ABSTRACTS OF ARTICLES OF THE JOURNAL "INFORMATION TECHNOLOGIES".
No. 4. Vol. 22. 2016

To the contents

A. D. Tarasov, Senior Lecturer of "Automated systems of information processing and control" department, e-mail: adtarasov@mail.ru, State agrarian university of Orenburg

Genetic Algorithm Effectiveness Used in Problem of Physical Defense System Designing

Is considered the problem of decision-making for potentially-dangerous objects physical defense system structure. The problem stage with experts knowledge using — the conceptual designing, which including object defense analysis, practical recommendations for physical defense system creation, options of engineering defense units structure. It is observed problem solving process of engineering defense units structure designing with Standard Genetic Algorithm using. Method based on object of defense representation in graph form. The example of engineering defense units structure searching with the program is shown. Algorithm parameters with the greatest influence on decision process were found. Experiments are described which show: these parameters must be determined with Adaptive Genetic Algorithm using.
Keywords: physical protection system, decision support system, standard genetic algorithm, adaptive genetic algorithm.

P. 243—249

A. I. Nikolaev, Research Intern, ainikolaev@hse.ru, Laboratory of Algorithms and Technologies for Networks Analysis, National Research University Higher School of Economics, Nizhny Novgorod

Efficient Approach for the Maximum Clique Problem Based on Machine Learning

In this paper a new approach for solving the maximum clique problem is presented. For a given graph the suggested approach uses machine learning technique to predict the fastest algorithm from several algorithms for the maximum clique problem. Then the chosen algorithm is applied for solving the maximum clique problem in this graph. The computational results show the efficiency of the proposed approach.
Keywords: maximum clique problem, exact algorithms, decision trees

P. 249—254

A. V. Dmitriev, Prof. Dr., S. V. Maltseva, Professor, O. A. Tsukanova, Graduate Student, National Research University "Higher School of Economics otsukanova@hse.ru, a.dmitriev@hse.ru, smaltseva@hse.ru

Modeling and Qualitative Analysis of a Social Microblogging Network as a Dynamical System

An approach to the investigation into the dynamics of social networks is described. The results of a qualitative analysis of the social network based on a microblogging principle are represented. The network interaction is approximated as a dynamic system characterizing the number of senders and recipients of small content elements.
The aim of the present paper is to make an investigation into the social network as a dynamical system, considering informational saturation and informational interaction between the senders and recipients. Systems without and with restrictions are investigated. It was determined that a social network, represented by a dynamic system considering saturation and interaction between the senders and recipients, is adequate. The parametric conditions of an asymptotic stability and forked bifurcation for the network dynamics are developed.
Keywords: mathematical modeling, social networks, microblogging, dynamical systems, non-linear dynamics, asymptotic stability, forked bifurcation

P. 255—260

V. I. Diagilev, Associate Professor, dvi-39@mail.ru, V. A. Kokovin, Associate Professor, kokovin@uni-protvino.ru, The Public Institution of Higher Education of Moscow Region the Dubna State University, branch "Protvino", Protvino, Russia, S. U. Uvaysov, Professor, National Research University The Higher School of Economics, s.uvaysov@hse.ru, Moscow, Russia, S. S. Uvaysova, Graduate Student, National Research University The Higher School of Economics, uvay@yandex.ru, Moscow, Russia

Computer Simulation of the Power Converter with Harmonic wave Output

This paper presents mathematical model of power converter (PC) circuit with output high-frequency sinusoidal voltage for in-feeding the piezoelectric transducers. These transducers can be used in various technological installations. To develop the model we took into account two aspects: the real schema attributes of power converter and peculiar properties of the oscillation circuit in a resonant mode and at load.
The real powerful power switches were chosen, named field MOSFET — transistors, rather than virtual model. Moreover, we also took into account the values of parasitic resistances and inductances of the power source, the output transformer and inductor. The implementation of second aspect gives an opportunity to assess the sine wave quality of PC output voltage, as well as get formulas for calculating the parameters of reactive power elements — inductor and a capacitor oscillation circuit.
The research of the proposed model showed that there are certain difficulties in calculating the PC equivalent circuit. The difficulties are because of the nonlinearity of the processes occurring in PC at variable load. The paper shows two approaches to solve these problems with the use of computer modeling. The example of PC scheme calculating for in-feeding the powerful piezoelectric transducers in resonant mode is also shown.
The research was based on proposed model. That allowed formulate recommendations for selection the converter circuit parameters with LC filters.
Keywords: computer modelling, the power converter, mathematical model, calculation procedure

P. 261—266

P. P. Oleynik, PhD, System Architect Software, Aston JSC, Associate Professor, Shakhty Institute (branch) of Platov South Russian State Polytechnic University (NPI), Russia, Rostov-on-Don, xsl@list.ru, V. I. Gurianov, PhD in Technical Science, Assistant Professor, Saint-Petersburg State Economic University, Branch in Cheboksary, Russia, Cheboksary, vg2007sns@rambler.ru

UML-Profile for Metamodel-Driven Design of Database Applications

Currently, most of the newly developed applications are database applications. At the same time for the functionality of the system most commonly used object-oriented programming languages that allow you to create re-usable pieces of code, which is achieved with the use of encapsulation, inheritance and polymorphism. Further development of the object-oriented approach is the extension of its principles for the development of the entire application. As a result, the developer provides a single development environment that allows you to create a target application. At the same time the developer tasked correct design and construction of domain models in terms of the selected tool.
This article presents a UML-profile that repeatedly used by the authors in the development of database applications in their own development environment. This profile allows you to metemodelno-oriented application design and describes the domain model in terms of the development of object metamodel. At the end of the article provides examples of the use of the profile and draw conclusions on ways of working.
Keywords: UML, UML Profile, Databases, Object-Oriented Design

P. 267—276

P. N. Bibilo, Professor, Head of Lab., I. P. Loginova, Associate Professor, Senior Researcher United Institute of Informatics Problem, National Academy of Sciences of Belarus, Minsk

The Creation of Energy-Intensive Tests for Combinational Logic Circuits According to the Results of Evaluation of Their Power Consumption

The decrease of energy consumption of logic circuits is a complex problem and it requires the decision in various aspects. One of these aspects is evaluation of power consumption in the mode of heightened consumption of energy. The design of logic circuits is performed using program-synthesizer that converts high level VHDL descriptions of algorithms for functioning circuits into structural description of logic circuits. The approximate evaluations of power consumption in logic CMOS circuits are implemented through modeling of such descriptions in logical simulation systems, for example in CAD ModelSim (Mentor Graphics). The more exact evaluation of power consumption in logic circuit projects are made by analog simulation systems. In both approaches, for every pair of input variable values, the designer gets the value of energy consumed by the circuit in the current time unit. This paper gives formalization of the problem of finding energy-intensive tests for combinational logic CMOS circuits. Also, it offers two algorithms for searching a given number of test vectors, using of which in logical re-simulation of such circuits provides the mode of their heightened consumption. Selection of test vectors is done by results of primary logic circuit simulation. Results of primary logic circuit simulation (both logical and analog) are represented by sequence of binary sets, each pair of neighboring sets related to the time activation being put in correspondence to power consumption of the circuit (i.e. the number equal power consumption in the time of activation is shown). The first algorithm operates on the set of all possible test vector pairs and intended for small-scale circuits. The second algorithm is fast and efficient, deals with pseudorandom sequences of input vectors and can be applied for arbitrary-dimension circuits. At the same time, when choosing the most energy-intensive pairs it cannot predict energy consumption of unused set pairs (in primary simulation), however it gives often enough up to 20—30 % increase of power average consumption even without re-simulation. The results of experiment on building of energy-intensive tests for combinational logic circuits from library CMOS gates are presented. Experiments shows that selection of average 10 % energy-intensive test sets from the whole sets of test vector pairs allows 50—70 % increasing power consumption of circuits in found energy-intensive tests by remodeling.
Keywords: combinational CMOS circuits, simulation, pseudorandom test vectors, estimating energy consumption

P. 277—283

S. M. Avdoshin, PhD, professor, Head of Software Engineering School, Faculty of Computer Science, HSE, savdoshin@hse.ru, A. V. Lazarenko, undergraduate student, Software Engineering School, Faculty of Computer Science, HSE, avlazarenko@edu.hse.ru, National Research University Higher School of Economics (HSE)

Technology of Anonymous Networks

This paper is an overview of currently used anonymous networks based on technology of onion routing and peer-to-peer networking. It describes key features of the networks and their comparative characteristics. The main purpose of every anonymous network is to protect information from the adversaries and provide users with great level of anonymity. All networks can be clustered on two classes: onion routing and its modifications and plain-old peer-to-peer networks. In the first class the major participant is Tor, which is based on the second generation of onion routing. On the other hand, P2P networks can be divided on 2 classes: traditional peer-to-peer and friend-to-friend. Friend-to-friend is a type of routing where users connects only to those users, who are considered as friends. The first class of peer-to-peer networks contains: Tarzan, MorphMix, Freenet, I2P, Netsukuku. The second class is represented by such networks as: Turtle, RetroShare. Current paper is focused only on those networks, which were successful on practice, or have strong impact on anonymous systems. Nowadays users have a wide specter of different solutions which can be used for protecting anonymity on the Internet. Anonymous networks differ by architectures, routing types and target audiences. Unfortunately, there is no any solution, which guarantees 100 % defense from adversaries. Every technology has its own weaknesses and vulnerabilities, allowing attacker to somehow deanonimize a particular user.
Keywords: anonymous networks, onion routing, peer-to-peer, layered encryption, invisible internet, overlay networks

P. 284—291

L. A. Lyutikova, Head of Departament, e-mail: lylarisa@yandex.ru, E. V. Shmatova, Junior Researcher, Institute of Applied Mathematics and Automation, Nalchik

Recognition Algorithms Analysis and Synthesis with Varied Values Logic

In this paper we carried out logical analysis of the given specified domain that presents an object the features of which describe it using varied values logic and a set of algorithms running within the given domain is studied. We developed logical procedure for constructing correct algorithms, to research the specified domain, simulate knowledge base for specific objects that minimize it and allocate a unique set of characteristics for each object.
Keywords: algorithm, training set, a set of data, knowledge base, the subject area, logic of varied values, clauses, the decision rule

P. 292—297

V. Ya. Tsvetkov, Deputy Director of the Center for Basic and Advanced Studies, e-mail: cvj2@mail.ru, NIIAS

Polysemy Information

This article describes the polysemy of the term information as the objective situation. This situation is due to the development of the term and its various applications in different areas. This article describes the "dogma of the one-dimensionality" as a typical cause of erroneous findings and conclusions. The article describes the evolution of the term "information" from the time of Aristotle to the modern days. The article describes the state of information theory and computer science. These sciences are not fully developed at the moment, this situation also contribute polisemii information.
Keywords: information, information science, information theory, computer science, polysemy, dogma-dimensionality

P. 298—302

A. I. Galushkin, Prof., Deputy Head of Chair of the Moscow Institute of Physics and Technology, Dolgoprudny, neurocomputer@yandex.ru, Prof. of High School of Economics, agalushkin@hse.ru, D. V. Pantiukhin, Senior Lecturer of High School of Economics, dpantiukhin@hse.ru, Lecturer of Chair of the Moscow Institute of Physics and Technology, Dolgoprudny, dim_beavis@mail.ru

Supercomputers and Memristors

The article describes the authors' opinion about the perspectives of development of one area of supercomputers — neurocomputers — with the advent of memristors. These perspectives are determined by ultra-high parallelism in connection with the transition to the neural network logical basis, the sharp decline in energy consumption in connection with the transition to the presentation of information in the form of a frequency sequence of narrow pulses, the transition to an analog implementation of the algorithms with a sharp increase in speed. Examples of such development of neurochips and neurosystems from recent years are described.
Due to developments in neural network theory, neuromathematics and neurocontrol, micro- and nanoelectronics and parallel hardware, it is possible today to change VonNeuman computing paradigm to neural network one. Neurocomputers with memristor will be important step to exascale computing.
This article show historical and today progress in neurocomputing area in a world and some projects of neurocomputers with memristors, including DARPA Neural network Study Projects, IBM TrueNorth neurochip and many approaches to memristor-based neural networks.
Authors postulate that major applications of such memristor-based neurocomputers will be:
— information security (intrusion, viruses and spam detection),
— (neuro-) cryptography tasks,
— mathematical physics tasks including adaptive computing mesh generation,
— image and 3D objects processing (detection, classification, clustering and oth.),
— brain modelling,
— and others applications.
It is also discussed about neurocomputing problem in Russia.
Authors hope that this article will stimulate discussion about such difficult problem as neurocomputers with memristor within Russia and abroad.
Keywords: neurocomputer, neurochip, memristor, exaflop computing

P. 304—312

A. I. Zadayanchuk¹, Student, e-mail: alex.goncharov@phystech.edu, M. S. Popova¹, Student, V. V. Strijov², Leading Scientist
¹Moscow Institute of Physics and Technology
²Dorodnicyn Computing Center, Russian Academy of Sciences

Selection of Optimal Time Series Classification Model

This paper solves the problem of selecting optimal stable models for classification of physical activity. We select optimal models from the class of two-layer artificial neural networks. There are three different ways to change structure of neurons: network pruning, network growing, and their combination. We construct models by removing its neurons. Neural networks with insufficient or excess number of neurons have insufficient generalization ability and can make unstable predictions. Proposed genetic algorithm optimizes the neural network structure. The novelty of the work lies in the fact that the probability of removing neurons is determined by the variance of parameters. In the computing experiment, models are generated by optimization two quality criteria — accuracy and stability.
Keywords: classification, artificial neural networks, accuracy, genetic algorithm, pruning criteria

P. 313—318

To the contents