Begell House Inc.
Journal of Automation and Information Sciences
JAI(S)
1064-2315
48
2
2016
Sixtieth Anniversary of the Journal
1
Estimate of Stationary Probabilities of States of the GI / G / ∞ Queueing System by Importance Sampling
2-10
Nikolay Yu.
Kuznetsov
V.M. Glushkov Institute of Cybernetics of National Academy of Sciences of Ukraine, Kiev
Igor N.
Kuznetsov
V.M. Glushkov Institute of Cybernetics of
National Academy of Sciences of Ukraine,
Institute of Physics and Technology of
National Technical University of Ukraine
"Kiev Polytechnic Institute", Kiev
The GI / G / ∞ queueing system in heavy traffic is investigated. The fast simulation method enabling one to construct asymptotically unbiased estimates of stationary probabilities is proposed. Numerical examples are
considered.
On the Question of Finding the Value of Routing Problem with Constraints
11-27
Alexander G.
Chentsov
N.N. Krasovsky Institute of Mathematics
and Mechanics of Ural Branch of Russian
Academy of Sciences, Ural Federal
University, Ekaterinburg
Alexey A.
Chentsov
N.N. Krasovsky Institute of Mathematics
and Mechanics of Ural Branch of Russian
Academy of Sciences, Ekaterinburg
The problem of sequential going around megapolises with the constraints of different types is considered. It is supposed that cost functions and "current" constraints can depend on the tasks list (the dependence on the
list of fulfilled and vice versa yet nonfulfilled tasks is possible). An approach to determination of a global extremum (a problem value) on the basis of the widely interpreted dynamic programming is proposed. Owing to this
approach, the storage-efficiency of computer is reached; this permits one to determine the extremum in the problem of the larger dimension and use it for heuristic algorithms testing. To construct the layers of the Bellman function, the shortened procedure, which enables one to decrease the computing complexity, is used (under the previous conditions, the construction of all the
array of the Bellman function values is not expected).
Cloud Technologies: New Facilities of Computing Cryptanalysis
28-35
Valeriy K.
Zadiraka
V.M. Glushkov Institute of Cybernetics of National Academy of Sciences of Ukraine, Kiev
Anton M.
Kudin
Institute of Physics and Technology of
National Technical University of Ukraine
"Kiev Polytechnic Institute", Kiev
Polina V.
Seliukh
Institute of Physics and Technology of
National Technical University of Ukraine
"Kiev Polytechnic Institute", Kiev
Inna V.
Shvidchenko
V.M. Glushkov Institute of Cybernetics of National Academy of Sciences of Ukraine, Kiev
Cloud computing as a new model of access to computing resources modifies classical problems of cryptography and cryptanalysis. Of special interest is an important evaluation parameter for cloud computing − the cost
of computing services offered in solving cryptanalysis problems of symmetric and asymmetric cryptosystems. The paper deals with cloud computing cost evaluation for the stages of directed enumeration of some common methods of cryptanalysis of symmetric and asymmetric cryptosystems. The perspectives of using cloud computing for cryptanalysis problems solution are considered.
Robust Control of Some Classes of Nonlinear Discrete-Time Objects Using Linear Controllers
36-53
Leonid S.
Zhiteckii
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
The problems of robust discrete-time control for two classes of nonlinear uncertain objects in the presence of arbitrary unmeasurable bounded disturbances with possibly unknown bounds and linear feedback are stated and solved. Namely, the problem of designing a linear nonadaptive robust controller for controlling a class of nonlinear multivariable static objects and the problem of designing a linear adaptive robust controller for controlling a class of nonlinear one-dimensional dynamic objects are
considered. Sufficient conditions guaranteeing the robust stability of the control systems and also some asymptotic properties of designed control algorithms are established.
Analysis of Parallel Algorithm of Empirical Models Synthesis on Principles of Genetic Algorithms
54-73
Mikhail I.
Gorbiychuk
Ivano-Frankovsk National Technical
University of Oil and Gas, Ivano-Frankovsk
Vera M.
Medvedchuk
Ivano-Frankovsk National Technical
University of Oil and Gas, Ivano-Frankovsk
Alla N.
Lazoriv
Ivano-Frankovsk National Technical
University of Oil and Gas, Ivano-Frankovsk
The developed method of synthesis of empirical models using the genetic algorithms significantly reduces the computing time for implementing empirical models comparing with the inductive method of self-organizing models. To improve the effectiveness of computing process the parallel algorithm was developed and analyzed that allowed us to estimate the reduction of computing time compared with the sequential or FIFO algorithm.
In-depth Classification Method on the Basis of the Mahalanobis Scalable Distance for Sets with a Priori Unequal Probabilities
74-82
Alexander A.
Galkin
Kiev National Taras Shevchenko University, Kiev
Nonparametric in-depth classification method was suggested in the case when data sets have a priori unequal probabilities and do not belong to common family of elliptic distributions. Universal in-depth classifier, which does not depend on displacement in model of location shift or violation of monotonic character of density function was developed. The Mahalanobis scalable distance is evaluated at every point with usage of the residual passage method.