Abstract
Constructive approaches, principles, and methods of optimization of identification in conditions of a priori insufficiency, parametric uncertainty, nonstationarity, low accuracy of data processing have been developed based on the use of natural information redundancy of random time series (RTS) and nonstationary objects in control systems of industrial and technological complexes. A technique has been developed for the accelerated computation of statistical parameters with insufficient statistics. The mechanisms of threshold control of values of elements of RTS, control of increments of the sequence of elements of RTS, inaccuracies of predictions for various models of identification of RTS are proposed. Received general and particular solutions of information and optimization problems are obtained based on a wide spectrum of statistical and dynamic prediction models, taking into account the conditions of a non-stationary process. Algorithms for adaptation of the level of location, compression or expansion of boundaries, combined adaptation of variables to the dynamics of RTS have been developed. A software package identification and forecasting of non-stationary objects has been developed. Defined conditions, the sequence of execution of program modules. Algorithmic synthesis of mechanisms for identifying RTS with a mechanism was carried out in a parallel computing environment on the NVIDIA CUDA platform a modified cyclic multigrid method. Implemented on C ++ modules in the mode with a four-core AMD Athlon 64X2 4800+ processor have been performed.
First Page
104
Last Page
112
References
1. D.Tikk, D.Baranyi, T.D.Gedeon, L.Muresan, Generalization of the rule interpolation method resulting always in acceptable conclusion.Tatra Mountains Math. Publ., 2001, 280 p.
2. A.Steinwolf, R.A.Ibrahim, “Numerical and experimental studies of linear systems subjected to non-Gaussian random excitations”, Probabilistic Engineering Mechanics, vol. 14, no. 4, pp. 289-299, 1999.
3. G.Mzyk, Combined Parametric-Nonparametric Identification of Block-Oriented Systems. Springer, 2014, 238 p.
4. Y.Boutalis, D.Theodoridis, T.Kottas, M.A.Christodoulou, System Identification and Adaptive Control. Theory and Applications of the Neurofuzzy and Fuzzy Cognitive Network Models. Springer, 2014, 313 p.
5. S.Fabrit, V.Kadirkamanathant, “Dual Adaptive Control of Nonlinear Stochastic Systems using Neural Networks”, Automatica, vol. 34, no. 2, pp. 245-253, 1998.
6. O.I.Djumanov, S.M.Kholmonov, “Methods and algorithms of selection the informative attributes in systems of adaptive data processing for analysis and forecasting”, Applied Technologies and Innovations”, Prague Development Center. – Prague, vol. 8, pp. 45-55, 2012.
7. I.I.Jumanov, A.R.Akhatov, “Fuzzy semantic hypernet for information authenticity controlling in electronic document circulation systems”, 4-th International Conference on Application of Information and Communication Technologies, Tashkent, 12-14 october, 2010, 978-1-4244-6904-8/10/$26.00 ©2010 IEEE.
8. S.G.Fabri, M.K.Bugeja, “Functional adaptive dual control of a class of nonlinear MIMO systems”, Transactions of the Institute of Measurements and Control, vol. 37, pp. 1009-1025, 2015.
9. S.Fabrit, V.Kadirkamanathant, “Dual Adaptive Control of Nonlinear Stochastic Systems using Neural Networks. Automatica, (1998), 34, p. 245-253.
10. Yu.N.Orlov, S.L.Fedorov, “Generasiya nestasionarnix trayektoriy vremennogo ryada na osnove uravneniya Fokkera-Planka” [Generation of non-stationary time series tractors based on the Fokker-Planck equation], Trudi MFTI, vol. 8, no. 2, pp. 126-133, 2016. (in Russian).
11. V.N.Afanasyev, M.M.Yuzbashev, Analiz vremennix ryadov i prognozirovaniye [Time-series analysis and forecasting]. Moskva: Finansi i statistika, 2001, 228 p. (in Russian).
12. I.I.Jumanov, S.M.Xolmonov, “Porogoviy kontrol tochnosti neprerivnoy informasii v avtomatizirovannix sistemax upravleniya texnologicheskimi prosessami” [Threshold control of continuous information accuracy in automated textual process control systems], V mejdunarodnoy nauchno-prakticheskoy konferensii «Nauka i obrazovaniye v sovremennom mire: vizovi XXI veka» Tezisi dokladov, NUR-SULTAN, 2019, pp. 308-313. (in Russian).
13. I.I.Jumanov, Z.T.Bekmurodov, “Identifikasiya sluchaynix vremennix ryadov na osnove neyro-nechetkoy seti dlya povisheniya dostovernosti prognoza” [Identification of random time series based on a neuro-fuzzy network to increase the reliability of the forecast], Trudi XI Mejdunarodnoy Aziatskoy shkoli-seminara «Problemi optimizasii slojnix sistem», Almati, 2015, pp. 258-264. (in Russian).
14. I.I.Jumanov, S.M.Xolmonov, N.M.Kayumova, “Povisheniye dostovernosti obrabotki dannix nestasionarnix obyektov na osnove primeneniya multikontekstnix rekurrentnix neyronnix setey” [Improving the reliability of data processing of non-stationary objects based on the use of multi-context recurrent neural networks], Nauka i mir, vol. 1, no. 3(55), pp. 52-54, 2018. (in Russian).
15. A.B.Barskiy, Neyronniye seti: raspoznavaniye, upravleniye, prinyatiye resheniy [Neural networks: recognition, management, decision-making]. Moskva: Finansi i statistika, 2004, p. 176. (in Russian).
16. N.Yusupbekov, S.Gulyamov, S.Kasymov, N.Usmanova, D.Mirzaev, “Software implementation of exchange processes in a distributed network environment of transmission and processing of information”, Journal of Automation, Mobile Robotics and Intelligent Systems, vol. 12(4), pp. 64-69 DOI: 10.14313/JAMRIS_4-2018/27.
Recommended Citation
Jumanov, Isroil Ibragimovich and Kholmonov, Sunatillo Maxmudovich
(2020)
"OPTIMIZATION OF IDENTIFICATION UNDER THE CONDITIONS OF LOW RELIABILITY OF INFORMATION AND PARAMETRIC UNCERTAINTY OF NON-STATIONARY OBJECTS,"
Chemical Technology, Control and Management: Vol. 2020:
Iss.
5, Article 19.
DOI: https://doi.org/10.34920/2020.5-6.104-111
