Begell House Inc.
Journal of Automation and Information Sciences
JAI(S)
1064-2315
40
6
2008
Presentation of Solutions of Linear Systems with Fractional Derivatives in the Sense of Riemann-Liouville, Caputo and Miller-Ross
1-11
Ivan I.
Matichin
V. M. Glushkov Institute of Cybernetics of National Academy of Sciences of Ukraine, Kyiv, Ukraine
Arkadiy A.
Chikriy
V.M. Glushkov Institute of Cybernetics of National Academy of Sciences of Ukraine, Kiev, Ukraine
Consideration is given to inhomogeneous linear systems of differential equations with classical Riemann-Liouville fractional derivatives as well as the regularized Caputo derivative and sequential Miller-Ross fractional derivatives. Using Laplace transform the solutions of such systems are presented in terms of matrix Mittag-Leffler functions as analogues of Cauchy formula for arbitrary Lebesque measurable and bounded functions of time in the right-hand side. These relations play a key role in solving related problems of mathematical control theory and theory of dynamic games.
To the Question of Stegosystems Stability for Detection of Fact of Hidden Message Communication for Two Special Cases
12-16
Valeriy K.
Zadiraka
V.M. Glushkov Institute of Cybernetics of National Academy of Sciences of Ukraine, Kiev
Lyubov L.
Nikitenko
V.M. Glushkov Institute of Cybernetics of National Academy of Sciences of Ukraine, Kiev, Ukraine
We consider two special cases of construction of stable stegosystems. The perfect stegosystem was constructed for theoretical information stegosystem model. For probabilistic model of stegosystem we consider the special case, when attacker does not know used containers.
Application of Group Method of Data Handling with a Shift Operator to Forecasting Financial Indices of Local Budgets
17-26
Valentin N.
Tomashevskiy
National Technical University of Ukraine "KPI", Kiev, Ukraine
Alexander N.
Vinogradov
National Technical University of Ukraine "KPI", Kiev, Ukraine
Yuriy A.
Oleinik
Research Institute of Systems Technologies of National Technical University of Ukraine "KPI", Kiev, Ukraine
Application of GMDH algorithm with a shift operator to forecasting financial indices of local budgets is offered. GMDH with the shift operator was updated by adding in the algorithm the criterion of variables balance. The examples of GMDH algorithm with the shift operator applied to forecasting a budget revenues of Volyn local budget are given. Comparison of results of GMDH operation with the available balance criterion and without it is made.
Solving the Problem of Discriminant Analysis under the Condition of Structural Uncertainty on the Basis of Group Method of Data Handling
27-40
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 search the optimal by complexity discriminant function. The criteria of quality for discriminant functions are described, which are developed within the framework of the group method of data handling: criterion based on partition of observations into initial and testing samplings, and criterion of sliding test. The problems of this class belong to pattern recognition problems under the condition of structural uncertainty, which were considered by A.G. Ivakhnenko as long ago as 60−70-th of the last century as urgent problems of engineering cybernetics.
Artificial Immune Systems as Means for Inductive Construction of Optimum Models for Complex Objects
41-59
Vladimir I.
Lytvynenko
Kherson National Technical University, Ukraine
We described mechanisms of negative and clonal selection of biological systems, as well as basic notions of theory of immune network. On the basis of immunologic metaphors we gave formal representation of algorithms of clonal selection and immune network. Starting from the idea of soft calculations relative to integration of different paradigms for amplification of desirable computational properties of algorithms we adduced results of integration of external criterion, which is used in group method of data handling, ways of coding decisions and mutation operators, which are used in genes expression programming. We suggested the developed hybrid algorithms of clonal selection and immune network, which are intended for solving problems of inductive modeling.
Genetic Algorithm for Solving the Problem of an Optimum Regression Model Construction as a Discrete Optimization Problem
60-71
Ivan M.
Melnik
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, Kiev, Ukraine
Consideration is given to the construction problem of optimum regressive model of a complex system characterized by m input (independent) variables and one output (dependent) variable of stochastic character. The problem consists in selecting from the set of independent variables such a subset which optimizes a given functional of model quality. The methods for solving this problem of discrete optimization as the search problem of the shortest path on a special graph are suggested. The main attention is focused on application of genetic algorithm ideas to heuristic search for optimum in this problem.
Fast Background Modeling Algorithm Based on Stability Intervals
72-79
Anton V.
Prokofiev
V.M. Glushkov Institute of Cybernetics of National Academy of Sciences of Ukraine, Kiev, Ukraine
A fast algorithm of per pixel background modeling for unimodal scenes is proposed. Ideas of recent values average algorithm and mixture of Gaussian densities algorithm were improved by the procedure of stability checking of the recent values of pixels enabling one to reduce the calculation amount and increase the performance.