dtЦифровая трансформацияDigital Transformation2522-96132524-2822Educational Establishment “Belarusian State University of Informatics and Radioelectronics”10.35596/1729-7648-2024-30-3-80-88dt-874Research ArticleТЕХНИЧЕСКИЕ НАУКИTECHNICAL SCIENCESПрактический подход к изучению эволюционных методов настройки весовых коэффициентов искусственных нейронных сетейPractical Approach to Studying Evolutionary Methods for Setting Weight Coefficients of Artificial Neural NetworksПетровД. О.PetrovD. O.

Петров Дмитрий Олегович, канд. техн. наук, доц. каф. «ЭВМ и системы»

Тел.: +375 29 523-87-23

224017, г. Брест, ул. Московская, 267 

Dmitriy O. Petrov, Cand. of Sci., Associate Professor at the Department of Computer and Computer Sciences

Tel.: +375 29 523-87-23

224017, Brest, Moskovskaya St., 267 

polegdo@gmail.comБрестский государственный технический университетBrest State Technical University2024290920243038088Copyright © Петров Д.О., 20242024Петров Д.О.Petrov D.O.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://dt.bsuir.by/jour/article/view/874

Описана проблематика разработки нейроконтроллеров для управления динамическими объектами, включающая в себя сложность формирования обучающих наборов данных. Указано, что одним из известных способов обучения управляющей объектом искусственной нейронной сети является нейроэволюционный подход, предполагающий использование генетического алгоритма для настройки синаптических весовых коэффициентов искусственной нейронной сети. Предложена идея использования средства демонстрации эволюционного подхода к настройке весовых коэффициентов искусственной нейронной сети для практического обучения студентов основам нейроэволюционного подхода. Разработано программное обеспечение для демонстрации нейроэволюционного подхода на примере эволюции искусственной нейронной сети заданной структуры, предназначенной для управления упрощенной компьютерной моделью автономного транспортного средства. Описан способ разрешения проблемы стагнации при использовании эволюционного подхода к обучению искусственной нейронной сети. Предложены варианты применения разработанного программного обеспечения при обучении студентов основам технологий искусственного интеллекта и эволюционным методам многокритериальной оптимизации.

The article describes the problems of developing neurocontrollers for controlling dynamic objects, including the complexity of forming training data sets. It is indicated that one of the known methods for training an artificial neural network controlling an object is the neuroevolutionary approach, which involves using a genetic algorithm to adjust the synaptic weighting coefficients of an artificial neural network. The idea of using a means of demonstrating the evolutionary approach to adjusting the weighting coefficients of an artificial neural network for practical training of students in the basics of the neuroevolutionary approach is proposed. Software has been developed to demonstrate the neuroevolutionary approach using the example of the evolution of an artificial neural network of a given structure intended to control a simplified computer model of an autonomous vehicle. A method for resolving the problem of stagnation when using the evolutionary approach to training an artificial neural network is described. Options for using the developed software in teaching students the basics of artificial intelligence technologies and evolutionary methods of multicriteria optimization are proposed.

генетический алгоритмнейроэволюциянейроконтроллерискусственная нейронная сетьмногокритериальная оптимизациястагнация эволюцииgenetic algorithmneuroevolutionneurocontrollerartificial neural networkmulticriteria optimizationstagnation of evolutionReferencesЧернодуб, А. Н. Обзор методов нейроуправления / А. Н. Чернодуб, Д. А. Дзюба // Проблемы программирования. 2011. № 2. С. 79–94.Сhernodub A. N., Dzyuba D. A. (2011) Review of Neurocontrol Methods. Programming Problems. (2), 79–94 (in Russian).Katoch, S. A Review on Genetic Algorithm: Past, Present, and Future / S. Katoch, S. S. Chauhan, V. Kumar // Multimed Tools Appl. 2021. Vol. 80. P. 8091–8126.Katoch S., Chauhan S. S., Kumar V. (2021) A Review on Genetic Algorithm: Past, Present, and Future. Multimed Tools Appl. 80, 8091–8126.McCall, J. Genetic Algorithms for Modelling and Optimisations / J. McCall // Journal of Computational and Applied Mathematics. 2020. Vol. 184, No 1. P. 205–222.McCall J. (2020) Genetic Algorithms for Modelling and Optimisations. Journal of Computational and Applied Mathematics. 184 (1), 205–222.Васенков, Д. В. Методы обучения искусственных нейронных сетей / Д. В. Васенков // Компьютерные инструменты в образовании. 2007. № 1. С. 20–29.Vasenkov D. V. (2007) Methods for Training Artificial Neural Networks. Computer Tools in Education. (1), 20–29 (in Russian).Maren, A. J. A Logical Topology of Neural Networks / A. J. Maren // Proceedings of the Second Workshop on Neural Networks, Auburn, USA, February 11–13, 1991. Auburn: Auburn University, 1991. P. 17–44.Maren A. J. (1991) A Logical Topology of Neural Networks. Proceedings of the Second Workshop on Neural Networks, Auburn, USA, Febr. 11–13. Auburn, Auburn University. 17–44.Мищенко, В. А. Использование генетических алгоритмов в обучении нейронных сетей / В. А. Мищенко, А. А. Коробкин // Современные проблемы науки и образования. 2011. № 6.Mishchenko V. A., Korobkin A. A. (2011) Using Genetic Algorithms in Training of Neural Networks. Modern Problems of Science and Education. (6) (in Russian).Шумков, Е. А. Использование генетических алгоритмов для обучения нейронных сетей / Е. А.Шумков // Научный журнал КубГАУ. 2013. № 91.Shumkov E. A. (2013) Using Genetic Algorithms for Training of Neural Networks. Science Magazine of KubGAU. (91) (in Russian).Montana, D. J. Training Feedforward Neural Networks Using Genetic Algorithms / D. J. Montana, L. Davis // IJCAI’89: Proceedings of the 11th International Joint Conference on Artificial Intelligence. 1989. Vol. 1. P. 762–767.Montana D. J., Davis L. (1989) Training Feedforward Neural Networks Using Genetic Algorithms. IJCAI’89: Proceedings of the 11th International Joint Conference on Artificial Intelligence. 1, 762–767.Yao, Xin. Evolving Artificial Neural Networks / Xin Yao // Proceedings of the IEEE. 1999. Vol. 87, No 9. P. 1423–1447.Yao Xin (1999) Evolving Artificial Neural Networks. Proceedings of the IEEE. 87 (9), 1423–1447.Picek, S. On the Recombination Operator in the Real-Coded Genetic Algorithms / S. Picek, D. Jakobovic, M. Golub // 2013 IEEE Congress On Evolutionary Computation, June 20–23, Cancun, Mexico, 2013. P. 3103–3110.Picek S., Jakobovic D., Golub M. (2013) On the Recombination Operator in the Real-Coded Genetic Algorithms. IEEE Congress On Evolutionary Computation, June 20–23, Cancun, Mexico. 3103–3110.Zankinski, I. Effects of the Neuron Permutation Problem on Training Artificial Neural Networks with Genetic Algorithms / I. Zankinski // International Conference on Numerical Analysis and Its Applications, June 15–22, Lozenetz, Bulgaria, 2016. P. 777–782.Zankinski I. (2016) Effects of the Neuron Permutation Problem on Training Artificial Neural Networks with Genetic Algorithms. International Conference on Numerical Analysis and Its Applications, June 15–22, Lozenetz, Bulgaria. 777–782.Haflidason, S. On the Significance of the Permutation Problem in Neuroevolution / S. Haflidason, R. Neville // GECCO’09: Proceedings of the 11th Annual Conference on Genetic and Evolutionary Computation, June 8, 2009. P. 787–794.Haflidason S., Neville R. (2009) On the Significance of the Permutation Problem in Neuroevolution. GECCO’09: Proceedings of the 11th Annual Conference on Genetic and Evolutionary Computation, June 8. 787–794.Pretorius, K. Neural Network Crossover in Genetic Algorithms Using Genetic Programming / K. Pretorius, N. Pillay // Genetic Programming and Evolvable Machines. 2024. Vol. 25. No 7.Pretorius K., Pillay N. (2024) Neural Network Crossover in Genetic Algorithms Using Genetic Programming. Genetic Programming and Evolvable Machines. 25 (7).Gomez, F. J. Active Guidance for a Finless Rocket Using Neuroevolution / F. J. Gomez, R. Miikkulainen // Proceedings of the 2003 International Conference on Genetic and Evolutionary Computation: Part II, July 12, San Francisco, California, USA, 2003. P. 2084–2095.Gomez F.J., Miikkulainen R. (2003) Active Guidance for a Finless Rocket Using Neuroevolution. Proceedings of the 2003 International Conference on Genetic and Evolutionary Computation: Part II, July 12, San Francisco, California, USA. 2084–2095.Gomez, F. J. Co-Evolving Recurrent Neurons Learn Deep Memory POMDPs // F. J. Gomez, J. Schmidhuber // GECCO’05: Proceedings of the 7th Annual Conference on Genetic and Evolutionary Computation, June 25–29, Washington DC, USA, 2005. P. 491–498.Gomez F. J., Schmidhuber J. (2005) Co-Evolving Recurrent Neurons Learn Deep Memory POMDPs. GECCO’05: Proceedings of the 7th Annual Conference on Genetic and Evolutionary Computation, June 25–29, Washington DC, USA. 491–498.

The authors declare that there are no conflicts of interest present.