|
Vol: 54(68) No: 4 / December 2009 Multiobjective Genetic Programming with Insular Evolution and Adaptive Migration Lavinia Ferariu Department of Automatic Control and Applied Informatics , “Gh. Asachi” Technical University of Iasi, Bd. D. Mangeron 53A, 700050, Iasi, Romania, phone: (004 0232) 278680, e-mail: lferaru@ac.tuiasi.ro Alina Patelli Department of Automatic Control and Applied Informatics , “Gh. Asachi” Technical University of Iasi, Bd. D. Mangeron 53A, 700050, Iasi, Romania, e-mail: apatelli@ac.tuiasi.ro Keywords: genetic programming, multiobjective optimization, migration, dominance analysis, nonlinear system identification. Abstract The paper introduces a genetic programming based approach, devoted to nonlinear system identification. The proposed method makes use of the nonlinear, linear in parameters formalism, for providing an efficient concomitant selection of model structure and parameters. In this context, QR decomposition is employed for a faster parameters’ computation and enhanced genetic operators are implemented, in order to ensure increased exploration capabilities. The tree-based individuals are evolved via a multiobjective optimization, which addresses to both accuracy and complexity. The authors present an original strategy, able to implement a dynamic symbiosis between the evolutionary search procedure and the decision mechanism. To enforce the selection of compact and accurate models, the priorities of the conflicting objectives are adapted by means of a migration scheme with multiple rates, targeted at efficiently exploiting the implemented fitness assignment mechanism. The performances of the design procedure are highlighted on two industrial identification problems. References [1] P. J. Flemming, R. C. Purshouse “Evolutionary Algorithms in Control Systems Engineering: A Survey”, Control Engineering Practice 10, 2002, pp. 1223-1241. [2] T. Back, D. Fogel, Z. Michalewicz, Evolutionary Computation-Advanced Algorithms and Operators, US: Institute of Physics Publishing, 2000. [3] D. Fogel, Evolutionary computation – Toward a New Philosophy of Machine Intelligence, 3rd Ed., IEEEPress, John Wiley and Sons, 2006. [4] J. R. Koza, Genetic Programming: On the Programming of Computers by Means of Natural Selection, Cambridge, MA, MIT Press, 1992, pp. 73-190. [5] J. R. Koza, F. H. Bennet III, D. Andre, M. A. Keane, “Genetic Programming: Turig’s Third Way to Achieve Machine Intelligence” in Evolutionary Algorithms in Engineering and Computer Science, Miettinen, Kaissa, Ed. Chichester, UK: Wiley & Sons, 1999, pp. 185-189. [6] H. Wey, S. A. Billings, J. Lui, “Term and Variable Selection for Nonlinear Models”, Int. J. Control 77, 2004, pp. 86-110. [7] K. Rodriguez-Vasquez, C. M. Fonseca, P. J. Flemming, “Identifying the Structure of Nonlinear Dynamic Systems Using Multiobjective Genetic Programming” in IEEE Transactions on Systems Man and Cybernetics, Part A – Systems and Humans, 34, 2004, pp. 531-534. [8] K. Rodriguez-Vasquez, P. J. Flemming, “A Genetic Programming/NARMAX Approach to Nonlinear System Identification” in Second International Conferenceon Genetic Algorithms in Engineering Systems: Innovations and Applications, GALESIA, 1997, pp. 409-414. [9] L. Ferariu, M. Voicu, “Nonlinear System Identification Based on Evolutionary Dynamic Neural Networks wih Hybrid Structure” in Proc. of IFAC Congress, Prague, 2005. [10] Deb, K.. Multi - Objective Optimisation using Evolutionary Algorithms, Wiley, USA, 2001. [11] J. Knowles, D. Corne, K. Deb, Multiobjective problem Solver from Nature, Springer Verlag, Netherlands, 2008. [12] O. Smart, H. Firpi, G. Vachtsevanos, “Genetic Programming of Conventional Features to Detect Seizure Precursors”, Engineering Applications of Artificial Intelligence 20, 2007, pp. 1070–1085. [13] A. V. A. Kumar, P. Balasubramaniam, “Optimal Control for Linear Singular Systems Using Genetic Programming” in Applied Mathematics and Computation 192, 2007, pp.78-89. [14] T. Marcu, L. Mirea, L. Ferariu, P. M. Frank, “Miscellaneous Neural Networks Applied to Fault Detection and Isolation of an Evaporation Station”, Proc. Of 4th IFAC Symposium on Fault Detection, Supervision and Safety for Technical Processes, SAFEPROCESS, Hungary, 2000. |