Begell House Inc.
Journal of Automation and Information Sciences
JAI(S)
1064-2315
28
5-6
1996
Numerical Aspects of Spectral Method of H∞-optimal Synthesis
1-12
Ya. M.
Bokova
E. I.
Veremei
Issues are considered related to the problem of mean-square synthesis in the absence of complete information about characteristics of disturbances. Numerical methods are proposed for synthesis of safeguarding controls in several specific situations, based on spectral theory of the mean-square synthesis, on one hand, and on the known methods of H∞-optimization, on the other hand.
Determination of the Parameters of a Step-Form Voltage of Multilevel Devices for Control of Asynchronous Motors
13-16
Vladimir I.
Gostev
State University of Information and Communication Technologies of Ministry of Transport and Communication of Ukraine, Kiev, Ukraine
Nikolay N.
Prisyazhnyuk
National Scientific Research Center of Defence Technologies and Military Security, Kyiv
S. V.
Yanchishin
A determination is made of a rational number n of steps and the number N of rectangular elements in a step-form voltage that appears on the winding of an asynchronous motor in devices for direct digital control of a motor. The choice of the numbers n and N is of great practical importance in the design of the force part of devices for direct digital control of a motor and is made from the results of solution of optimization problems by the Hooke-Jeeves method.
A Method of Analyzing Random Signals in State Diagnosis
17-22
B. I.
Adasovskiy
T. B.
Kirillova
Glushkov Institute of Cybernetics, Kiev
M. A.
Adasovskaya
Glushkov Institute of Cybernetics, Kiev
A precedent method for analyzing the correspondence between random signals of multidimensional dynamic objects and a desired solution of state diagnostic problems is proposed.
The Asymptotics of Bounded Controls in Optimal Nonlinear Elliptic Problems
23-35
Vladimir E.
Kapustyan
Dnepropetrovsk State University, Dnepropetrovsk Technical University of Railway Transport; National Technical University of Ukraine "Kiev Polytechnic Institute", Kiev
Irina V.
Shapoval
Institute for Problems of Material Science of National Academy of Sciences of Ukraine, Kiev
Issues in the construction and substantiation of the asymptotics of bounded controls in elliptic optimal nonlinear in the state systems are investigated. The results of a numerical experiment for a model problem that allow one to judge the effectiveness of the obtained asymptotics are presented.
Multivalued Mappings and Their Selectors in Game Control Problems
36-49
Alexey A.
Chikriy
V.M. Glushkov Institute of Cybernetics of National Academy of Sciences of Ukraine, Kyiv
A pursuit method in the class of counter controls with switching that is based on the use of calibrating functions and their inverses is proposed. Guaranteed times for different schemes are compared. It is shown that this method yields, in particular, a complete substantiation of the parallel-pursuit rule which is well known in practice.
Steering a Solid body in a Resistive Medium to a Terminal Set
50-59
Dmitriy V.
Lebedev
International Research and Training Center of Information Technologies and Systems of National Academy of Sciences of Ukraine and Ministry of Education and Science of Ukraine, Kyiv
An algorithmic solution of the problem of steering a solid body in a resistive medium into the ε−neighborhood of some point which is evolving in an inertial space is proposed. The control algorithms are synthesized by Lyapunov's second method while taking the constraints on the control parameters into account.
Computational Elements Built of Nondissipative Logic-Dynamic Systems
60-64
O. Yu.
Golovko
Glushkov Institute of Cybernetics, Kiev
Elements that realize standard logic operations are determined on the basis of a proposed Hamiltonian of a triple nonlinear interacting oscillators. An n-bit adder with carryover is constructed, and a sufficient condition for laying out the functional diagram of a discrete automaton in a plane is found.
Synthesizing Spaces for Restoring Membership Functions in Fuzzy Image Recognition Problems
65-71
V. I.
Vasil'ev
Glushkov Institute of Cybernetics of National Academy of Sciences of Ukraine, Kiev
V. I.
Sushko
Glushkov Institute of Cybernetics, Kiev
A method for specifying and restoring membership functions in fuzzily specified image (concept) recognition problems is proposed. The problem of restoring a membership function reduces a procedure of search for metric attributes subspace in which the order of the training sample objects, when ordered by their distance from the standard object of the given concept, coincides with the order of objects that is defined on the set of membership in this concept. The operational quality and reliability of the restored membership function on new data are guaranteed to be no worse than pre-specified values.
Discrete Analysis of the Frequency Characteristics of Random Signals
72-76
Inna D.
Ponomareva
International Research and Training Center of Information Technologies and Systems of National Academy of Sciences of Ukraine and Ministry of Education and Science of Ukraine, Kyiv, Ukaine
Gennadiy V.
Tsepkov
International Research and Training Center of Information Technologies and Systems of National Academy of Sciences of Ukraine and Ministry of Education and Science of Ukraine, Kyiv, Ukraine
The Use of Cluster Analysis for Partitioning Mixtures of Multidimensional Functional Characteristics of Complex Biomedical Systems
77-83
Evgen A.
Nastenko
Bogomoletz Institute of Physiology, NAS of Ukraine, Kyiv, Ukraine
A cluster-analysis method with constant cluster inertia is proposed. Its effectiveness for partitioning mixtures of functional characteristics of complex biomedical systems is demonstrated. The families of characteristics can be either trivial or nontrivial.
Use of a Contrast Criterion in a Sequential Procedure for Processing Spectral Power Density Estimates
84-92
B. M.
Koychu
The System Institute of the State Standards Bureau of Ukraine
V. N.
Sivers
The System Institute of the State Standards Bureau of Ukraine
Applying the Regularization Method to Estimation Problems
93-103
Vyacheslav F.
Gubarev
Institute of Space Research of National Academy of Sciences of Ukraine and State Space Agency of Ukraine, Kiev, Ukraine
Nikolay N.
Aksenov
Space Research Institute of National Academy of Sciences of Ukraine and National Space Agency of Ukraine
The problem of estimating the state variables of a dynamic object on the basis of an approach which uses the methods for solving ill-posed problems is examined and solved. The obtained results are illustrated by a computational experiment.
Some Optimization Problems in Ecology
104-110
A. I.
Yegorov
Dnepropetrovsk Institute of Railroad Engineering
Issues in determining the diffusion of coefficient of additives in the atmosphere far from the point source are analyzed. Optimal control theory methods for distributed-parameter systems are applied.
Synthesis of Robustly Optimal Multidimensional Control Systems
111-115
Vsevolod M.
Kuntsevich
Institute of Space Research of National Academy of Sciences of Ukraine and State Space Agency of Ukraine, Kiev, Ukraine
Alexey V.
Kuntsevich
V.M. Glushkov Institute of Cybernetics of National Academy of Sciences of Ukraine, Kyiv
The problem of synthesizing a system for controlling multidimensional objects (continuous and discrete) when set estimates of their parameters are available is posed as a problem of determining the control which provides the best proximity of the synthesized system motion to the motion of a specified (standard) system. In such a posing, this problem is formulated ultimately as a minimax problem, the solution of which reduces to the solution of a standard problem of minimizing a convex function under convex constraints.
Factorizing about the Imaginary Axis Matrices Which Do Not Satisfy the Coercitivity
116-130
Vladimir B.
Larin
S.P. Timoshenko Institute of Mechanics of National Academy of Sciences of Ukraine, Kyiv, Ukraine
The problem of factorization of a rational matrix about the imaginary axis is examined. Attention is concentrated on the case in which the matrix to be factored becomes singular at an infinitely distant point on the imaginary axis. The factorization algorithm is based on the construction of a stabilizing solution of a matrix algebraic Riccati equation. A generalized variant of the Anderson-Moore factorization identity is presented.
Optimal Evader Strategy in a Positional Group Pursuit Problem with Simple Players Motions
131-145
A. Ye.
Perekatov
Alexey A.
Chikriy
V.M. Glushkov Institute of Cybernetics of National Academy of Sciences of Ukraine, Kyiv
A positional group pursuit positional problem with a single evader and several pursuers who act independently and move along pursuit curves is considered. The optimal evader control is found in the class of constant controls which take into account group and individual pursuit zones. A methodology for constructing a control that yields a longer game duration than any of the constant controls is proposed on this basis. The results of a computational experiment are presented.
Simplified Linear-System Restorability and Controllability Criteria and Their Application in Robotics
146-151
Evgeniy M.
Potapenko
Zaporozhye National Technical University, Zaporozhye
Examining Bp Modification in Neural Arrays of Different Dimensionalities
152-158
Walid A.
Salameh
Dept. of Mathematical Sciences-Computer Science Branch JUST, P.O. Box 3030, Irbid-Jordan
The backpropagation neural network is the most popular network architecture. A better weight update for this learning rule would be possible if we could compensate for future changes to these weights in earlier layers. To do so, we address some modifications of the backpropagation learning algorithm that use the expected value of the source. These modifications are examined in neural arrays of different dimensionalities by means of computer simulation. A method of using these networks as pattern classifiers is proposed and simulated.
The Convergence of a Global Optimization Algorithm
159-165
E.
Senkiene
Institute of Mathematics and Information Sciences, Vilnius
The convergence of a numerical global optimization (annealing) algorithm when the function being optimized is discrete is examined.
Observability in Variable-Structure Systems
166-169
Fedor A.
Sopronyuk
Yuriy Fedkovich Chernovtsy National University
Ye. N.
Timofeyeva
Chernovtsy University
A theorem on the necessary and sufficient conditions for complete observability of linear dynamic variable-structure systems is proven. A recurrent filter and a recurrent matrix Riccati equation are constructed.