Population in an Evolutionary Trap: Simulation of Natural Exploration

• Autorzy: Obuchowicz A.
• Języki publikacji: EN

Abstrakty

EN

The aim of the presented paper is to introduce and compare three novel mechanisms: Simple Variance Adaptation (SVA), Forced Direction of Mutation (FDM) and Deterioration of the Objective Function (DOF), which accelerate the global optimization ability of evolutionary algorithms. The evolutionary algorithm considered here, called Evolutionary Search with Soft Selection (ESSS), is based on the simplest selection-mutation model of phenotype evolution [5]. The comparison analysis is made using two kinds of simulation experiments. The first one is focused on the saddle crossing ability of the algorithm considered. In the second one, chosen global optimization problems are solved.

Słowa kluczowe

EN

global optimization; evolutionary search; variance adaptation; forced direction of mutation; fitness erosion

• Wydawca: Polska Akademia Nauk, Wydział IV Nauk Technicznych
• Czasopismo: Bulletin of the Polish Academy of Sciences. Technical Sciences
• Rocznik: 2003
• Tom: Vol. 51, nr 1
• Strony: 60--104
• Opis fizyczny: 15 rys., 4 tabele, bibliogr. 19 poz.

Twórcy

autor

Obuchowicz A.

• Institute of Control and Computation Engineering, University of Zielona Góra, ul. Podgórna 50, 65-246 Zielona Góra, (Instytut Sterowania i Systemów Informatycznych Uniwersytetu Zielonogórskiego)

Bibliografia

• [1] T. Bäck, H.-P, Schwefel, An overview of evolutionary computation, Evol. Comput. 1, 1, (1993) 1-23.
• [2] A. Chorążyczewski, R. Galar, I. Karcz-Dulęba, Considering phenotypic evolution in the space of population states, Proc. 5th Int. Conf, Neural Networks and Soft Computing, Technical University Press, Częstochowa (2000) 615-620.
• [3] L. J. Fogel, A. J. Owens, M. J. Walsh, Artificial intelligence through simulated evolution, Wiley, New York 1966.
• [4] D. B. Fogel, An introduction to simulated evolutionary computation, IEEE Trans. Neural Networks, 5 (1994) 3-14.
• [5] R. Galar, Handicapped individua in evolutionary processes, Biological Cybernetics, 51 (1985) 1-9.
• [6] R. Galar, Evolutionary search with soft selection, Biological Cybernetics, 60 (1989) 357-364.
• [7] R. Galar, I, Karcz-Dulęba, The evolution of two: an example of space of states approach, Proc. 3rd Annual Conf. Evolutionary Programming, San Diego CA, World Scientific (1994) 261-268.
• [8] D. E, Goldberg, Genetic algorithms in search, optimization, and machine learning, Addison-Wesley Publishing Company, Inc., Reading, MA, 1989.
• [9] J, H. Holland, Adaptation in natural and artificial systems, The University of Michigan Press, Ann Arbor, MI, 1975.
• [10] I. Karcz-Dulęba, Simulating evolutionary processes as a tool of global optimization in Rn (in Polish), Ph.D. thesis, Technical University Press, Wrocław, 1992.
• [11] I. Karcz-Dulęba, Some convergence aspects of evolutionary search with soft selection method, Proc. 2nd Conf. Evolutionary Algorithms and Global optimization, Warsaw University of Technology Press (1997) 113-120.
• [12] Z. Michalewicz, Genetic algorithms + data structures = evolution programs, Springer-Verlag, Berlin Heidelberg 1996.
• [13] A. Obuchowicz, The true nature of multi-dimensional Gaussian mutation, Int, Conf. Artificial Neural Networks and Genetic Algorithms — ICANNGA’2000, Prague, Czech Republic, to be published.
• [14] A. Obuchowicz, J. Korbicz, The evolutionary search with soft selection and deterioration of the objective function, Proc. 6th Int. Symp. Intelligent Information Systems — IIS’97, Warsaw: ICS PAS Press (1997) 228-295.
• [15] A. Obuchowicz, J. Korbicz, Evolutionary search with soft selection and forced direction of mutation, Proc. 7th Int. Symp. Intelligent Information Systems — IIS’98, Warsaw: ICS PAS Press (1998) 300-309.
• [16] A. Obuchowicz, K. Patan, About some modifications of evolutionary search with soft selection algorithm, Proc. 2nd Conf. Evolutionary Algorithms and Global optimization, Warsaw University of Technology Press (1997) 193-200.
• [17] A. Obuchowicz, K. Patan, An algorithm of evolutionary search with soft selection for training multilayer feedforward neural networks, Proc. 3rd Conf. Neural Networks and Their Applications, Technical University Press, Częstochowa (1997) 123-128.
• [18] I. Rechenberg, Cybernetic solution path of an experimental problem, Roy. Aircr. Establ., libr. Transl. 1122, Farnborough, Hants., UK, 1965.
• [19] X. Yao, Y. Liu, Evolutionary programming made faster, IEEE Trans. Evolutionary Computation, 3, 2, (1999).