ABSTRACTS OF ARTICLES OF THE JOURNAL "INFORMATION TECHNOLOGIES" N. 3, 2014

To the contents

I. M. Zatsman

Computer and Information Science: Technological Prerequisites for Formation

Conceptual and program documents of the 7 Framework Programme of the Åèãîðåàï Union in which need of new paradigm development for information technology creation of future generations are considered. Gorn's and Shreider's approaches to the coalescence of computer and information sciences in order to form a unified knowledge domain "computer and information science" are analyzed. The description of characteristic features of this knowledge domain as possible paradigm is provided. In interests of further development of the theoretical foundations for computer and information science, the hermeticity axiom of its media (mental, social and communication, digital electronic) is formulated. Examples from medical informatics and computer linguistics, which illustrate a role of this axiom in the course of information technology creation, are considered.
Keywords: information science, computer science, informatics as computer and information science, future generation information technology

P. 3 – 12

A. N. Bozhko, B. S. Syusyukalov

Mathematical Models for Basing and Redundancy in Mechanical Systems

The paper deals with modeling of positional mechanical linkages in technical systems. We consider a hypergraph model that adequately describes the conditions basing components in the construction. In terms of this model can be expressed many design situations arising in the process of technological preparation of assembly production. We consider the effect of overbasing, which modeling is represented as a redundant relationships hypergraph. The article offers a heuristic algorithm of removing excess edges of hypergraph, which produces not excessive structure with the maximum freedom assembly.
Keywords: assembly, mechanical connection, coupling, basing, graph, hypergraph, contraction of the vertices, assembly sequence, assembly unit, the contraction of vertices, overbasing

P. 13 – 18

A. V. Vasyukov, A. S. Ermakov, I. B. Petrov, A. Ð. Potapov, A. V. Favorskaya, A. V. Shevtsov

Combining the Grid-Characteristic Method and the Method of Smoothed Particle for Computer Modeling of Elastoplastic Bodies

In this paper the combined method for the modeling of elastoplastic bodies is considered. This method is designed to combine the advantages of two methods: the SPH method and grid-characteristic method. There are two families of methods, which are optimal for the two different groups of tasks. However, the real problem may often be mixed, which would require to make substantial compromises during the choosing a numerical method. In order to solve such problems the combined numerical method GCM-SPH is developed, connecting the advantages and partially eliminating the disadvantages of two basic techniques.
Keywords: grid-characteristic method, method of smoothed particle hydrodynamics, numerical simulation, unstructured grid, the combined method GCM-SPH, high-performance computing systems, spatial dynamic problems

P. 19 – 24

A. V. Ermachikhin, V. G. Litvinov

LabVIEW in Modern Measurement Industry (Summary)

This paper provides a brief description of the graphical programming environment of LabVIEW, the history and contribution to the development of modern laboratory stands. The basic application and a hierarchy of certification experts. The examples of programs for the management and automation of scientific experiments.
Keywords: LabVIEW, measurement automation, certification, labview portal

P. 25 – 29

A. A. Tokmakova

Stochastic Optimization Algorithm for Feature and Object Selection with Application to Credit Scoring

We consider feature and object selection problem for logistic regression with application to credit scoring, and propose the selection procedure that involves stochastic optimization. To improve prediction quality, we suggest to cluster the objects and to train the model separately on different clusters. To estimate clusterization quality, we propose the modification of the ROC curve for multiclass case. Another improvement comes from the binary representation of ordered and nominal features, which allows additional dimension reduction by dummy variables grouping. The performance of the procedure is evaluated on consumer loans data.
Keywords: feature selection, object selection, credit scoring, stochastic optimization, nominal scales

P. 30 – 35

L. A. Ismagilova, E. V. Orlova

Methodological Basis and Tools for Fiscal Modeling and Forecasting

The problem of modeling and forecasting fiscal flows under uncertainty and risk is considered in the article. A set of principles was used for development an intelligent technology in taxation management system. The example of the tax potential assessment and tax revenues prediction of the administrative-territorial unit is carried out.
Keywords: fiscal and tax planning, modeling, tools, uncertainty, methods of artificial intelligence

P. 35 – 42

A. D. Ivannikov

Subject Devoted Internet Portals as an Instrument for Digital Content Aggregation

The methods to create subject devoted web portals for efficient access to electronic resources of specified domain are analyzed. Subject portals are compared with universal search engines in terms of obtained resources list relevance. The principles of aggregated resources effective categorization that make the attribute — contextual search possible are discussed. As an example, the system of Russian Federal educational portals is described.
Keywords: aggregation of digital content, subject devoted portals, attribute search, categorization of resources, educational portals

P. 43 - 48

A. V. Ivaschenko, D. G. Peysakhovich

Proactive Dispatching in the Integrated Information Space for 5PL Providers in Transportation Logistics

Considering the modern trends in transportation logistics in this paper it is proposed and validated a new approach for cargo flow scheduling automation in the integrated information space of 5 Party Logistics operator. There is described the technology of proactive dispatching and the multi-actor integrated information space structure that allows providing such an approach for transportation logistics. Its implementation is based on the information management technology for interaction of the users of logistics services.
Keywords: transportation logistics, 5PL, intelligent scheduling systems, proactive dispatching, multi-agent systems, scheduling by conditions

A. V. Skorokhodov

Aircraft Detection by its Audio Signal Spectrum using Neural Network Classifier

An algorithm for automatic identification of civil jet aircraft by its audio signal spectrum using a self-organizing neural network is developed. Shown that the range of 50—3000 Hz is an informative spectrum part for aircraft detection. The proposed algorithm can identify the aircraft at a wind speed of 5 m/s, and creating noise in the observation point greater than less than 3—5 dB background noise. The neural network architecture description and a learning algorithm are presented. The program test results of aircraft detection departing from Sheremetyevo International Airport (Moscow) are discussed.
Keywords: aircraft, identification, neural network, spectrum

P. 56 – 60

D. I. Ignatyev, A. N. Khrabrov

Modeling of Unsteady Longitudinal Aerodynamic Characteristics of Aircraft Using Neural Networks

Mathematical models of unsteady aerodynamics are required for flight dynamic problems. Various dynamic experiments in wind tunnels with different values of kinematic parameters are used for the model development. A technique for modeling of aircraft unsteady aerodynamic characteristics using neural networks is presented in the paper. Comparison of multilayer perceptron and NARX is fulfilled. NARX is shown to be more appropriate for unsteady aerodynamics modeling in the flight dynamics problems. To improve accuracy of the models that are developed using various experiments the training algorithm for heteroscedastic data based on Bayesian regularization is proposed. The algorithm is shown to improve the accuracy of neural network models.
Keywords: unsteady aerodynamic characteristics, recurrent neural network NARX, neural network training, Bayesian regularization, heteroscedastic data

P. 61 – 69

A. V. Kinarskiy, S. V. Zhernakov

Bench Debugging of Aviation Gas Turbine Engine Parameters Based on the Neuronet Technology

One of the approaches, based on the neuronet technology, to debugging of aviation gas turbine engine parameters during bench testing was described. The engineering technique for bench debugging of gas turbine engine parameters was developed.
Keywords: debugging, aviation gas turbine engine, neural networks

P. 69 – 72

To the contents