Systems and Means of Informatics

2025, Volume 35, Issue 3, pp 33-53

NEURAL NETWORKS BAYES SYNTHESIS OF MULTIDIMENSIONAL LINEAR STOCHASTIC SYSTEM

I. N. Sinitsyn
V. I. Sinitsyn
E. R. Korepanov
T. D. Konashenkova

Abstract

New method of optimal synthesis of multidimensional linear stochastic system on Bayes criterion (BC) based on quantitative estimate of output stochastic process (StP). Stochastic system is described by Pugachev equation for input and output StP. Input StP contains useful signal and random additive multidimensional normal noise with zero mathematical expectation and known matrix of covariance functions. Random noise does not depend upon vector of random parameters of useful signal. Distribution of random vector parameters is known. Model of BC optimal estimate of output StP is constructed on the basis of wavelet canonical expansion (CE) of random noise and wavelet CE of input StP. For finding unknown parameters in optimal output StP estimate, the architecture of multilayer wavelet neural networks (WNN) is developed. The WNN training algorithm for inverse error prevalence by method steepest descent is used. Formulae for mathematical expectation, second initial probabilistic moment, and error covariance matrix of BC optimal estimate of output StP is obtained. Numerical example illustrates CE WNN preference with wavelet CE.

[+] References (22)

Pugachev, V. S. 1962. Teoriya sluchaynykh funktsiy i ee primenenie k zadacham avtomaticheskogo upravleniya [Theory of random functions and its application to automatic control problems]. Moscow: Fizmatgiz. 884 p.
Bakulin, M.G., V.B. Kreyndelin, V.A. Grigoriev, V. O. Aksenov, and A. S. Schesnyak. 2020. Bayesian estimation with successive rejection and utilization of a priori knowledge. J. Commun. Technol. El. 65(3):255-264. doi: 10.1134/S1064226920030031. EDN: FUEROQ.
Yasinskii, S.A., A.N. Grigorchuk, V.N. Sykashew, and S.L. Dobrovolsky. 2022. Primenenie bayesovskogo printsipa i metoda maksimal'nogo pravdopodobiya dlya obnaruzheniya i razlicheniya signalov [Application of the Bayesian principle and the maximum likelihood method for detecting and distinguishing signals]. Trudy TsNIIS. Sankt-Peterburgskiy filial [Central Research Institute of Communications. St. Petersburg Branch] 1(13):106{112. EDN: CGPLYM.
Sinitsyn, I.N. 2023. Kanonicheskie predstavleniya sluchaynykh funktsiy. Teoriya i primeneniya [Canonical expansions of random functions. Theory and applications]. 2nd ed. Moscow: TORUS PRESS. 816 p.
Sinitsyn, I.N. 2023. Developing the theory of stochastic canonic expansions and its applications. Pattern Recognition Image Analysis 33(4):862{887. doi: 10.1134/S1054661823040429. EDN: QOWFBU.
Biryukov, I. D. 2024. Sovmestnyy kvazioptimal'nyy algoritm obrabotki radiosignalov istochnikov radioizlucheniya aviatsionnym sredstvom radiotekhnicheskogo nablyudeniya [Joint quasi-optimal algorithm for processing radio signals of radio emission sources by aviation radio surveillance facilities]. Vestnik RAEN [Bulletin of the Russian Academy of Natural Sciences] 24(3):48^8. doi: 10.52531/1682-1696-2024- 24-3-48-58. EDN: FVQSXJ.
Lutin, V. I., F. A. Desyatirikov, A. S. Belyaev, E. V. Lutina, and E. N. Desyatirikov. 2024. Sintez algoritma poiska zadannogo segmenta v posledovatel'nosti dannykh konechnoy dliny [Synthesis of an algorithm for searching a given segment in a data sequence of finite length]. Informatsionnye tekhnologii [Information Technologies] 30(7):335{341. doi: 10.17587/it.30.335-341. EDN: OFJDPY.
Sinitsyn, I.N., I.V. Sergeev, E.R. Korepanov, and T.D. Konashenkova. 2017. Instrumental'noe programmnoe obespechenie analiza i sinteza stokhasticheskikh sistem vysokoy dostupnosti (IV) [Software tools for analysis and synthesis of stochastic systems with high availability (IV)]. Sistemy vysokoy dostupnosti [Highly Available Systems] 13(3):55{69. EDN: ZSQHLR.
Sinitsyn, I.N., I.V. Sergeev, E.R. Korepanov, and T.D. Konashenkova. 2018. Instrumental'noe programmnoe obespechenie analiza i sinteza stokhasticheskikh sistem vysokoy dostupnosti (V) [Software tools for analysis and synthesis of stochastic systems with high availability (V)]. Sistemy vysokoy dostupnosti [Highly Available Systems] 14(1):59{70. EDN: YVXNBW.
Daubechies, I. 1992. Ten lectures on wavelets. Philadelphia, PA: Society for Industrial and Applied Mathematics. 352 p.
Sinitsyn, I.N., V.I. Sinitsyn, E.R. Korepanov, and T.D. Konashenkova. 2024. Modelirovanie nestatsionarnogo stokhasticheskogo protsessa posredstvom ego kanonicheskogo razlozheniya na osnove veyvlet-neyronnoy seti [Nonstationary stochastic process modeling by canonical expansion and wavelet neutral network]. Sistemy i Sredstva Informatiki - Systems and Means of Informatics 34(2):21{39. doi: 10.14357/08696527240202. EDN: YFHFIN.
Sinitsyn, I.N., V.I. Sinitsyn, E.R. Korepanov, and T.D. Konashenkova. 2025. Algoritm modelirovaniya vektornogo stokhasticheskogo protsessa posredstvom ego kanonicheskogo razlozheniya na osnove mnogosloynoy veyvlet-neyronnoy seti [Modeling algorithms for vector stochastic process by canonical expansions based on multilayer wavelet neural network]. Sistemy i Sredstva Informatiki - Systems and Means of Informatics 35(2):17{30. doi: 10.14357/08696527250202. EDN: TFDTXJ.
Sinitsyn, I.N., V.I. Sinitsyn, E.R. Korepanov, and T.D. Konashenkova. 2020. Optimization of linear stochastic systems based on canonical wavelet expansions. Automat. Rem. Contr. 81(11):2046-2061. doi: 10.1134/S0005117920110077. EDN: EOJMHD.
Sinitsyn, I., V. Sinitsyn, E. Korepanov, and T. Konashenkova. 2022. Bayes synthesis of linear nonstationary stochastic systems by wavelet canonical expansions. Mathematics 10(9): 1517. 14 p. doi: 10.3390/math10091517. EDN: GACWCC.
Sinitsyn, I., V. Sinitsyn, E. Korepanov, and T. Konashenkova. 2023. Synthesis of nonlinear nonstationary stochastic systems by wavelet canonical expansions. Mathematics 11 (9):2059. 18 p. doi: 10.3390/math11092059. EDN: LZQEDA.
Sinitsyn, I.N., V. I. Sinitsyn, E.R. Korepanov, and T.D. Konashenkova. 2023. Sintez optimal'nykh po funktsional'nym beyesovym kriteriyam stokhasticheskikh sistem metodom veyvlet kanonicheskikh razlozheniy [Methodological support of functional Bayes synthesis of stochastic systems based on wavelet canonical expansions]. Sistemy vysokoy dostupnosti [Highly Available Systems] 19(4):37-50. doi: 10.18127/j20729472-202304-03. EDN: WYQSQT.
Sinitsyn, I.N., V. I. Sinitsyn, E.R. Korepanov, and T.D. Konashenkova. 2024. Beyesov sintez mnogomernoy stokhasticheskoy sistemy vysokoy dostupnosti metodom veyvlet kanonicheskikh razlozheniy [Bayes synthesis of multidimensional stochastic system with high availability by wavelet canonical expansions]. Sistemy vysokoy dostupnosti [Highly Available Systems] 20(1):55-66. doi: 10.18127/j20729472-202401 -06. EDN: OMIDZW.
Sinitsyn, I.N., V.I. Sinitsyn, E.R. Korepanov, and T.D. Konashenkova. 2024. Neyrosetevoy sintez optimal'noy lineynoy stokhasticheskoy sistemy po kriteriyu minimuma srednekvadratichnoy oshibki [Neural network synthesis of an optimal linear stochastic system according to the criterion of minimum mean square error]. Sistemy i Sredstva Informatiki | Systems and Means of Informatics 34(3):87-108. doi: 10.14357/08696527240307. EDN: FVQWBN.
Sinitsyn, I.N., V.I. Sinitsyn, E.R. Korepanov, and T.D. Konashenkova. 2024. Neyrosetevoy algoritm sinteza optimal'noy lineynoy stokhasticheskoy sistemy vysokoy dostupnosti po energeticheskomu kriteriyu [Neural network algorithm for the synthesis of an optimal linear stochastic system with high availability according to the energy criterion]. Sistemy vysokoy dostupnosti [Highly Available Systems] 20(4):5-14. doi: 10.18127/j20729472-202404-01. EDN: PRWOPG.
Haykin, S. 1999. Neural networks. A comprehensive foundation. 2nded. Prentice-Hall of India Pvt. Ltd. 842 p.
Terekhov, S. A. 2001. Veyvlety ineyronnye seti [Wavelets and neural networks]. Nauch- naya sessiya MIFI-2001: III Vseros. nauchn.-tekhn. konf. "Neyroinformatika-2001": lektsii po neyroinformatike [Scientific session MEPhI-2001: 3rd All-Russian Scientific and Technical Conference "Neuroinformatics-2001:" Lectures on neuroinformatics]. Moscow: MIFI. 142-181.
Veitch, D. 2005. Wavelet neural networks and their application in the study of dynamical system. Networks 1(8):313-320.

[+] About this article