S.V. Chernomordov1, O.V. Druzhinina2, O.N. Masina3, A.A. Petrov4

1, 3, 4 Bunin Yelets State University (Yelets, Russia)

2 FRС «Computer Science and Control» of RAS (Moscow, Russia)

The development of instrumental and methodological support for modeling controlled technical systems (CTS) is an actual problem. Promising methods for solving this problem include methods of neural network modeling and machine learning. The purpose of this paper is to develop an approach to construct models of intelligent control of technical systems and to the application of machine learning methods in the problems of modeling for controlled technical systems. The implementations and concretizations of the algorithms of the Gym modular library of the high-level Python language for solving problems of neural network modeling and reinforcement machine learning are considered. Methods for training neural networks in modeling technical systems are characterized. The possibility of using heterogeneous computing to improve the performance of computing systems is analyzed. A machine learning algorithm is developed for the implementation of an unmanned aerial vehicle (UAV) model. An approach to creating a combined reinforcement machine learning algorithm and optimization is developed. The practical significance of the results lies in the fact that the proposed algorithmic support can be used in the development of neurocomputer systems. The obtained results can be used in the problems of CTS modeling. In addition, the results can be used in various problems of neural network modeling and machine learning.

Chernomordov S.V., Druzhinina O.V., Masina O.N., Petrov A.A. Application of machine learning methods in problems of neural network modeling of controlled technical systems. Neurocomputers. 2022. V. 24. № 1. Р. 25-35. DOI: https://doi.org/10.18127/j19998554-202201-03 (in Russian)

1. Siniak N.G., Marina B.A. A review in big data and its role in real estate industry decision making. Real Estate: Economics, Management. 2020. № 2. P. C 22–28.
2. Druzhinina O.V., Masina O.N., Petrov A.A. Up-to-date software and methodological support for studying models of controlled dynamic systems using artificial intelligence. Lecture Notes in Networks and Systems (LNNS). Springer. 2021. V. 1225. P. 470–483.
3. Aggarwal C. Neural Networks and Deep Learning. Springer International Publishing. 2019.
4. Geron A. Hands-on Machine Learning with Scikit-Learn, Keras, and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems. OReilly. 2019.
5. Starostin N.V., Pankratova M.A. Mnogourovnevye algoritmy dekompozicii grafa dannyh dlja parallel'nyh vychislenij na geterogennoj vychislitel'noj sisteme. Voprosy atomnoj nauki i tehniki. Ser. Matematicheskoe modelirovanie fizicheskih processov. 2016. № 1. S. 60–68 (in Russian).
6. Hajkin S. Nejronnye seti: polnyj kurs. M.: Vil'jams. 2006 (in Russian).
7. Voroncov K. V. Mashinnoe obuchenie (kurs lekcij). [Jelektronnyj resurs]. URL=http://www.machinelearning.ru/ wiki/index.php?title=Mashinnoe_obuchenie_(kurs_lekcij,_K.V.Voroncov) (data obrashhenija: 24.12.2021) (in Russian).
8. Dulhare U.N., Bin Ahmad K.A., Ahmad K. Machine Learning and Big Data: Concepts, Algorithms, Tools and Applications. Wiley. 2020.
9. Voroncov K.V. Matematicheskie metody obuchenija po precedentam (teorija obuchenija mashin) [Jelektronnyj resurs]. URL=http://www.machinelearning.ru/wiki/images/6/6d/voron-ml-1.pdf (data obrashhenija: 24.12.2021) (in Russian).
10. Druzhinina O.V., Korepanov Je.R., Belousov V.V., Masina O.N., Petrov A.A. Opyt razrabotki metodov i sredstv nejrosetevogo modelirovanija nelinejnyh sistem na baze otechestvennoj vychislitel'noj platformy «Jel'brus 801-PC». Nelinejnyj mir. 2020. T. 18. № 2. S. 5–17 (in Russian).
11. Kilin G.A., Zhdanovskij E.O. Preimushhestva ispol'zovanija obuchenija s podkrepleniem dlja obuchenija nejronnyh setej. Materialy Vseross. nauch.-tehnich. konf. «Avtomatizirovannye sistemy upravlenija i informacionnye tehnologii» (g. Perm', 17 maja 2018 g.). V 2-h tomah. Perm': PNIPU. 2018. S. 152–158 (in Russian).
12. Radford A., Metz L., Chintala S. Unsupervised representation learning with deep convolutional generative adversarial networks // ArXiv: 1511. 06434v2 [cs. LG] 7 Jan 2016.
13. Ajrapetov A.Je., Kovalenko A.A. Issledovanie generativno-sostjazatel'noj seti. Politehnicheskij molodezhnyj zhurnal. 2018. № 10(27).
    S. 1–7 (in Russian).
14. Masina O.N., Petrov A.A., Druzhinina O.V., Rapoport L.B. Modelirovanie upravljaemyh sistem s primeneniem metodov stabilizacii i algoritmov poiska optimal'nyh traektorij: Ucheb. posobie. Elec: Eleckij gosud. un-t im. I.A. Bunina. 2021 (in Russian).
15. Antonov A.S., Afanas'ev I.V., Voevodin V.V. Vysokoproizvoditel'nye vychislitel'nye platformy: tekushhij status i tendencii razvitija. Vychislitel'nye metody i programmirovanie. 2021. T. 22. № 2. S. 135–177 (in Russian).
16. Bizli D. Python. Podrobnyj spravochnik. SPb: Simvol-Pljus. 2010 (in Russian).
17. Kurbatov T.G., Nekrasova V.Je. Primenenie mashinnogo obuchenija s podkrepleniem dlja reshenija zadach platformy Openai Gym. Sb. materialov VII Vseross. nauch.-tehnich. konf. «Studencheskaja nauka dlja razvitija informacionnogo obshhestva» (g. Stavropol', 26–28 dekabrja 2017 g.). Stavropol': Severo-Kavkazskij federal'nyj un-t. 2018. S. 182–184 (in Russian).
18. Control theory problems from the classic RL literature. [Jelektronnyj resurs]. URL=https://gym.openai.com/envs/#clas-sic_control (data obrashhenija: 20.12.2021).