Computational Model of Cumulative Learning. Machine Learning 5(4), 383-406.] ------------------------------ From: capo@imdes02.dees.unict.it Date: Fri, 7 Jan 1994 19:20:49 +0100 Subject: Information request I would like to know if it exists an optimization software tool, GAs based, able to manipulate "about" one thousand variables. Thanks in advance Riccardo Caponetto PhD Student at Universita' di Catania, Italy Email dees@dees.unict.it (please specify my name) ------------------------------ From: jvr@micas.inescn.pt (Joao Vasco Ranito) Date: Fri, 7 Jan 94 16:38:22 GMT Subject: Distributed GA Hi. Does anyone have some papers on Distributed Genetic Algorithms available by ftp? It is not easy, here in Portugal, to get Proceedings and so on, so we are more or less confined to ftp. If someone could give me some pointers, I would be very gratefull. Oh! And I will publish the list on the Digest, of course... My e-mail is: jfn@micas.inescn.pt Thanks in advance. Joao Neto Joao Filipe Neto e-mail: jfn@micas.inescn.pt INESC fax : + 351-2-318692 Largo Mompilher, 22 tel : + 351-2-2094015 4000 Porto Portugal ------------------------------ From: Nici Schraudolph Date: Sat, 15 Jan 94 19:21:19 PST Subject: GAucsd port to Borland C/C++? > From: osherson@idiap.ch (Daniel Osherson) > > For various reasons, I'd like to port the system to > Borland C/C++. Since the code you provide is in an older > version of C, this appears to be rather difficult. > > Does there happen to be a version of GAucsd tailor-written > for Borland C/C++ ? Has anybody out there done this port? - Nici. ------------------------------ From: john@cs.ucsd.edu (John McInerney) Date: Thu, 6 Jan 94 15:58:41 -0800 Subject: WCCI Tutorial Abstracts ***************************************************************************** WCCI TUTORIAL ABSTRACTS ***************************************************************************** FRESH, NEW, CUTTING EDGE TUTORIALS on ... # NEURAL NETWORKS # FUZZY TECHNOLOGY # EVOLUTIONARY PROGRAMMING At the ... 1994 IEEE World Congress on COMPUTATIONAL INTELLIGENCE Walt Disney World Dolphin Hotel Orlando, FLA June 26 - July 2, 1994 [Ed's Note: This message has been shortened due to space constraints. The full message, containing the descriptions of each tutorial, is available from the ftp server, ftp.aic.nrl.navy.mil in the file /pub/galist/info/conferences/WCCI-tutorial-94. -- Connie] ***************************************************************************** "Evolution Strategies: A Thorough Introduction" Professor Thomas Baeck Computer Science Department, LS XI University of Dortmund, Dortmund, Germany In addition to Genetic Algorithms and Evolutionary Programming, the Evolution Strategy (Evolutionsstrategie) by Rechenberg and Schwefel forms the third major representative of Evolutionary Algorithms. Since its development in the 1960s at the Technical University of Berlin (Germany) for solving experimental optimization problems, the computer algorithm has been successfully applied to numerous hard continuous parameter optimization problems (an application field where Evolution Strategies reveal their strengths in comparison to the more familiar Genetic Algorithms). The tutorial presents a thorough introduction to Evolution Strategies, with special emphasis on the history of evolution strategies, detailed presentation and explanation of the algorithm, genetic operators and parameter settings, self-adaptation of strategy parameters, theory of evolution strategies, selected application examples of evolution strategies, evolution strategies for neural networks and fuzzy logic, guidelines for practitioners, and comparison to genetic algorithms and evolutionary programming. ***************************************************************************** "An Introduction to Evolutionary Computation" David B. Fogel, Ph.D. 1591 Calle De Cinco La Jolla, CA The impact of evolutionary thinking on biology cannot be underestimated. Indeed, many biologists have remarked that the study of life cannot be conducted reasonably in the absence of an evolutionary paradigm. But evolutionary thought extends beyond an ordering principle of biology. Evolution is a process that can be simulated on a computer and used for solving difficult engineering problems and gaining insight into natural evolved systems. This tutorial, aimed at researchers in neural networks and fuzzy systems, will introduce methods of evolutionary computations These include genetic algorithms, evolution strategies and evolutionary programming, as well as related techniques. The fundamental philosophical foundations of the methods will be discussed and applications will be described, including synergistic efforts of combining evolutionary optimization with connectionist and fuzzy systems. ***************************************************************************** "Genetic Algorithms and Their Applications" Dr. Lawrence "David" Davis, President Tica Associates Cambridge, MA 02139 Genetic algorithms are techniques for optimization and machine learning that have been applied to a wide range of real-world problems. This tutorial consists of an overview of genetic algorithms, a discussion of techniques for applying them, a survey of areas in which they have been applied, and several application case studies. Particularly stressed in the tutorial will be traditional and nontraditional genetic algorithms for numerical function optimization; the use of order-based genetic algorithms for combinatorial optimization; and techniques for hybridizing genetic algorithms with other optimization algorithms. ***************************************************************************** "Genetic Programming" Dr. John R. Koza Consulting Professor Computer Science Department, Stanford University Genetic programming extends the genetic algorithm to the domain of computer programs and genetically breeds populations of computer programs to solve problems. Genetic programming can solve problems of system identification, optimal control, pattern recognition, equation solving, game playing, optimization, and planning. Starting with hundreds or thousands of randomly created programs, the population is progressively improved by applying Darwinian fitness proportionate reproduction and crossover (sexual recombination). Many problem environments have regularities, symmetries, and homogeneities that can be exploited in solving the problem. The recently developed facility of automatic function definition enables genetic programming to dynamically decompose a problem into simpler subproblems, solve the subproblems, and assemble original problem. Experimental evidence suggests that automatic function definition reduces the computation effort needed to solve a problem and produces a simpler and more understandable overall solution. Portions of videotapes on genetic programming will be shown. ***************************************************************************** "Genetics-Based Machine Learning in Rule-Based and Neural Systems" Professor Robert E. Smith Department of Engineering Science and Mechanics The University of Alabama, Tuscaloosa, Alabama This tutorial covers the application of genetic algorithms (GAs) in machine learning. Machine learning is introduced in the framework of control, with an emphasis on reinforcement learning, where the system must learn through a exploration. A brief overview of GAs is also provided. Given this background, the tutorial discusses rule-based, neural, and fuzzy techniques that utilize GAs. A rule-based technique, the learning classifier system (LCS), is shown to be analogous to a neural network. The integration of fuzzy logic into the LCS is also discussed. Research issues related to GA-based learning are overviewed. The application potential for genetics-based machine learning is discussed. ***************************************************************************** "Genetic Algorithms: Theoretical Foundations and Experimental Evaluation" Professor Darrell Whitley, Computer Science Colorado State University, Fort Collins, CO 80523 The principle of hyperplane sampling will be examined, as well as exact theoretical models of a canonical genetic algorithm. Other topics include: deception, remapping hyperspace, stochastic hill-climbing versus hyperplane sampling and the case against gray coding for test functions. Holland's schema theorem and the K-arm bandit analogy will be reviewed and critiqued. Alternative forms of the genetic algorithm such as Genitor, CHC, Evolution Strategies and parallel genetic algorithms will be reviewed. The practical implications of the existing theory will be explored with respect to implementing and applying genetic algorithms to complex problems. Examples are given where simple theoretical insights result in improved search on problems of more than 500 variables. ***************************************************************************** "An Introduction to Fuzzy Logic" Professor James Bezdek Department of Computer Science University of West Florida, Pensacola, Florida ***************************************************************************** "Fuzzy Sets in Constraint Satisfaction" Didier Dubois Institut de Recherche en Informatique de Toulouse Universite Paul Sabatier, Toulouse Cedex - France ***************************************************************************** "Fuzzy Logic in Computer Vision" Professor James M. Keller Electrical and Computer Engineering Department University of Missouri-Columbia, Columbia, MO ***************************************************************************** "Fuzzy Logic Applications to Artificial Intelligence and Intelligent Control Systems" Enrique H. Ruspini Artificial Intelligence Center SRI International Menlo Park, CA ***************************************************************************** "Fuzzy Neurocomputations" Witold Pedrycz Dept. of Electrical and Computer Eng. University of Manitoba, Winnipeg ***************************************************************************** "Fuzzy Data Analysis" Prof. Dr. Dr.h.c.Hans-Jurgen Zimmermann Professor of Operations Research RWTH Aachen Aachen, Germany ***************************************************************************** "Applications of Neural Networks to Virtual Reality" Professor Thomas P. Caudell Department of Electrical Engineering and Computer Engineering University of New Mexico, Albuquerque, New Mexico ***************************************************************************** "Computational Studies Of Biological Neural Networks: Introduction And Applications To Vision And Sensory-Motor Control" Paolo Gaudiano Department of Cognitive and Neural Systems Boston University, Boston, MA ***************************************************************************** "Hybrid Systems: Neural, Symbolic, and Fuzzy" Lawrence O. Hall and Abraham Kandel Computer Science and Engineering Department University of South Florida, Tampa, Fl. ***************************************************************************** "Practical Applications of Neural Network Theory" Dr. Robert Hecht-Nielsen HNC, Inc San Diego, CA ***************************************************************************** "Basics of Building Market Timing Systems: Making Money with Neural Networks" Casimir C. Klimasauskas NeuralWare, Inc. ***************************************************************************** "Learning Algorithms In Neural Networks" Professor Jacek M. Zurada Computer Science and Engineering University of Louisville, Louisville, KY For more information contact: IEEE ICNN IEEE World Congress on Computational Intelligence Meeting Management 2603 Main Street, Suite # 690 Irvine, California 92714 1-800-321-MEET FAX 714 752 7444 e-mail: 74710.2266@compuserve.com ------------------------------ End of Genetic Algorithms Digest ******************************