ABSTRACTS OF ARTICLES OF THE JOURNAL "INFORMATION TECHNOLOGIES" N. 1, 2014

To the contents

M. A. Baranov

Composite Key Terms in Task of Clustering Text Documents

In the first time the problem of text document clustering using phrases consisting of more than two words is considered in terms of vector space model. It is shown that best results are achieved when the clustering process combined with a single word phrases consisting of two or three words is used.
Keywords: clustering text documents, greedy algorithm, composite key terms

Valeeva A. F., Goncharova Yu. A., Koshcheev I. S.

The Development of the Logistic Transport System

In this article the prototype of the logistic transport system for solving the problem of goods delivery to customers (3L-EVRP, Three-Dimensional Extended Vehicle Routing Problem) is considered. Mathematical model for 3L-EVRP problem and the algorithms of the optimization kernel of system such as a population based ant colony optimization algorithm for looking for rational routes and evolutionary algorithm (1+1)-EA3D for looking for rational packing of contents with cuboid and cylindrical forms, taking into account required restriction (the consideration of the center of gravity, palletization etc) are presented. Some numerical experiments on random generated data, taking into account such restrictions as vehicle capacity, time windows, split delivery, delivery period, multi depot and so on, and tested examples, taken from International OR-library tests, are resulted.
Keywords: routing, goods displacement in transportation means, transportation logistics, logistic transport system, population based ant colony optimization algorithm, evolution algorithm

P. 9—14

A. N. Chetyrbotsky

Numerical Modeling of the Earth's Mantle Viscosity Methods Radial Basis Functions in MATLAB

We consider the form of the analytic representation of characteristics of internal friction (viscosity) in the Earth's mantle. The issues of harmonization of statistical samples that are used to carry out in MATLAB parametric identification of the model of mantle viscosity. To construct the model of the machine is used radial basis functions. The methods of the graphical representation of the results of computational experiments.
Keywords: Earth's mantle, the viscosity of the mantle, statistical sampling, parametric identification

G. G. Bulychev

Numerical Simulation of Dynamics and Dynamical Destruction of Arch and Bridge Structures

In this article the project’s architecture, mathematical model, method and results of the numerical simulation of dynamics and dynamic destruction of arches and bridges is considered. For simulation the numerical spatial characteristics method was developed and numerically realised by the author as applied to the given class of problems, and an author's machine graphics was used for it. The nature of destruction of structures and magnitudes of its limiting breaking loads was determined, structure's optimisation on a dynamic strength was conducted.
Keywords: a numerical simulation, a mechanics of the deformed solid body, dynamics, fracture, frame structures

P. 19—24

S. V. Luchkova, T. O. Peremitina, I. G. Vashchenko

Application of Software Multidimensional Data Analysis Based on Fuzzy and Statistical Modeling

In this paper we consider an integrated approach to the analysis of multidimensional data with missing values. This approach includes data recovered by fuzzy system based on evolutionary strategy and identifying dependencies of object properties by factor analysis with "varimax" rotation. The model of data recovery and algorithm of factor analysis are described. Experimental researches results are presented.
Keywords: fuzzy system, data recovered, factor analysis, integrated approach, physic-chemical properties of oil, microbiological properties of produced water

P. 24—30

M. V. Tarasuk

Choice of the VPN Topology Taking into Account Requirements of the Confidentiality of the Telecommunication Loading Characteristics

In this paper the model of the amount information estimation in the packets time delays, transfer throw untrusted telecommunication cannel, is proposed. The principles of the construction virtual channels topology for the special crypto security network on the base of the VPN technology are define for considering security requirements of the telecommunication loading characteristics such network.
Keywords: covert channel, virtual channel, virtual private network, crypto-router, traffic masquerading

G. Ch. Habibakova

A Fuzzification Method of Attributes of Data Warehouse of Decision Support System in the Foreign Policy

The paper presents a fuzzification process of attributes of data warehouse of decision support system in the foreign policy. To this end, the chosen algorithm CLARANS was applied for clustering of data warehouse. The values of membership function were derived with medoids, which were obtained after clustering. The paper deals with the case, where the number of term-sets generated on different dimensions are various.
Keywords: OLAP-cube, fuzzification, fuzzy-set, linguistic variable, term-set, clustering, data warehouse, attribute, medoid

N. T. Fisun, G. V. Gorbanj

Models and Methods for Building OLAP System for Object-Oriented Databases

Aspects of building OLAP systems in an object-oriented DBMS are discussed in the article. A scheme of relationship of facts with dimensions in the object-oriented data model and metaclass diagram for OLAP implementation using metadata are proposed.
Keywords: database OLAP, relational database, object-oriented database, fact table, dimension, measure, multidimensional array, class, instance of the class, property, reference, relationship, metadata, notation IDEF1

N. A. Solomeshch, I. V. Brailovskiy

Analysis of Video Quality of Wavelet Video Codec for Unreliable Communicational Channel

This paper examines dependency of video quality from the level of interference in communicational channel in case of wavelet based video codec. To analyze this dependency authors conducted computational experiments based on open source wavelet-based video codec Schrodinger. The results or this research allows adapting parameters of the wavelet video codec for minimizing visual quality degradations for streaming over unreliable communicational channel.
Keywords: video coding, wavelet transform, communicational channels with interference, bit rate, rate distortion function, image and video quality, SSIM, robustness

P. 46—51

A. V. Savchenko

Speech Segmentation in the Problems of Automatic Voice Control Detection and Recognition

The problem of automatic detection and recognition of the voice commands in the continuous speech in control application are investigated. The phoneme segmentation algorithm on the nasis of the probabilistic neural network with homogeneity testing is applied to extract a vowel from the syllable. The speech signal is proposed to be marked as a part of the voice control command if the ratio of durations of the homogeneous segment and the whole syllable exceeds a fixed threshold. The experimental research shows that the proposed approach extracts the control command containing more than two syllables in 90 % cases if the commands are pronounced in isolated syllable mode.
Keywords: automatic speech recognition, speech segmentation, probabilistic neural network with homogeneity testing, voice control system, words phonetic coding method

P. 53—57

E. D. Aved'yan, V. E. Lugansky

An Approach to the Problem of Filling Gaps in Tables and Lines Based on the Multilayer and CMAC Neural Networks

In this paper, is claimed that the problem of detection and filling gaps arises in many practical applications. A brief review of recent works on this problem is given. These works describe the possible applications and a wide range of approaches to their solution. From this review, it follows that the approach based on the multilayer neural network have great potential in solving this problem. We show that this problem is especially effective solved using neural network CMAC. We give a brief description of this neural network and an algorithm for solving the problem of detection and recovery of missing data. The simulation results show that the neural network CMAC solves the problem of filling gaps in the table with the precision that depends on the properties of recovered data.
Keywords: filling gaps in tables and lines, neural network CMAC, algorithm, simulation results

P. 58—66

E. A. Sirota

The Unstable Time Sequence Forecasting Based on the Hierarchic Multilayer Neural Network

The methods of the construction of the hierarchic multilayer neural network for unstable time sequence forecasting are provided. The analysis of the opportunity of using this hierarchic model for blood and plasma consumption process is made. The argumentation of the hierarchic multilayer neural network using advances before usual neural network in solving unstable time sequence forecasting problem is given.
Keywords: hierarchic multilayer neural network, unstable time sequence, forecasting, forecasting accuracy

P. 67—71

To the contents