Search academic texts

Information × Registration Number 0214U002711, 0113U003453 , R & D reports Title The development of software for control and identification problems solving in large dimension systems on the base of supercomputers technology. popup.stage_title Head Kuntsevich Vsevolod Mykhailovich, Registration Date 30-05-2014 Organization Space Research Institute of National Academy of Sciences of Ukraine and State Space Agency of Ukraine popup.description2 Object of investigation: methods and algorithms of reachable (feasible) set estimating for linear and nonlinear dynamic systems, system identification methods by frequency characteristics under conditions of bounded uncertainty. Mission of investigation: determination of reachable sets for linear and nonlinear dynamic system classes, including systems with convex a priory estimations of nonlinear functions; formulation the problem of identification for multidimensional SISO systems under conditions of bounded uncertainty, development the method for it solving on the basis of the frequency characteristics, which are gained in active experiment. Methods of investigation: methods from Linear Algebra, Set theory, Convex Analysis, Optimization and Mathematical Programming, Computer Simulation, randomization. Results obtained: 1. An iterative method of finding ellipsoidal estimate for reachable set has been proposed for the wide nonlinear systems class. The upper bound invariant radius estimation for a reachable set has been founded. 2. The problem of identification for multidimensional SISO systems under conditions of bounded uncertainty has been defined. The requirements to solution has been formulated. Based on the frequency system identification method, more general approach to the process of mathematic modeling has been developed. The impact of computational errors on results of identification has been researched and limitation of applicability has been demonstrated. It has been revealed, that for model stability, the complexity (dimensionality) of model has to be concerted with the level of uncertainty. Via computer simulations with randomizations, a conceptual partition of space by model stability has been performed. The algorithm of structural identification, which guarantee solution stability, has been proposed. Product Description popup.authors Бабій Надія Анатоліівна Волосов Віктор Вікторович Губарев Вячеслав Федорович Компанієнко Лілія Григорівна Кравченко Аліна Вікторівна Кунцевич Всеволод Михайлович Мельничук Сергій Вікторович Царук Ніна Петрівна Шевченко Володимир Миколайович popup.nrat_date 2020-04-02 Close