Optimizing complex systems by intelligent evolotion: the LEMd method and case study

• Autorzy: Michalski R.S.
• Języki publikacji: EN

Abstrakty

EN

Most methods of evolutionary computation follow a Darwinian-type model that proceeds through random mutations or recombinations of the genetic material and natural selection of individuals carried out according to the principle of the survival of the fittest. In such a model, the creation of new individuals is not guided by any reasoning process or "external mind", but rather by random or semi-random changes. Recently, a new, non-Darwinian approach to evolutionary computation bas been proposed, called Learnable Evolution Model (LEM), in which the evolutionary process is guided by computational intelligence. In LEM, a new way of creating individuals is proposed, namely, by hypothesis formation and instantiation. In numerous experiments, LEM bas consistently and significantly outperformed compared conventional Darwinian-type algorithms in terms of the evolution length (the number of fitness evaluations) in solving complex function optimization problems. Based on the LEM ideas, we developed a method, called LEMd, which is tailored to problems of optimizing very complex engineering systems. This article provides a brief description of LEMd and its application to the development of a specialized system, ISHED, for the optimization of evaporator designs in cooling systems. According to experts in cooling systems, ISHED-developed designs have matched or outperformed the best human designs. These results and those from the experimental testing of learnable evolution on problems with hundreds of variables suggest that LEMd may be an attractive new tool for optimizing very complex engineering systems.

Słowa kluczowe

EN

engineering design optimization; evolutionary computation; function optimization; learnable evolution model; machine learning (ML); genetic algorithms

• Wydawca: Polska Akademia Nauk, Wydział IV Nauk Technicznych
• Czasopismo: Bulletin of the Polish Academy of Sciences. Technical Sciences
• Rocznik: 2006
• Tom: Vol. 54, nr 4
• Strony: 505--513
• Opis fizyczny: Bibliogr. 17 poz., rys.

Twórcy

autor

Michalski R.S.

• Machine Learning and Inference Laboratory, George Mason University, M.S.T.C8, Fair-fax, VA, 22030, USA Institute of Computer Science, Polish Academy of Sciences, 21 Ordona St., 01-237 Warszawa, Poland,

Bibliografia

• [1] M.W. Kirschner and J.C. Gerhart, The Plausibility of Life: Resolving Darwinin’s Dilemma, Yale University Press, 2005.
• [2] Z. Michalewicz, Genetic Algorithms + Data Structures = Evolution Programs, Springer-Verlag, London, 1996.
• [3] P. Bentley, “From coffee tables to hospitals: generic evolutionary design”, in Evolutionary Design by Computers, pp. 405–423, ed. P. Bentley, Menlo Park, CA: Morgan Kaufmann, 1999.
• [4] P. Bentley, Evolutionary Design by Computers, Menlo Park, CA: Morgan Kaufmann, 1999.
• [5] A. Oyama, “Multidisciplinary optimization of transonic wing design based on evolutionary algorithms coupled with CFD solver”, Eur. Congress on Computational Methods in Applied Sciences and Engineering, 2000.
• [6] P. Bentley and D. Corne, Creative Evolutionary Systems, Morgan Kaufmann, 2002.
• [7] R.S. Michalski, “Learnable evolution model: evolutionary processes guided by machine learning”, Machine Learning 38, 9–40 (2000).
• [8] J. Wojtusiak and R.S. Michalski, “The LEM3 implementation of learnable evolution model and its testing on complex function optimization problems”, Proc. Genetic and Evolutionary Computation Conference, GECCO 2006 PO6–7, (2006).
• [9] R.S. Michalski, and K.A.Kaufman, “The AQ19 system for machine learning and pattern discovery: a general description and user’s guide”, Reports of the Machine Learning and Inference 512 Bull. Pol. Ac.: Tech. 54(4) 2006 Optimizing complex systems by intelligent evolution: the LEMd method and case study Laboratory MLI 01-2, George Mason University, Fairfax, VA, 2001.
• [10] J. Wojtusiak, “AQ21 user’s guide,” Reports of the Machine Learning and Inference Laboratory MLI 04-3, George Mason University, Fairfax, VA, 2004.
• [11] R.S. Michalski, “Attributional calculus: a logic and representation language for natural induction”, Reports of the Machine Learning and Inference Laboratory MLI 04-2, George Mason University, Fairfax, VA, 2004.
• [12] P.A. Domanski, D. Yashar, K. Kaufman, and R.S. Michalski, “An optimized design of finned-tube evaporators using the learnable evolution model”, Int. Journal on Heating, Ventilating, Air-Conditioning and Refrigerating Research 10, 201–211 (2004).
• [13] P. Domanski, “EVSIM-an evaporator simulation model accounting for refrigerant and one dimensional air distribution”, NISTIR, 89–4133 (1989).
• [14] R.G. Reynolds, “An introduction to cultural algorithms, Proc. 3rd Annual Conf. on Evolutionary Programming, 131–139 (1994).
• [15] S. Saleem and R. Reynolds, “Cultural algorithms in dynamic environments”, Proc. Congress on Evolutionary Computation, 1513–1520 (2000).
• [16] K. Rasheed, “GADO: A genetic algorithm for continuous design optimization”, Technical Report DCS-TR-352, Department of Computer Science, Rutgers University, New Brunswick, NJ, 1998.
• [17] W.E. Hart, N. Krasnogor, and J.E. Smith, Recent Advances in Memetic Algorithms, Springer-Verlag, Berlin, 2005.