Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Research › peer-review
Parameter Estimation Problems in Markov Random Processes. / Karelin, Vladimir; Fominyh, Alexander; Myshkov, Stanislav; Polyakova, Lyudmila.
Computational Science and Its Applications- ICCSA 2019 - 19th International Conference, Proceedings. ed. / Beniamino Murgante; Osvaldo Gervasi; Elena Stankova; Vladimir Korkhov; Sanjay Misra; Carmelo Torre; Eufemia Tarantino; David Taniar; Ana Maria A.C. Rocha; Bernady O. Apduhan. Springer Nature, 2019. p. 691-701 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11622 LNCS).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Research › peer-review
}
TY - GEN
T1 - Parameter Estimation Problems in Markov Random Processes
AU - Karelin, Vladimir
AU - Fominyh, Alexander
AU - Myshkov, Stanislav
AU - Polyakova, Lyudmila
N1 - Conference code: 19
PY - 2019/1/1
Y1 - 2019/1/1
N2 - Problems of convergence and stability of Bayesian estimates in the identification of stochastic control systems are considered. The informational measure of the mismatch between the estimated distribution and the estimate is the main apparatus for establishing the fact of convergence. The choice of a priori distribution of parameters is not always obvious. The Kullback-Leibler information number is taken as such measure. The convergence of the estimates of the transition function of the process to the non-stationary transition function is established in this paper. The problem of synthesis of optimal strategies for dynamic systems in which there is no part of the main information needed for constructing the optimal control is also considered. It is assumed that the system contains at least one unknown parameter belonging to some parameter space. Therefore, the class of control systems considered in the article is the class of parametric adaptive systems.
AB - Problems of convergence and stability of Bayesian estimates in the identification of stochastic control systems are considered. The informational measure of the mismatch between the estimated distribution and the estimate is the main apparatus for establishing the fact of convergence. The choice of a priori distribution of parameters is not always obvious. The Kullback-Leibler information number is taken as such measure. The convergence of the estimates of the transition function of the process to the non-stationary transition function is established in this paper. The problem of synthesis of optimal strategies for dynamic systems in which there is no part of the main information needed for constructing the optimal control is also considered. It is assumed that the system contains at least one unknown parameter belonging to some parameter space. Therefore, the class of control systems considered in the article is the class of parametric adaptive systems.
KW - Bayesian probability theory
KW - Kullback-Leibler information number
KW - Parameter estimation
UR - http://www.scopus.com/inward/record.url?scp=85068591717&partnerID=8YFLogxK
UR - http://www.mendeley.com/research/parameter-estimation-problems-markov-random-processes
U2 - 10.1007/978-3-030-24305-0_51
DO - 10.1007/978-3-030-24305-0_51
M3 - Conference contribution
AN - SCOPUS:85068591717
SN - 9783030243043
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 691
EP - 701
BT - Computational Science and Its Applications- ICCSA 2019 - 19th International Conference, Proceedings
A2 - Murgante, Beniamino
A2 - Gervasi, Osvaldo
A2 - Stankova, Elena
A2 - Korkhov, Vladimir
A2 - Misra, Sanjay
A2 - Torre, Carmelo
A2 - Tarantino, Eufemia
A2 - Taniar, David
A2 - Rocha, Ana Maria A.C.
A2 - Apduhan, Bernady O.
PB - Springer Nature
T2 - 19th International Conference on Computational Science and Its Applications, ICCSA 2019
Y2 - 1 July 2019 through 4 July 2019
ER -
ID: 43689589