Search academic texts

Information × Registration Number 0223U004846, 0123U102943 , R & D reports Title Analytical methods and machine learning in control theory and decision-making in conditions of conflict and uncertainty popup.stage_title Head Norkin Volodymyr I., Доктор фізико-математичних наук Registration Date 06-12-2023 Organization V.M. Glushkov Institute of Cybernetics of the National Academy of Sciences of Ukraine popup.description2 Research object: Problems and methods of optimization of non-smooth non-convex functions under constraints, problems and methods of stochastic programming, Markov fields with local interaction, conflict-controlled dynamic processes. Purpose of the work: The purpose is to develop a set of mathematical methods for analysis, decision-making, and management of dynamic processes that function in conditions of conflict interaction and uncertainty. Research methods: methods of variational analysis, theories of random processes and fields, theory of differential games, optimization methods. Results and novelty. I. In the field of optimization theory: new methods of exact projective penalty functions, smoothing variants of adaptive stochastic gradient methods, and a new combined variant of the stochastic method of branches and boundaries have been developed. II. In the field of stochastic control theory: new robust methods of identification and learning with dependent observations have been developed; a new approach is proposed for the search for optimal control of stochastic processes and fields, which boils down to the search for analytical parametric controls and reduction of the problem to a finite-dimensional problem of stochastic programming; III. In the field of theory of decision-making in conflict situations: new models of conflict-controlled processes described by systems of differential equations, by non-stationary and functional differential equations, in particular with delay, with fractional derivatives of various types have been developed; new modifications of the resolving functions method have been developed. Product Description popup.authors Bila Halyna D. Bogdanov Oleksandr V. Bulovatski Maxym Sergiovych Guro Dmytro Anatolievych Dunaevskiy Maksim Serhiyovich Knopov Pavlo S. Kozyrev Anton Yurievych Norkin Volodymyr I. Samosyenok Olexandr S Chykrii Arkadii O. popup.nrat_date 2023-12-06 Close