Begell House Inc.
Journal of Automation and Information Sciences
JAI(S)
1064-2315
37
9
2005
On Optimal Control of Singular Mixed Systems
1-11
10.1615/J Automat Inf Scien.v37.i9.10
Victor N.
Mizernyi
Vinnitsa National Technical University, Ukraine
Louisa
Toscano
University of Naples "Federico II", Italy
Analysis of optimal control problems for objects, whose mathematical model is described by equations of different types (with partial derivatives, ordinary differential equations, integral equations etc.) Conditions of solvability and properties of solutions of extremal problems are studied.
Guaranteed Estimation of the Phase State and Parameters of Linear Dynamic Systems
12-18
10.1615/J Automat Inf Scien.v37.i9.20
Vsevolod M.
Kuntsevich
Institute of Space Research of National Academy of Sciences of Ukraine and State Space Agency of Ukraine, Kiev, Ukraine
We give solution of a guaranteed estimation problem of the state vectors and parameters of linear controlled objects, their mathematical models being given in the form of vector-matrix difference equations. We show that while the certain condition holds true, there exists
a principal possibility to improve the method of determining estimates of the vector state and parameters, proposed in [1, 2]. This method is based on the scheme of “inverse recalculation”. The proposed method is generalized for a class of nonstationary systems with the bounded rate of parameters changing.
On Further Development of Pseudoinverse Approach to Identification of Kernels of Matrix Functional Transformers
19-30
10.1615/J Automat Inf Scien.v37.i9.30
Vasiliy B.
Zvaridchuk
Kiev National Taras Shevchenko University, Kiev, Ukraine
Vladimir Antonovich
Stoyan
Kyiv National Taras Shevchenko University, Kyiv, Ukraine
Problems of identification of matrix functions, transforming vector static input into vector dynamic output, are solved. The study is grounded on a linear functional model which, however, allows deviations from linearity. The issue of nonidentifibility of the model due to linear dependence of input measurements is studied in detail.
Identification for Systems of Stochastic Dynamic Discrete Models with Determinate Coefficients
31-46
10.1615/J Automat Inf Scien.v37.i9.40
Alexander P.
Sarychev
Institute of Technical Mechanics of National Academy of Sciences of Ukraine and National Space Agency of Ukraine, Dnepropetrovsk, Ukraine
We consider the problem of estimation of coefficients in the system of stochastic dynamic discrete models, for which variations of states are defined by different sets of preceding states and different variables of state. Quality functional for the system of stochastic dynamic discrete models was introduced. We obtained conditions of optimality, suggested iterative scheme for determination of coefficients, which minimize the introduced functional.
Stochastic Evaluation of Aircraft Collision Risk under Cooperative Air Traffic Control
47-54
10.1615/J Automat Inf Scien.v37.i9.50
Vladimir P.
Kharchenko
National Aviation University, 1 Kosmonavta Komarova Ave., Kyiv, 03058,
Ukraine
Alexander G.
Kukush
National Aviation University, Kiev, Ukraine
Vladimir N.
Vasylyev
National Aviation University, Kiev, Ukraine
We obtained equations for estimate of probability of potential conflict and collision of aircrafts with taking into account stochastic character and correlation dependence on time of deviations from preset trajectories of flight under cooperative air traffic control. Here stabilization of the given parameters of trajectories is taken into account, which is provided by onboard system of flight control. We determine coefficients of equations, which are required for numerical solving of the problem.
Estimation by Information Criteria
55-62
10.1615/J Automat Inf Scien.v37.i9.60
Oleg L.
Levoshych
Kiev National Taras Shevchenko University, Kiev, Ukraine
For constructing criteria of estimation of random vectors the mean mutual information between an estimate and the value, being estimated, is used. The theorem about reduction of
a stochastic maximin estimation problem to the corresponding determinate problem is proved.
Vector Multiconnected Markov Chains
63-69
10.1615/J Automat Inf Scien.v37.i9.70
Vladimir S.
Mukha
Byelorussian State University of Informatics and Radioelectronics, Minsk, Belorussia
A multidimensional matrix mathematical model of a vector multiconnected Markov chain is developed. An example of analysis of such chain is studied.