In the article in order to determine the most effective global optimum multiextremal multivariate functions in the general case containing the points of discontinuity of the first and second kind, proposed modification of the genetic method. Numerical evaluation of the effectiveness of finding the global optimum of the proposed modification of the genetic method in comparison with a standard genetic algorithm and its known modifications made to a select group multiextremal test functions.
hybrid genetic algorithm, gradient learning, selection with prediction and parabolic interpolation, global optimization, multiextremal function.
1. Tenenev, V. A. Hybrid genetic algorithm with additional training of the leader / V. A. Tenenev, N. B. Paklin // Intellectual systems in production. - 2003. - No. 2. - pp. 181-206.
2. Dmitriyev, S. V. Direct optimization methods in a hybrid genetic algorithm / S. V. Dmitriyev, V. A. Tenenev // Intellectual systems in production. - 2005. - No. 2. - pp. 11-22.
3. Dmitriyev, S. V. Optimization of multiextreme functions by means of hybrid genetic algorithms / S. V. Dmitriyev, V. A. Tenenev // Proceedings of Institute of Mathematics and Informatics. Izhevsk. - 2006. - No. 2 (36). - pp. 163-166.
4. Vasilyev, F. P. Numerical methods of the extreme tasks decision / F. P. Vasilyev. - M.: Science, 1980. - 520 p.
5. Polyansky, I. S. Method for one-dimensional unconstrained optimization in problems of isolation partial beams of multibeam antenna mirror type / I. S. Polyansky // Modern problems of science and education. - 2012. - No. 4. - Access Mode: www.science-education.ru/104-6880.
6. Taha, Hemdi A. Introduction to operation research /Hemdi A. Taha. - M.: Williams, 2001. - 912 p.
7. Anderson, T. Statistical time-series analysis /T. Anderson; transl. from English I. G. Zhurbenko, V. P. Nosko, under the editorship of Yu. K. Belyaev. - M.: World, 1976. - 755 p.
8. Herrera, F. Tackling real-coded genetic algorithms: operators and tools for the behaviour analysis / F. Herrera, M. Lozano, J. L. Verdegay // Artificial Intelligence Review. - 1998. - Vol. 12. - № 4. - P. 265-319.
9. Eremenko, V. T. Methodological aspects of the synthesis of optimal tree in data acquisition and processing / V. T. Eremenko, I. S. Polyansky, I. I. Besedin // Bulletin of computer and information technologies. - 2013. - No. 11. - pp. 11-15.
10. Polyansky, I. S. Algoritm for distribution of heterogeneous discrete limited resources in the physical security system / I. S. Polyansky, I. I. Besedin // Information systems and technologies. - 2013. - No. 4(78).
