ISSN 0236-235X (P)
ISSN 2311-2735 (E)


Next issue

Publication date:
16 March 2021

Articles of journal № 3 at 2021 year.

Order result by:
Public date | Title | Authors

1. Software implementation of the data mining module based on case-based reasoning for distributed intelligent systems [№3 за 2021 год]
Authors: A.P. Eremeev , Varshavskiy P.R., Polyakov S.A.
Visitors: 1467
The paper discusses the problems of implementing the Case-Based Reasoning (CBR) module that al-lows performing all stages of the case-based learning cycle (CBR-cycle) for distributed intelligent sys-tems. Nowadays, there is a steady trend towards widespread using of methods and tools for data mining (DM) for solving various applied problems. More and more large companies are using intelligent sys-tems and DM tools to improve their business efficiency and reduce costs. Due to developing Internet technologies and the availability of cloud computing, a promising direction in the field of artificial in-telligence (AI) is the creation of distributed intelligent systems (DIS). DIS systems are characterized by the distribution of computing and information resources, which increases adaptability, reliability, as well as the overall performance of the system due to the ability to simultaneously process large data sets. DIS systems consist of autonomous nodes (agents) that can act independently and exchange information with each other asynchronously. Since there are agents capa-ble of implementing certain intelligent functions, this is an indication of the close connection between DIS systems and multi-agent systems (MAS). The work focuses on one of the DM methods based on the use of previously accumulated experi-ence (cases). The indicated case method (CBR-method) allows solving a new problem using (adopting) the solution of a similar and already known problem. To solve data analysis problems based on cases, there is a modular case system (CBR-module) that implements the main stages of the CBR-cycle. The implemented system allows working with cases presented in a parametric and structured (based on on-tologies) form. The CBR-module is a web application implemented in the Python 3.7.5 programming language using the Flask web framework and the Owlready2 library to work with ontologies. The de-veloped CBR-module focuses on integrating into a DIS system to perform data mining based on cases. The implemented CBR-module has been tested on the example of solving the classification task us-ing data sets taken from the open repository of the University of California (UCI Machine Learning Re-pository).

2. The program for forming a stack of artificial intelligence methods to solve applied problems [№3 за 2021 год]
Authors: A.Yu. Puchkov , E.I. Lobaneva , M.A. Vasilkova
Visitors: 1281
The paper proposes an algorithm for automated selection of a stack of artificial intelligence methods for data processing based on taking into account their characteristics, such as volume, presentation format, as well as the specifics of the problem solved based on these data. The purpose of the study is to develop a system of assessments and criteria to select a stack (a list of methods reflecting different levels of their suitability for solving an applied problem) of artificial in-telligence methods, which makes it possible to effectively implement the required target parameters of data processing, as well as creating an algorithm to form such stack. The task of the study was to analyze the problem of choosing a paradigm of artificial intelligence methods, to determine an assessment system that most fully reflects the features of the subject area for which the paradigm is selected, to develop a program that forms a stack of data processing methods. Considering the variety of artificial intelligence methods, and with the aim of concretizing the research task, machine learning methods were considered. To provide the possibility of applying simpler analy-sis methods to the obtained data, a choice of statistical methods also related to machine learning tech-nology was provided. The developed algorithm for choosing a stack of artificial intelligence methods contains a multi-stage procedure. It includes the following: assessment of the parameter importance in terms of their im-pact on the target parameter; the use of a fuzzy inference system that, ranks the proposed set of pro-cessing methods based on the specified expert assessments of data characteristics; selection of web-sites and search in them for subject area keywords reflecting the task for the solution of which the pro-cessed data is used. The paper presents the program work results written in the MatLAB language and implementing the proposed algorithm, which demonstrate the correct choice of artificial intelligence methods based on the given characteristics of the input information.

3. An intelligent approach to online reporting and short-term forecasting of time series [№3 за 2021 год]
Authors: V.S. Tormozov , A.L. Zolkin , А.U. Mentsiev
Visitors: 1368
The problem of predicting future values of the complex system characteristics remains important. Nowadays, along with traditional methods for predicting time series, the theory of artificial neural net-works is used, which has proven itself well in the control field, where the use of human intelligence was previously necessary, in particular in solving prediction problems. The interest in neural networks is due to both theoretical and applied advances in this field. The paper considers the development of a software system for informing and forecasting the time series using the example of the exchange rate on the Forex stock exchange using a mathematical appa-ratus of artificial neural networks. The paper presents the designed information and functional diagram of the prediction system and a server, the conceptual and physical data model used for research of the neural network structure. A prediction server in the C++ programming language is implemented, as well as an access request processing server based on the Tomcat Web server, which generates a jsp page with current currency quotes and a forecast result. A software system has been designed and de-veloped to inform and predict currency quotes in the Forex market online. The developed and imple-mented system strictly meets the set requirements. Based on the experimental results, it can be concluded that with the help of a model for predicting the values of currency quotes, it is possible to make fairly reliable short-term forecasts. For the most testing samples, the forecast showed the correct direction of short-term change in the studied quota-tion. The constructed forecasting model helps making fairly reliable short-term forecasts.

4. An approach to modeling the functioning of interdependent energy systems under disturbances and its software support [№3 за 2021 год]
Authors: Edelev А.V., Beresneva N.M.
Visitors: 1293
The purpose of the study is to develop an approach to modeling the functioning of interconnected en-ergy systems under disturbance conditions, which has a universal character in relation to different classes of disturbances and energy complexes of different levels of the territorial hierarchy. The relevance of the work is determined by the need to ensure the energy sector sustainable func-tioning affected by various disturbances ranging from high fluctuations in energy prices and worsening weather conditions up to large-scale natural disasters and man-made disasters. The novelty of the approach in comparison with others is its focus on using high-performance com-puting to speed up calculations and to divide the energy infrastructure model into structural and func-tional parts. The structural part represents the infrastructure topology in the form of a network. The functional part describes the distribution of energy flows across the infrastructure and its response to the impact of disturbances. The division of the energy infrastructure model into structural and functional parts is a fundamental principle in modeling not only energy systems, but also energy complexes, which are con-sidered as metasystems of interconnected energy systems in the proposed approach. The energy complex model is built by combining the structural and functional parts of the models of individual energy systems based on predetermined relationships between them. The universal nature of the proposed approach is also supported by its software implementation using the meta-programming tools of the C++ language, which significantly saves time when creating software packag-es designed to assess the adaptability of energy complexes. The paper demonstrates the application of the approach on the development of a country's energy model that is used in a distributed package of applied programs for global or spatial analysis of vulner-ability, search for critical elements of the energy complex.

5. Methods for accelerating the preparation and embedding of a digital watermark using mobile devices based on Arnold and wavelet transforms [№3 за 2021 год]
Authors: Zotin A.G., Proskurin A.V.
Visitors: 1207
In recent years, digital watermarking technology has become increasingly important due to the explo-sive growth of data transmitted through unprotected communication channels. Digital watermarks can reduce the amount of transmitted information and be used to protect images (hosts) from illegal use. For a more effective implementation of the latter task, various studies are being carried out in order to improve robustness, imperceptibility and security of the watermark. In most cases, this leads to an in-crease in computational costs, which makes it difficult to use digital watermarking in mobile devices. This work proposes several ways to reduce computational costs and computation time of digital wa-termarking methods based on Arnold and wavelet transforms. The first way consists in linear interpre-tation of digital watermark and a host, so it avoids the use of double cycles. The second way is to use lookup tables (LUT) to replace direct calculations. One of these tables allows performing the Arnold transform in certain time regardless of the number of iterations. Iterations for each block are deter-mined using hash code of the secret key and specially formed tables. The third way of digital water-marks embedding time reduction is multithreaded execution implemented using the OpenMP technolo-gy. In combination with linear interpretation, this results in accelerations of 1.90, 2.56 and 3.01 times for two, three and four threads, respectively.

6. Approaches to providing data visualization on devices using modern real time operating systems [№3 за 2021 год]
Authors: P.S. Bazhenov , Giatsintov A.M., Mamrosenko K.A.
Visitors: 1266
The paper considers various approaches to developing display controller driver software for embedded systems, which usually use System-on-chip (SOC) solutions and real-time operating systems (RTOS). It also describes the main principles and design decisions of the chosen RTOS, such as portability, flexi-ble scheduling, responsiveness, etc., as well as used standards (C language, POSIX 1003.1, ARINC 653). The authors list general operating principles of the display controller hardware including support for several displays and overlays that can be used for displaying multiple video streams on one screen or achieving the effect similar to chroma keying. The proposed method for developing display controller drivers defines the main parameters of the display controller hardware and the steps necessary to show an image on a screen correctly. The meth-od also takes into account the features related to using hardware interrupts and estimating the frequen-cy required for display controller to show an image on a screen correctly with the defined screen mode. Contrary to the known methods, the proposed method takes into account various features of the many real-time operating systems: lack of dedicated API to interact with graphics hardware and resources, seamless access to hardware registers from user space and so on. The paper considers several ap-proaches to debugging software on the target hardware as well as using prototyping systems based on FPGA. Prototyping systems usually introduce additional challenges to debugging, such as low simula-tion speed. The proposed method was tested during the development of the display controller driver for the home made RTOS, specifically – an X Window System display driver.

7. The method of constructing software for military simulation complexes [№3 за 2021 год]
Author: Kostrov А.S.
Visitors: 1172
Nowadays, in order to increase the efficiency and validity of decisions taken by military com- mand and control bodies, there are military simulation complexes. They are a combination of hardware and software that ensure calculations and simulation of military operations. The analysis of the struc-ture and functionality of the software of modern modeling complexes has shown that their efficiency when used for their intended purpose is largely affected by the duration of input of initial data on po-tential adversary’s actions. However, in the general software architecture accepted for supply to the subdivisions of the control bodies, modeling complexes, a specific technological toolkit that allows minimizing the time spent on information input under the conditions of limited periods of software ap-plication has not been implemented. The paper discusses the issues of building software in military simulation complexes, which make it possible to reduce the time spent on entering initial data in the process of simulating combat operations in the aerospace sphere. The purpose of this paper is to assess the effect of replacing the actual values of the input parame-ters about the strike of a probable aerospace enemy with default values on the indicators of the combat operation simulation results. Scientific novelty is the development of a new method for constructing software for military simulation complexes. There are numerical examples of calculating default values provided. The preferred option for building software is proposed on the example of introducing tech-nological developments into the general software architecture of the modeling complex.

8. The models of analyzing and forecasting load of radar station computing and power resources [№3 за 2021 год]
Authors: A.A. Nepryaev , Shpigar A.N.
Visitors: 1233
The paper proposes a method for solving the problem of distribution of computing and power resources of a multifunctional multichannel radar station with an active phased antenna array. This method is based on using a cognitive radar resource management system that plans its real-time actions based on analyzing an external situation and an internal state. The main component of the cognitive control system is a radar scheduler based on Orman's algorithm, which takes into account available time, power and computing resources and provides an optimal distribution of the radar final resources. To solve the problem of analyzing and forecasting the distribution of power resources, there is a transceiver module thermal model; for computing resources, there is a model of the primary radar plot extraction. The transceiver module thermal model is designed to simulate thermal processes occurring during the sounding signal emission and in pauses. The processing pipeline model calculates the time required for the automated system to process the received radar information. It allows forecasting the load level of computing resources when planning the radar operation modes. The purpose of the paper is to demonstrate the results of modeling a radar scheduler showing how it is possible to analyze and forecast the available time and power resources using the developed models, as well as how to redistribute them between radar tasks. It is shown how the radar scheduler generates a time diagram and places items of temporary task bundles on the timing pipeline taking into account computing and energy resources.

9. Comparative analysis of video conferencing tools for distance learning [№3 за 2021 год]
Author: Sukhanov M.B.
Visitors: 1824
Suddenly appeared pandemic problem has led to the need to switch to distance learning and to choose video conferencing means for conducting lectures and practical classes remotely. Video conferencing software was chosen as the object of the research. The analysis of publications on the research topic has shown that nowadays it is difficult to find reasonable recommendations in the scientific literature on the choice of modern informatization tools for distance learning, and the prospects for using web and video conferencing platforms by both re-searchers and participants are poorly studied. The video conferencing software of foreign companies is very popular. The paper compares the functionality and features of video conferencing software. It considers the requirements for a computer and mobile devices, their technical characteristics are the most important for ensuring high-quality video conferencing. There is an example of studying the performance dynam-ics of the main computer resources during the Zoom video conference. The performance data is ob-tained using the “Task Manager” program, which is a part of the Windows 10 operating system. It is shown that video conferencing programs consume a lot of computer resources (CPU, RAM). The paper identifies the trends in using video conferencing software in Russian distance education – the transition to using Microsoft corporate solutions, the joint use of Office 365 and Microsoft Teams, and the widespread use of Zoom software. Since video recordings of lessons are intellectual property, import substitution of foreign video conferencing tools in the field of education is advisable. The advantages of such tools should be high reliability, security, assessment of the student satisfaction level, and a cognitive interface.

10. A mathematical model of vacuum freeze-dryer with uneven vapor distribution over the chamber volume [№3 за 2021 год]
Authors: Mokhova E.K., Gordienko M.G.
Visitors: 1293
A vacuum freeze-drying process is used to dehydrate thermolabile substances, and is widely used in biotechnology, pharmaceutical and food industries to obtain dry tissues, drugs and products without losing their structural integrity and biological activity. Nowadays, there is a need to study the influence of pressure gradients in the vacuum chamber on the drying rate of materials located at different dis-tances from the condenser. Uneven removal of solvent vapor from the surface of the material during the drying process can lead to significant differences in drying rates in different parts of the chamber. Within the framework of this work, the authors jointly used two models: a mathematical model for calculating the kinetics of the vacuum freeze-drying process (native software); Computational Fluid Dynamics (CFD) model for predicting water vapor pressure above a material surface (Ansys Fluent 17.0 software package). The data between the models was transferred by transferring the values of the mass flow of evaporated moisture calculated by the drying kinetics model, through UDF files at speci-fied time intervals. UDF data was used as boundary conditions. The obtained distributions of the water vapor partial pressure were transferred to the kinetic model as boundary conditions for the next cycle of iterations. The joint solution of two mathematical models made it possible to obtain the kinetic char-acteristics of the drying process for 2 mass. % aqueous sodium alginate taking into account the uneven distribution of water vapor over the vacuum chamber volume. The paper shows that for the ScanVac Coolsafe 100-9 dryer model at the considered mass flows, there is a rapid equalization of the pressure profile along the height of the working chamber, which ensures uniform conditions for material drying. Based on the results of a series of computational experiments, the authors give recommendations on the thermal regime of vacuum freeze-drying of medical materials based on alginate hydrogel.

| 1 | 2 | Next →