The paper presents results on modeling reflection, understood in a broad sense as the presence of an internal representation of the external world in an active agent that influences its behavior. The ability of the simplest neural network model objects of homogeneous and heterogeneous (modular) structure to solve tasks requiring the presence and use of stable internal representations of external stimuli is revealed. It is determined that these representations are decodable, i.e. based on the current type of neural activity pattern of a neural network, it is possible to determine which specific stimulus or time series of stimuli is currently being processed in it. The authors' initial assumptions made on the basis of general considerations regarding the effectiveness of neural networks of various structures in reflection tasks and the corresponding results are presented. In particular, the following effects are shown: 1) positions in the odd-even game are asymmetric under the condition of limited computational capabilities of the players; 2) formally similar tasks on reflection (the odd-even game and responding to fixed time series of stimuli according to the rules of this game) differ in the requirements for players; 3) decodable patterns of neural activity present not only in neural networks trained to respond to stimuli, but also in networks with random weight coefficients; 4) the accuracy of decoding the neural activity of recurrent neural networks with temporal heterogeneity exceeds the accuracy of the response of these networks when processing series of stimuli; 5) patterns of neural activity in homogeneous recurrent neural networks are more difficult to decode than in heterogeneous networks of comparable size. These effects illustrate the rich internal and behavioral dynamics of the simplest recurrent neural networks, which, on the one hand, is promising for research and practical purposes, and on the other hand, complicates the prediction and interpretation of their behavior.

Overfitting an artificial neural network model is the result of training taking into account both essential and insignificant features, noise. Regularization methods are intended to minimize the influence of random noise and to identify regular features. There are a number of regularization methods for 2nd generation artificial neural networks (dropout, L1-regularization, L2-regularization, etc.). But these conventional regularization methods are not suitable for the 3rd generation of neural networks, spiking neural networks (SNN), which provide more energy-efficient and biologically plausible computations. Information in SNN is transmitted using short pulses (spikes), and training occurs locally. The biological concept of brain neurons "use it or lose it" is that if a synaptic connection is not used, it weakens and disappears. The application of the biological concept to the SNN consists in imparting a temporal dependence to the synaptic weights of the network, which reduces the weight value proportionally to the "silence" time of the synaptic connection. In this paper, a new method of weight regularization for SNN is proposed, based on the pruning of unused weights during the network training, which occurs due to the weights receiving a dependence on the time elapsed since the spike. In the experiments, a two-layer SNN was used, trained according to a combined Hebbian rule, previously developed by the authors on the basis of local learning rules STDP (spike-timing-dependent plasticity) and all-LTD (all-long-term-depression rule). For training and testing SNN, the MNIST dataset (images of handwritten digits) was used: 15,000 images for training and 1,500 images for testing, only 3 classes of images out of 10 possible were used in the experiments.

The ionosphere is a part of the Earth's atmosphere with a high concentration of free electrons and ions. The characteristic features of the ionosphere include variability and heterogeneity. One of the heterogeneities is the so-called auroral oval, which determines the range of the polar lights. Recognition of the auroral oval is an important task for predicting auroral storms, since they affect the operation of long-range communication systems, navigation, communication between satellites and the ground, making it difficult or impossible. Thus, there is a need to detect and predict the movement of the auroral oval in order to be aware of the area of their possible influence in certain periods of time. Based on the available set of images obtained in the SIMuRG system, which are based on GNSS datasets, it is proposed to use the LSTM model and the CNN architecture. The paper reviews existing implementations and proposes a method for predicting auroral oval movements in images using a Convolutional LSTM architecture that combines time series processing and computer vision. The result is a machine learning model that can make predictions based on even small amounts of data.

In this paper, the classical genetic algorithm is used to solve the problem of optimum placement of nodes (hubs) in networks of wireless sensors on a 3D building model, which allows taking into account not only signal attenuation in the walls, but also in interfloor ceilings. To design the structure of an optimal WSN based on the genetic algorithm, a program in Python was developed. The results of model calculations of optimum placement of hubs are presented.

Relevance. Wireless sensor networks (BSN) due to such characteristics as low energy consumption by nodes, low cost, self-organization, distribution and small size, the ability to determine the location of events have been applied in various fields. The increase in the amount of information transmitted to the BSS has created a number of problems that have become the focus of research in the field of wireless sensor networks. They are aimed at finding ways to expand the functionality of network nodes, especially when used in environments characterized by uncertainty. The main ones are: the network node must be capable of pre-processing information, including clustering, aggregation and merging of heterogeneous data, studying the situation and making decisions, both independently and as part of a group, taking into account the data received from other network nodes. In addition, the node management system must perform distributed processing of complex requests within the network and routing with optimized energy consumption. The implementation of these requirements is possible by improving the intellectual component of the management system of both the network as a whole and an individual node in particular. The purpose of the work. Development of methods that extend the functionality of a wireless sensor network node control system based on fuzzy logic inference and pattern theory. Main results: Due to the complexity of structuring and mathematically describing the behavior of the environment in the node's control area, it is proposed to use effective patterns of response to changes in its state. For this purpose, linguistic variables and fuzzy production rules are introduced in information models of the environment. This approach made it possible to ensure a relevant correlation of the coordinate vector of the situation with a particular response pattern and reduce the amount of calculations. The rules contain references to primary data processing tasks and response procedures by means of an autonomous network node. It is proposed to use meta-guidelines to update the rule base. Meta-guidelines define the parameters of membership functions and the rules of inference when calculating attribute values. This made it possible to supplement data transfer with the transfer of knowledge required for the necessary configuration of the parameters of the algorithms of local nodes. With this approach, the network node has knowledge about itself and the environment, and becomes capable of independent decision-making as part of a group. To increase the robustness of data processing algorithms in nodes, it is proposed to use the concept of a "typical situation". As an example, the paper shows the use of the proposed approach in the tasks of integrating data from several sensors of a node and detecting an intrusion object.

The purpose of the research presented in the article was to develop digital models of traction power supply systems that make it possible to adequately determine the modes of single-phase ground faults in 6-10 kV power lines located in areas of increased electromagnetic influences of traction networks. For their implementation, the Fazonord software product, version 5.3.3.0-2024, was used, which makes it possible to determine power transmission line modes for various methods of connecting transformer neutrals to the ground: isolated neutral, as well as grounding through resistors with low and significant resistance. Simulation of single-phase faults was carried out in the Fazonord software package, version 5.3.3.0-2024, for a system including the following elements: three 110 kV power lines, two substations with 40,000 kV·A transformers, a 25 kV traction network with two contact pendants, four rail threads and a 10 kV line mounted on contact network supports on the field side. Traction loads were created by the movement of two trains weighing 3192 tons in an odd direction and the same number of trains weighing 4192 tons in an even direction. The methodology presented in the article and the computer models developed on its basis make it possible to adequately determine the modes of single-phase ground faults in technological power lines of railway transport. The technique is applicable to power lines and traction networks of any design and can be used in practice to determine single-phase fault modes, configure protection devices against these types of damage, create means of their identification and localization, and select the most effective method of neutral grounding.

The article considers the problem of calculating the quality of water in water supply systems with closed circulation loops. The conducted review of works on modeling water quality in water supply systems showed that there are many factors affecting the water quality in water supply systems. The most widespread in practice is water disinfection with chlorine. Bilateral restrictions are imposed on the concentration of chlorine in water, and operating organizations are obliged to maintain the required concentration of chlorine to the tap of each consumer. In the process of transporting water, the concentration of chlorine decreases due to interaction with the pipe material and dissolved substances. Thus, one of the main indicators of water quality can be its age. Existing models for calculating the water age do not allow calculating the age of water in water supply systems in the presence of closed circulation loops. A topological algorithm for calculating the age of water in networks with closed circulation loops is proposed. The software implementation of the algorithm is integrated into the “ANGARA-VS” information and computing complex and tested on conditional and real examples of water supply systems. The software implementation of the algorithm has shown high computational efficiency and can be applied in practice when analyzing the modes of water supply systems. It is shown that there are only a few ways to manage water quality in water supply systems. A new approach to water quality management based on the creation of closed circulation circuits is proposed. This approach requires minimal capital and operating costs compared to

 The article presents a parametric model for optimizing the production of crop products taking into account multi-level trends in crop yields and different predecessors. An algorithm for solving a parametric problem is proposed, which is tested on data on the production of crop products of a municipal district and an agricultural organization. The application of the developed model allows us to assess the prospects of production activities of enterprises under average, favorable and unfavorable conditions with different combinations of predecessors.

This paper explores the application of SQL for modeling finite state machines in the management of technological processes. Finite state machines, as a mathematical model, are widely used for automating sequential operations and managing complex systems. Here, a method is proposed in which the relational data model serves as the foundation for implementing a finite state machine, and SQL is used to define the transition logic between states through tables and triggers. The article discusses the design of state and transition tables, where each state is characterized by attributes such as timestamp and status, and transitions are defined by conditions and events that trigger state changes. Triggers created in the MySQL database environment automate the system’s logic, enabling state transitions when specified conditions are met, thereby implementing the model without the need for additional external software. The proposed solution is flexible and scalable, simplifying the addition of new states and transitions, as well as adapting the system to evolving requirements. The use of SQL and the relational model also facilitates integration with other analytical tools, enabling data collection and analysis for process optimization. This SQL-based approach to finite state machine modeling in a relational database makes the control system for technological processes more efficient, easy to maintain, and conducive to further analysis, which is particularly relevant in modern automated production environments.

Solving the diverse tasks of operational and strategic management of natural and man-made safety of territories optimal using a new structure of information resources. The management information support processes have defined using the ontology. The digital transformation of decision-making, in addition to intelligent data processing and integrated monitoring, requires changes in business processes of information using

The article discusses key indicators for assessing the efficiency of the region's innovation infrastructure
using regression analysis.
The work highlights the stages of forming a system of key KPI indicators. The first step was to define goals and objectives for creating a multi-level system of KPI performance indicators. The second step is to determine the key indicators for further analysis in order to explore which indicators may reflect the effectiveness of the region’s innovation infrastructure. For the article, the following indicators were selected: the level of innovation activity in the region (x1), the number of new jobs created in the innovation sector (x2), the volume of innovative products or services launched on the market (x3). The third step was to collect data on selected indicators and then select a research method - regression analysis.
The article examines the process of forming a multi-level system of key indicators (KPIs) to assess the effectiveness of regional innovation infrastructure. The main components of the innovation infrastructure were identified, and indicators for each of them were determined.
A regression model was used as the main method of analysis, which made it possible to identify the most significant factors influencing the development of innovation infrastructure. The main conclusion of the article is that the development of a multi-level KPI system is a necessary condition for the effective management of innovation infrastructure and achieving goals in this area.

The paper deals with the issue of concentration of data from the digital educational footprint and display of its dynamics. The comparative analysis of visualization methods is made and it is proposed to consider the modification of the method of cognitive maps of knowledge diagnosis with the account of displaying the dynamics of the educational process. The essence of the method is revealed and the stages of map formation are shown. Illustrations are prepared on the basis of experimental data of master's students of Siberian Federal University in the intelligent tutoring system AESU. In conclusion, possible directions for further improvement of the method of cognitive maps of knowledge diagnosis are given.

The paper describes a decision support system (DSS) for actions in conditions of a threat of track erosion for the Ulaanbaatar Railway. Information about the road is briefly presented, and the geoclimatic conditions of its operation are described. A mathematical model is given for kilometer-by-kilometer calculation of the threat of track erosion depending on the meteorological forecast, track condition and the presence of artificial structures. The composition of the DSS, its functionality and general principles of operation are considered.

This article considers the problem of effective management of assessment of the quality of students' training in the educational program as a whole or in its individual components. Traditional methods of creating assessment funds often face difficulties, such as ambiguity of evaluation criteria and limited adaptability to the needs of participants of the educational process. The article proposes the introduction of an automated system for the creation and management of funds of assessment tools for educational programs of the university. This will allow to more accurately and objectively assess the achievement of planned learning outcomes (competencies, indicators of competencies achievement, knowledge, skills, abilities, skills and/or practical experience) for an educational program or its individual components, reducing the human factor in the assessment process. The use of modern information technologies helps to improve the quality of education and optimize the learning process.

The article analyzes the design of an electric jib crane with a magnetic gripper. Cyclograms of the crane movement mechanism operation have been developed. An algorithm for the operation of the PLC of the crane travel mechanism control system has been developed. The work program of the PLC of the crane travel mechanism control system has been developed. A technical implementation of the method of controlling the mechanism of movement of a crane is proposed. An experimental oscillogram of the angular velocity motor of the movement mechanism В of a jib crane with a magnetic gripper is determined, confirming the operability of the developed algorithms and programs.