This article is devoted to the development of a method for detecting defects on the surface of a product based on anomaly detection methods using a feature extractor based on a convolutional neural network. The method involves the use of machine learning to train classification models based on the obtained features from a layer of a pre-trained U-Net neural network. As part of the study, an autoencoder is trained based on the U-Net model on data that does not contain images of defects. The features obtained from the neural network are classified using classical algorithms for identifying anomalies in the data. This method allows you to localize areas of anomalies in a test data set when only samples without anomalies are available for training. The proposed method not only provides anomaly detection capabilities, but also has high potential for automating quality control processes in various industries, including manufacturing, medicine, and information security. Due to the advantages of unsupervised machine learning models, such as robustness to unknown forms of anomalies, this method can significantly improve the efficiency of quality control and diagnostics, which in turn will reduce costs and increase productivity. It is expected that further research in this area will lead to even more accurate and reliable methods for detecting anomalies, which will contribute to the development of industry and science.

Keywords: U-Net, neural network, classification, anomaly, defect, novelty detection, autoencoder, machine learning, image, product quality, performance

The problem of planning the sending of messages in a cellular network to destinations with known needs is considered. It is assumed that the costs of transmitting information on the one hand are proportional to the transmitted volumes and the cost of transmitting a unit of information over the selected communication channels in cases of exceeding the traffic established by the contract with the mobile operator, and on the other hand are associated with a fixed subscription fee for the use of channels, independent of the volume of information transmitted. An indicator of the quality of the plan in this setting is the total cost of sending the entire planned volume of messages. A procedure for reducing the formulated problem to a linear transport problem is proposed. The accuracy of the solution obtained on the basis of the proposed algorithm is estimated.

Keywords: single jump function, transport problem, minimum total cost criterion, computational complexity of the algorithm, confidence interval

The article is dedicated to the development of an automated system aimed at creating a program of works for the maintenance of road surfaces. The system is based on data from the diagnostics and assessment of the technical condition of roads, in particular data on the assessment of the International Roughness Index (IRI). The development of a program of works for the maintenance of road surfaces is carried out based on the analysis of the IRI assessment both in the short term and on the time horizon of the contractor's work under the contract. The system is developed on the principle of modular programming, where one of the modules uses polynomial regression to predict the IRI assessment for several years ahead. The analysis of the deviation of the predicted IRI value from the actual one is the basis for the selection of works included in the program. The financial module allows the system to comply with the budget framework limited by the contract and provides an opportunity to evaluate the effectiveness of planning by calculating the difference between the cost of road surface maintenance and the contract value. Practical studies demonstrate that the system is capable of effectively and efficiently planning road surface maintenance works in accordance with the established contract deadlines.

Keywords: road surface, automated system, modular programming, machine learning, recurrent neural network, road condition, international roughness index, road diagnostics, road work planning, road work program

The article describes the integration aspects of the Telegram bot implemented on the 1C: Enterprise platform, into the information system for processing the results of sports competitions. The basic functionality of user interaction with the bot is considered. A diagram of the system states in the process of user interaction with the bot is provided, illustrating the possible transition states when the user selects certain commands or buttons. A diagram of the sequence of the registration process for participants of events using a Telegram bot is presented, illustrating the transmission of messages using post and get requests.

Keywords: processing the results of sports competitions, Telegram bot, messenger,1C: Enterprise platform, state processing, information systems in the field of sports

The article considers mathematical models for the collection and processing of voice content, on the basis of which a fundamentally logical scheme for predicting synthetic voice deepfakes has been developed. Experiments have been conducted on selected mathematical formulas and sets of python programming language libraries that allow real-time analysis of audio content in an organization. The software capabilities of neural networks for detecting voice fakes and generated synthetic (artificial) speech are considered and the main criteria for the study of voice messages are determined. Based on the results of the experiments, a mathematical apparatus has been formed that is necessary for positive solutions to problems of detecting voice deepfakes. A list of technical standards recommended for collecting voice information and improving the quality of information security in the organization has been formed.

Keywords: neural networks, detection of voice defects, information security, synthetic voice speech, voice deepfakes, technical standards for collecting voice information, algorithms for detecting audio deepfakes, voice cloning

One of the key directions in the development of intelligent transport networks (ITS) is the introduction of automated traffic management systems. In the context of these systems, special attention is paid to the effective management of traffic lights, which are an important element of automated traffic management systems. The article is devoted to the development of an automated system aimed at compiling an optimal program of traffic light signals on a certain section of the road network. The Simulation of Urban Mobility (SUMO) traffic modeling package was chosen as a modeling tool, BFGS (Broyden-Fletcher-Goldfarb-Shanno) optimization algorithm was used, gradient boosting was used as a machine learning method. The results of practical research show that the developed system is able to quickly and effectively optimize the parameters of phases and duration of traffic light cycles, which significantly improves traffic management on the corresponding section of the road network.

Keywords: intelligent transport network, traffic management, machine learning, traffic jam, traffic light, phase of the traffic light cycle, traffic flow, modeling of the road network, python, simulation of urban mobility

It is propossed to use foggy calculations to reduce the load on data transmission devices and computing systems in GIS. To improve the accuracy of estimating the efficiency of foggy calculations a non-Markov model of a multichannel system with queues, "warming up" and "cooling" is used. A method for calculating the probalistic-temporal characteristics of a non-Markov system with queues and with Cox distributions of the duration of "warming up" and "cooling" is prorosed. A program has been created to calculate the characteristics of the efficiency of fog calculations. The silution can be used as a software tool for predictive evaluation of the efficiency of access to geographic information systems, taking into account the features of fog computing technology and the costs of ensuring information security.

Keywords: fog computing, model of a multi-channel service system with queues, “warming up”, “cooling down”, geographic information systems, Cox distribution

The paper discusses a method for constructing a nonlinear software reliability efficiency function. The proposed algorithm is based on the use of information about the values ​​of reliability criteria, as well as some expert judgments. This approach differs significantly from previously proposed models for assessing software reliability, which are based on a probabilistic approach. In the proposed method, in addition to objective information, subjective expert assessments are taken into account, which allows for a more flexible assessment of the reliability of software products.

Keywords: software reliability, probabilistic models, statistical models, partial performance criteria, linear programming, vector optimization, decision theory

The article consider the influence of illumination and distance on the recognition quality for various models of neural networks of embedded systems. The platforms on which the testing was carried out, as well as the models used, are described. The results of the study of the influence of illumination on the quality of recognition are presented.

Keywords: artificial intelligence, computer vision, embedded systems, pattern recognition, YOLO, Inception, Peoplenet, ESP 32, Sipeed, Jetson, Nvidia, Max

5G wireless networks are of great interest for research. Network Slicing is one of the key technologies that allows efficient use of resources in fifth-generation networks. This paper considers a method of resource allocation in 5G wireless networks using Network Slicing technology. The paper examined a model for accessing radio network resources, which includes several solutions to improve service efficiency by configuring the logical part of the network. This model uses network slicing technology and elastic traffic. In the practical part of the work, transition intensity matrices were constructed for two different configurations.

Keywords: queuing system, 5G, two - service queuing system, resource allocation, Network Slicing, elastic traffic, minimum guaranteed bitrate

The article examines methods for assessing the structural stability of raster images. The study proposes a comprehensive approach, including texture analysis, color characteristics, and object shape analysis. The author presents experimental results demonstrating the effectiveness of the proposed method on various types of images. The findings obtained enable the optimization of processes for processing and storing graphical information, which is important for various fields, including medicine, geology, and computer vision.

Keywords: raster image, filtering, morphology, relative mean square error rrmse, OpenCV, Python

With the development of scientific and technological progress, the use of modern data forecasting methods is becoming an increasingly necessary and important task in analyzing the economic activity of any enterprise, since business operations can generate a very large amount of data. This article is devoted to the study of methods for forecasting financial and trade indicators using neural networks for enterprises of the Krasnodar Territory. The indicators under consideration are the company's revenue for the reporting period, the number of published (available for sale) goods, as well as the number of ordered goods during the day, week and month. In this study, a multilayer perceptron is considered in detail, which can be used in revenue forecasting tasks using neural networks, and neural network predictive models "MLP 21-8-1", "MLP 21-6-1", and "MLP 20-10-1" are built based on data from the online auto chemistry store Profline-23.

Keywords: automated neural networks, marketplaces, forecasting, neural network models, mathematical models, forecasting methods

The article discusses the features and prospects of implementing distributed management of critical urban infrastructure facilities based on the principles of autonomy. Based on the analysis, the main technologies, directions of development and features of energy transfer in an urban environment are highlighted, contributing to the introduction of distributed management of urban infrastructure facilities. The study focuses on the analysis of the distributed structure of integrated security of critical urban infrastructure facilities and the development of general principles of distributed management of critical infrastructure facilities using the «Autonomous Building» technology. t is shown that the reliable and safe functioning of critical infrastructure facilities in the city is ensured through the synthesis of special technical systems for complex protection of the facility from major security threats based on the combined use of elements of life support and safety systems. At the same time, technical life support systems for autonomous objects of critical infrastructure of the city are built on the basis of the combined use of autonomous energy sources, including non-renewable energy sources, on the principles of joint operation of electric and static power converters, storage, frequency regulation and energy conversion, and technical safety systems of autonomous objects are built using combined optical and electronic means event detection and recognition with the ability to control the full spectrum of electromagnetic radiation.

Keywords: distributed management, technology, energy, energy transfer, urban infrastructure, critical facility, electrification, decentralization, automation, autonomy

The present paper examines the actual problem of using graphics processing units (GPUs) in computing processes that are traditionally performed on central processing units (CPUs). With the development of technology and the advent of specialized architectures and libraries, GPUs have become indispensable in areas requiring intensive computing. The article examines in detail the advantages of using GPUs compared to traditional CPUs, justifying this with their ability to process in parallel and high throughput, which makes them an ideal tool for working with large amounts of data.are accidents caused by violations of rules and regulations at work sites, among them cases related to non-compliance with the rules of wearing protective helmets. The article examines methods and algorithms for recognizing protective helmets and helmets, and assesses their effectiveness.

Keywords: graphics processors, GPU, CUDA, OpenCL, cuBLAS, CL Blast, rocBLAS, parallel data processing, mathematical calculations, code optimization, memory management, machine learning, scientific research

Equipping roads with communications is complicated by the almost complete lack of roadside infrastructure, including power lines, as well as difficult terrain. When emergencies occur on this kind of country roads, residents are forced to seek help from nearby settlements that are well-connected. Therefore, providing suburban routes with communications is a key social task. Using an existing base station as an example, this article calculates the attenuation and propagation range of a radio signal for LTE technology and GSM technology, provides a comparative analysis, and uses methods of mathematical modeling and system analysis.

Keywords: LTE, GSM, Okumura-Hata model, Lee model, Longley-Rice model

The development of digital technologies stimulates widespread automation of processes in enterprises. This article discusses the problem of determining the values of the oil indicator of a transformer from the resulting image using computer vision. During the study, the device of the MS-1 and MS-2 oil indicators was studied and the features that must be taken into account when recognizing the device in the image and determining its value were considered. Based on the processed material, a method for recognizing device elements in an image has been developed using the OpenCV library and the Python programming language. The developed method determines instrument readings at different angles of rotation and in different weather conditions, which confirms the effectiveness of the proposed method.

Keywords: technical vision, oil indicator, contour recognition, OpenCV library

This study examines the combination of several non-destructive partial discharge (PD) detection methods to improve the accuracy of their detection. An analysis of methods for detecting PD in high-voltage insulation, a consideration of their features, and an analysis of the possibility of combining them to achieve the most accurate measurements were carried out. Analysis of the practical effectiveness of combining methods based on the developed variations of installations operating on the principles of two or more detection methods. Options for installations for PD detection that combine two or more detection methods are considered. A conclusion is given about the possibility of combining various methods of detecting partial discharges, taking into account the peculiarities of this type of combination. The simplest and most effective at the moment is the use of measuring cells that combine electromagnetic and acoustic detection methods.

Keywords: partial discharges; non-destructive testing of insulation; high voltage insulator; diagnostic methods for insulators

The paper describes a multifactorial nonlinear regression model of revenue dynamics of the mining and metallurgical company Severstal, based on retrospective information for 2009-2021. Production volumes by type were used as independent variables: hot-rolled and cold-rolled sheet, galvanized sheet and sheet with other metal coating, rolled products, large diameter pipes, other pipe products and profiles. The criteria of multiple definition and Fisher, as well as the average absolute approximation error, were used as criteria for the adequacy of the model. A model competition was held to select the best regression dependence. As a result, a model is constructed containing inverse transformations of two independent variables in the right part.

Keywords: regression model, least squares method, adequacy criteria, mining and metallurgical company, revenue, model competition

The article describes the automation of the audio recording recognition process in order to identify the ordered song on the radio station. The Golos Russian speech recognition model from the SberDevices was used. An algorithm for correcting the text obtained as a result of audio recording analysis using the Golos model based on the Levenshtein distance method has been developed. For recognized requests from radio listeners, interaction with the DIGISPOT II database is organized (formation and execution of queries to search for artists and their songs).

Keywords: speech recognition, Golos, Digispot II

Stepper motors are often used in automated laser cutting systems. The control circuit of a stepper motor requires a special electronic device - a driver, which receives logical signals as input and changes the current in the motor windings to provide motion parameters. This research study evaluated stepper motor drivers to determine the feasibility of their use - PLDS880, OSM-42RA, OSM-88RA. To control the system, software code was written, which was connected to the controller via a link board. With each driver, in different modes, optimal parameters were selected (initial speed, final speed and acceleration), that is, the movement of the carriage without stalling for ten passes with a minimum travel time. The results of the experiments are presented in the form of tables.

Keywords: laser, laser cutting, automation, technological process, stepper motor, performance, driver, controller, control circuit, optimal parameters

In the work, based on the previously constructed multifactor dynamic regression model of water level in the Iya River (Eastern Siberia), the authors forecast this indicator for June 2023 in three options: pessimistic, optimistic and neutral (base). A comparison of the forecasting results with the actual value of the water level confirmed the high adequacy of the model and good prospects for its future successful use to solve a wide range of applied and practical problems.

Keywords: regression model, river water level, lag time, seasonal variable, forecast, adequacy, criteria

This paper analyzes the performance of solving the classification problem using various open-source artificial intelligence and machine learning libraries in the field of marketing and customer relationship management; based on the results of experiments, the best library is selected for the purpose of introducing artificial intelligence into domestic CRM systems based on numerical performance indicators.

Keywords: artificial intelligence, machine learning, big data, classification, marketing, customer relationship management, import substitution, open source

The integration of blockchain technology with the Internet of Things (IoT) offers transformative potential for various sectors. This article delves into 16 distinct methods of integrating blockchain with IoT, emphasizing that there isn't a one-size-fits-all solution. Each method has its unique advantages and potential drawbacks, necessitating careful consideration based on specific IoT system requirements. For instance, integrating with public blockchains like Bitcoin and Ethereum offers transparency and decentralization but faces scalability issues. Sidechains introduce flexibility but might pose security risks. Blockchain platforms like Hyperledger Fabric expedite development but can lead to vendor lock-in. The diversity of these methods underscores the importance of a well-thought-out approach tailored to the specific needs of the IoT system in question. As technology evolves, we anticipate more innovative approaches to emerge, emphasizing the continuous need for research, experimentation, and collaboration to fully harness the potential of IoT and blockchain integration.

Keywords: blockchain , IOT , BIoT, smart contracts, IoT security

An approach to solving the problem of justifying a project for the development of a complex system, taking into account the possibility of realizing the values of its characteristics and the presence of restrictions on their values is proposed . The general goal is formulated in terms of the theory of fuzzy sets and the theory of possibilities. Based on the global development goal of a complex system, local goals and objectives are determined. The problem of taking into account factors influencing the achievement of a global goal is considered. Achieving the general goal is associated with taking into account restrictions on the values of system indicators, as well as the need to determine their optimal values. Unlike existing models, the proposed approach allows us to consider a more general relationship between the elements of the system and take into account the fuzzy nature of the implementation of the parameters. To solve this problem, an effective decomposition method has been developed.

Keywords: diversification of management, production diversification, financial and economic purposes of a diversification, technological purposes of ensuring flexibility of production

The study presents an extensive analysis of methods for low-level optimization of the matrix multiplication algorithm for computing systems with shared memory. Based on a comparison of various approaches, including block optimization, parallel execution with OpenMP, vectorization with AVX and the use of the Intel MKL library, significant improvements in the performance of the resulting software implementations are revealed. In particular, block optimization reduces the number of cache misses, parallelism effectively uses multicore, and vectorization and Intel MKL demonstrate maximum acceleration due to more efficient software optimizations. The obtained results emphasize the importance of careful selection of optimization methods and their compliance with the architecture of the computing system in order to achieve the required performance parameters of the designed software.

Keywords: low-level optimization, block optimization, parallel execution, OpenMP, vectorization, AVX, Intel MKL, performance, benchmarking, matrix multiplication