The article proposes a new method for teaching private classifiers, as well as a way to aggregate their forecasts as part of a committee. The training is based on the hypothesis of iterative re-entry of biological neural networks and uses the principal component method for its implementation. For private classifiers, the areas of competence are defined in the aggregate covering the training set of examples. It is shown that the iterative learning process converges in several steps, ensuring 100% recognition accuracy in the area of competence of the private classifier. The aggregation of forecasts is implemented according to the principle of maximum projection of the image onto its own subspaces of classes of private classifiers. Examples of the use of the committee of competent classifiers for the MNIST dataset are given. A continuous learning model of the classifier committee is proposedthat is suitable for building self-learning recognition systems. The neural network implementation of classifiers in the class of self-similar neural networks is considered.

The XOR problem is a classic problem in machine learning. Most machine learning methods yield unsatisfactory results when addressing the XOR problem. Attempts to improve the quality metrics of decision rules by transitioning from linear separators to polynomial ones, or by increasing the depth and number of trees, reduce the interpretability of the solution and can lead to overfitting. Therefore, this problem is usually solved using neural networks with the backpropagation method. Currently, there is a significant interest in finding alternatives to the neural network approach for solving classification tasks and the backpropagation method for training neural networks. This article proposes solving the XOR problem based on committee constructions in the form of mathematical programming tasks. Their effectiveness largely depends on the structure of the original data, which can only be understood in the process of solving the problem. Classical committees of unanimity, majority, and seniority only achieve high-quality metrics of the decision rule when the data structure is relatively simple. Therefore, it is proposed to augment the set of committee constructions with an XOR committee. The article presents its geometric interpretation, mathematical model, program listing in IBM ILOG CPLEX package codes, and a comparison with the quality metrics of solutions by other machine learning methods. Translating the process of building committee constructions into mathematical programming tasks provides additional opportunities for controlling the quality of the decision rule during its construction, rather than post facto, as occurs when using programs from standard libraries for machine learning tasks.

This paper proposes a method for aligning coordinate systems in robotic technological complexes (RTCs) using machine vision and a deep learning model, without the use of special calibration objects. The method is based on "eye-to-hand" calibration and employs the YOLOv5 model for object detection and classification in the robot's working area. The proposed approach allows for automatic transformation of object coordinates from the camera coordinate system to the coordinate systems of the robot and other RTC components, ensuring precise interaction between them. Simulation results confirmed the effectiveness of the method and its suitability for industrial applications. The method reduces reconfiguration time and enhances the flexibility of RTCs, which is especially important in multi-nomenclature small-batch production environments.

The purpose of this article is to eliminate the shortcomings in the neural network training algorithm, which include insufficiently accurate determination of the direction of movement [4], slow convergence to an extremum, and the need to use a sufficiently large number of initial simplexes. It is proposed to introduce an additional search direction into the neural network training algorithm, in relation to solving the problem of parametric optimization of artificial neural networks (ANN) contained in links with pulse width modulation (PWM) of automatic control systems. Due to the fact that ANNs are used in PWM, the tasks of training and parametric optimization are equivalent and ultimately come down to determining the weighting coefficients of the ANN.To achieve this goal, the following tasks were set and solved: 1) existing approaches used in direct search methods to improve their characteristics are analyzed; 2) conducting experiments on the use of the most common approaches, in the context of the problem of parametric optimization of systems with PWM controllers; 3) development of recommendations for their use. Ultimately, the above makes it possible to resolve the problems of speed and the number of initial simplexes that arise when solving the problem of parametric optimization of automatic control systems with a device that performs PWM using ANN. Based on the above, we can talk about the relevance of the presented article.

This article is devoted to the development of a new structural specification of regression models. Previously, the author introduced non-elementary linear regressions, in which explanatory variables are transformed using such non-elementary operations as minimum, maximum and modulus. In this article, to transform explanatory variables in a regression model, it is proposed to use the operations of rounding their values to the nearest integer downwards (floor) or up (ceiling). In mathematics and digital signal processing, this conversion process is called quantization. The well-known uniform quantizer with a rounding boundary of 0.5 is considered. A non-elementary linear regression with quantized explanatory variables is proposed. The ranges of possible values of quantization steps size for a model with one explanatory variable are determined. Based on this, an algorithm has been developed for approximate estimation using the ordinary least squares method of the proposed structural specification parameters. Using artificially generated statistical data in the Gretl package, computational experiments were carried out that confirmed the correctness of the above mathematical reasoning. All non-elementary linear regressions with quantized variables obtained during the experiments turned out to be more adequate than classical linear regressions.

Abstract. The article presents a functional diagram of the control subsystem of the electric drive of a walking excavator. Structural-parametric models of АC electric drives of an excavator with a pulse-phase control system and a pulse-frequency control system have been developed. Simulation of АC electric drives of an excavator with a pulse-phase control system and a pulse-frequency control system in the Matlab/Simulink environment was carried out. A comparison of the results of modeling control systems is given.

The paper proposes a refined model of heat transfer during continuous extrusion by the Conform method, taking into account technological and energy-power parameters, and the influence of heat transfer in the tool, which allows using an accurate description of the geometry of the deformation zone, and conducting mathematical and computer modeling of the metal temperature distribution. Numerical experiments have been carried out confirming the correspondence of the temperature field to the distribution of metal flow velocities.

The process of creating intelligent energy systems (IES) is accompanied by the widespread use of information technologies, among them data processing technologies occupy an important place: cloud technologies used to process large amounts of data, as well as edge computing, which allows you to quickly and efficiently process data locally. The article compares these technologies according to various criteria, and identifies the areas of their application in the development of the IES. Examples of using these technologies in the implementation of control functions for new facilities and structures created and functioning within the framework of digital transformation and intellectualization of power industry such as a smart home, an active distribution network, a dispatcher simulator, a virtual power plant, etc. are considered.

The objective of this paper is to develop an integrated approach to structural and parametric optimization of microgrids, taking into account the impact of various disruptions. The proposed multi-level optimization framework enables decision-makers to identify a range of promising microgrid designs within an acceptable timeframe, thereby ensuring a stable and continuous power supply to users under specified adverse conditions. In this framework, stochastic "weather generators" are crucial for generating time series of natural and climate data with varying durations that closely resemble real-world values in a given geographic location. This is done for the purpose of evaluating the effectiveness of intermediate and final computational results.

Abstract. In many cases, it is advisable to supply electricity and heat to consumers in remote and hard-to-reach areas using renewable energy sources. Designing an optimal autonomous energy system is associated with a number of difficulties: the stochastic nature of the potential of renewable energy sources, a variety of technical and economic parameters and technological limitations of the equipment. The long life cycle of an energy system and the multiplicity of goals pursued during its creation or development lead to the need for a multi-criteria consideration of the problem. The article provides an overview of methods and software for selecting configurations of energy systems, and shows the relevance of developing multi-criteria approaches. A two-stage approach to the multi-criteria selection of the configuration of an autonomous energy system is proposed. At the first stage, the nPro software tool is used, providing optimization of configurations of various energy systems, including wind turbines, photovoltaic converters, heat pumps, solar collectors, electric and thermal energy storage devices. At the second stage, a multi-criteria assessment of the formed energy systems is carried out using the TOPSIS method. To improve the validity of the solutions obtained, the weights of the criteria are determined based on an objective assessment using the entropy method, as well as a subjective method. An example of the approach application is considered for the remote settlement of Ust-Sobolevka, located in Primorsky Krai. As a result, ten configurations for autonomous electricity and heat supply were optimized and their multi-criteria assessment was carried out taking into account four criteria: capital costs, the levelized cost of electricity and heat, and carbon dioxide emissions. The most preferable configuration has relatively low capital costs and carbon dioxide emissions, as well as the best estimates of the levelized cost of electricity and heat among the options considered.

The relevance of building digital twins in the field of wind power plants (WPP) is due to the rapid development of technologies and the need to improve the efficiency of energy systems operation. Digital twins allow creating virtual models of real objects, which opens new horizons for process optimization, condition monitoring and prediction of plant operation. In the conditions of global transition to sustainable energy, the importance of digital modeling increases, as it contributes to more efficient use of resources, cost reduction and increased reliability of energy systems.
The paper discusses the process of forming a digital model of a wind power plant, which is a key element in the development of a digital twin. The paper describes the main approaches and methods used to create an accurate virtual replica of a real wind turbine, including the selection of modeling parameters such as geometric dimensions, materials and mechanical properties, as well as the definition of boundary conditions that allow for the most accurate reproduction of operational features.
Special attention is paid to the comparison of weather conditions, including wind speed time series, which is critical for estimating the utilization factor of the installed capacity of the wind turbine. These weather data allow for accurate modeling of real plant conditions and for adjusting model parameters to changing climatic conditions. The paper also discusses modern technologies and tools used to ensure a high degree of correspondence between the digital model and the physical object, such as data acquisition systems and analysis software. The prospects for the application of digital twins in the wind energy industry are emphasized in the context of optimizing plant performance, predicting power generation and improving overall operational efficiency, which is an important step towards the sustainable development of renewable energy sources.

The present geopolitical situation has caused a decrease in gas export volumes to Western countries. In this context, it is essential to explore the potential for redirecting gas flows to the east, including the domestic market and exports. Additionally, it is crucial to actively develop LNG supplies by sea. The future expansion of the gas system is planned using optimization calculations. It is impossible to build a detailed computational model of a system that sufficiently accurately describes all its components. The calculation scheme must have fewer nodes and connections compared to the detailed scheme, while still preserving all of its main characteristics. Therefore, it is crucial to clarify the existing methods for aggregating the schemes and determining the indices of their components. The proposed methodology for aggregating the initial calculation scheme allows designing an expanded calculation scheme of the Unified gas system, which can then be used to perform calculations for the future development of the gas industry in Russia and its individual regions. The paper explores potential directions for the expansion of the gas transmission system and provides a detailed gas supply scheme for the eastern part of the country. The calculation scheme was developed for Russia’s Unified gas system with a specific focus on Eastern Siberia and the Far East to investigate the region’s development in conjunction with the country’s gas system.

The article developed an ontology of the self-cleaning ability of the atmosphere from anthropogenic pollutants and identified the influence of significant factors on it. According to the website https://rp5.ru in the created database in the My SQL DBMS the average annual and average monthly values of meteorological indicators of accumulation of MPZA, MPZA2, MPZA3, and self-purification of the atmosphere PSA, UMPA were calculated and analyzed for the large industrial cities of Angarsk, Bratsk, Irkutsk with a high level of air pollution for 2010-2023. It has been shown that with a determination coefficient of 95.8-99.9%, instead of the MPZA3 indicator, the simpler MPZA2 indicator can be used, and instead of the difficult-to-calculate PSA indicator, the 1/MPZA indicator can be used. In Irkutsk in May, only in 7.1% of cases are favorable conditions for the dispersion of pollution formed in the atmosphere, while in Angarsk and Bratsk there are no favorable conditions for their dispersion.

This paper provides an overview of the current state in the field of dynamic cognitive (system) modeling and the corresponding software tools for analyzing heterogeneous factors affecting the development of complex, poorly formalized systems in the energy sector. Existing methods for analyzing multifactor dynamic systems, as well as the application of modern simulation tools for conducting qualitative analysis, are reviewed. Approaches to dynamic cognitive modeling are presented in the context of its application for analyzing the influence of complex factors and assessing systemic effects on scientific and technological development in the energy sector. The results of the study can be used both for the development of software tools for dynamic cognitive modeling and for improving approaches to the analysis of heterogeneous factors influencing the development of energy systems.

The paper examines issues related to the presentation of information on publications of scientific employees of ISEM SB RAS, as well as on scientific publications of the Institute in information and analytical materials of the Scilit system. Specific examples show the advantage of complete and correct setting of metadata for scientific publications, as well as the problems that arise when handling bibliographic information carelessly.

The article substantiates a need to change the existing approaches to organizing the students practice in technical universities. The results of survey analysis of senior students and employees, whose responsibilities include organization and management of various types of production practices, at technical universities, as well as publications in media about holding events at the local, regional and federal levels on the problems of filling the labor market, showed the need to solve the tasks of targeted training at the state level. The authors propose to change the approach to targeted education in universities and turn to targeted training of specialists for enterprises and organizations by allocating students to industrial (including pre-graduate) practice on a competitive base, taking into account their knowledge and skills, as well as the current needs of the receiving enterprise. A purpose of the research is to develop an approach to organizing a targeted training of specialists, as close as possible to the needs of enterprises and meeting the needs of graduates, as well as effective in terms of training costs. Within the proposed approach, an algorithm for allocating the students to practice has been developed, the novelty of which lies in organization of the trilateral interaction between universities, enterprises and students within the process of students competitive selection for practice and a possibility of graduates further employment on the practice bases using a single information platform. The algorithm, presented in the form of a BPMN model, is the basis for generating the functional requirements to the information system. A necessity of ontological modeling of the students practice allocating business process for preparing a practice-oriented final qualification work, which will allow determining the thesaurus of information system concepts and semantic relations between them, is substantiated.