Vol 12, No 3 (2025)

The paper is devoted to the development and study of cellular automata approximation methods using machine learning models. Cellular automata are models used to study the dynamics of complex systems based on simple interaction rules. In recent years, machine learning models have become powerful tools in the field of data processing. The paper examines approaches to predicting cellular automata rules using machine learning models, considers their advantages and limitations, and proposes metrics for assessing the quality of cellular automata state predictions and the dependence of cellular automata state prediction on the number of cellular automata rule models entering the input for training. The study aims to understand how machine learning models can be used to analyze and model complex systems based on cellular automata, as well as possible prospects for the development of this approach. Based on the proposed metrics, a comparative analysis of the effectiveness of various machine learning models in predicting cellular automata rules is carried out.

The article is devoted to the description of a new variation of stochastic block cellular automata – the so-called Markov automata, a distinctive feature of which is the dynamic and stochastic formation of blocks. Examples of the simplest models of physical processes built on the basis of this type of automata are given. The expressive possibilities of the introduced model are considered in the article. In particular, through comparison with the Turing machine, the algorithmic universality of Markov automata is shown, which allows them to theoretically perform arbitrarily complex processing of symbol chains. On the other hand, the presence of the so-called mixing substitution subsystem in the system of automata rules leads to a different type of behavior of these automata, the dynamics of which is described by classical kinetic equations for chemical reaction systems. It is shown that the use of special separating symbols (membranes) in the automaton allows combining several different types of behavior in different parts of the same automaton, as well as organizing information interaction between these parts. This technique opens up the possibility of modeling the simplest biological systems – cells. Using the example of a two-dimensional version of the proposed model, it is shown how the basic one-dimensional model can be extended to the case of higher dimensions.

The relevance of the article is due to the following circumstances: the need to form a new concept of cybersecurity of the digital ruble in the system of destructive events in the digital space associated with an increase in the number of computer attacks on the credit and financial sector; secondly, the growth of means of destructive information impact on financial organizations; the tasks of identifying and preventing cyber threats in the cybersecurity system of the financial sector are becoming relevant; the need to ensure security and protection of information in the digital banking space from destructive events such as hacker attacks, viruses, fraud, as well as the importance of developing systems for monitoring constructive events, for example, process optimization, improving the quality and security of financial payments on the digital ruble platform. The purpose of the study is a technical solution to ensure the cybersecurity of its infrastructure. The following tasks were completed: the concepts of “trust infrastructure”, “cybersecurity” and its varieties were clarified; an analysis of cyberspace tools in terms of cybersecurity of the trust infrastructure in the security system of the digital ruble was carried out. Proposals aimed at improving the cybersecurity of the trust infrastructure in the security system of the digital ruble were formulated.

Graph visualization is a crucial tool for the visual analysis of interconnected data; however, traditional methods often fail to efficiently represent nested, multi-level hierarchical structures. This study proposes a method based on classical force-directed layout algorithms, specifically designed for visualizing metagraphs through a planetary interaction model of vertices. The distinctive feature of the proposed approach lies in incorporating the intrinsic weight of each vertex. This adaptation allows for the preservation of nested structures, improved graph readability, and scalability of the visualization as data complexity increases. Simulation results demonstrate the method's effectiveness across various types of nested data, including file systems, organizational hierarchies, and biological ontologies.

The paper presents a module of the device for searching the degree of optimal placement (DOPPS) for onboard computers, implementing hardware and software reorganization of the topology of the onboard computer subsystem responsible for determining the parameters of speed, altitude, and incident flow pressure under combined interference. The “early cutoff” algorithm on the Kintex-7 FPGA checks up to 1.2 · 10⁶ routes in 0.55 μs and switches the channel in less than 0.72 μs, which satisfies the 1 μs limit established by DO-178C and ARINC 664. HIL tests showed a 15% decrease in the integral “perturbed cost” ΔL and an increase in the probability of successful transmission Ps to 0.96. At the same time, the dynamic power of the CPU decreased by 1.1 W, and the peak temperature of the crystal did not exceed 55 °C. The solution is suitable for serial implementation in UAVs and modernization of manned systems without modification of certified software.

The article is devoted to the study of the current state of the industrial Internet of Things (IIoT), its advantages and disadvantages, and the identification of development prospects. The industrial Internet of Things is fundamentally changing the economic model of supplier-consumer interaction. This makes it possible to automate the process of monitoring and managing the lifecycle of equipment, organize efficient chains from supplier enterprises to consumer companies, switch to “sharing economy” models, and much more. The article presents a model of a modern 12-layer IIoT architecture. The main advantages of IIoT are highlighted, such as increased efficiency, reduced errors, increased worker safety, and energy savings. It has been revealed that IIoT-based industrial enterprise management allows for real-time monitoring of industrial systems, supply chain management, and analysis of large amounts of data, which helps improve productivity, manage inventory and energy consumption more efficiently. At the same time, the risks and problems associated with the widespread use of IIoT have been identified, and the main ones are information security and a shortage of qualified personnel. The main directions of the industrial Internet of Things development are defined as conclusions.

The aim of the study is to develop a hybrid blockchain architecture to overcome the key problems of logistics supply chains: low transparency, insufficient data security and a high share of manual document flow. The work applies methods of system analysis of existing blockchain platforms (IBM Food Trust, TradeLens, VeChain), assessment of their architectural features and limitations, as well as design of a hybrid model based on Hyperledger Fabric with dynamic smart contracts. The results: the proposed multi-level architecture combining private channels for confidential data and a public ledger for verification of key events; a module for dynamic generation of smart contracts, reducing development time by 40%; scenarios for automation of logistics processes (document flow, customs clearance, SLA management), ensuring a reduction in operating costs by 30-50% and acceleration of transaction processing up to 320 TPS. In conclusion, recommendations for integration with IoT and regulatory systems, as well as directions for further research in the field of data privacy are formulated.

The article presents a method for detecting information security (IS) threat indicators in critical information infrastructure (CII) facilities using a digital twin (DT) with an adaptive mechanism. It addresses the limitations of traditional IS approaches under conditions of scarce real attack data, challenges in testing on operational CII facilities, and difficulties in identifying targeted, evasive threats. A dual-loop method (DT loop and CII facility loop) integrated with a three-level adaptation mechanism (operational, tactical, strategic modes) is proposed. The method encompasses stages of synthetic data generation, model training/testing in the DT, detection/classification at the facility, and defines adaptation trigger. Key advantages include the ability to safely generate threat scenarios and train in the virtual DT environment, automated maintenance of threat detection models. Validation results on a synthetic model of energy facility control system show significant improvement in quality metrics after adaptation.

The article presents a concept of a software and analytical complex (SAC) designed to support the balanced development of industrial ecosystems. The architecture of the complex integrates modules for data collection, analytics, visualization, recommendation generation, and scenario analysis, enabling the consolidation of heterogeneous information on enterprises, processes, resources, and interconnections into a unified environment. Special attention is given to mechanisms of data consolidation and the integration of analytical tools that provide a comprehensive assessment of ecosystem sustainability. To ensure semantic consistency and reveal hidden dependencies, the complex incorporates an ontological knowledge model. The proposed SAC is aimed at enhancing interorganizational coordination, optimizing resource utilization, and supporting decision-making under uncertainty and dynamic external conditions.

In the context of digitalization of education, there is an increasing need for intelligent decision support systems that ensure informed, adaptive and personalized management of the educational process in a digital educational environment. At the same time, the key areas are timely assessment, forecasting of learning outcomes and personalization of educational routes in a digital educational environment. The implementation of these directions is impossible without an intelligent automated information system capable of providing adaptive feedback in a digital educational environment. The purpose of the study is to describe the architecture and modules of an automated decision support information system for implementing adaptive feedback. Unlike known implementations of automated information systems, the proposed project uses a modular architecture that integrates an analytical module and adaptive feedback to support decision-making in a digital educational environment. The description of the modules and the analysis of the implemented solutions in the automated information system of decision support are given. The article implements the design stages with a description of the adaptive assessment sequence diagram, class diagram, and deployment diagram. A functional architecture is built with decomposition by levels of representation in the designed decision support system. The presented draft of the system in the article will ensure the formation of personalized and informational support for the user in the decision-making process in the digital educational environment.

This paper presents a robust mathematical model for task prioritization under conditions of multicriteria complexity, changing input parameters, and partial data incompleteness—common challenges in modern distributed and streaming digital environments. The proposed model automatically calculates criterion weights based on statistical variability (e.g., standard deviation) and dynamically adjusts them using feedback from task execution outcomes. Unlike traditional approaches such as AHP and TOPSIS – which require complete data and manual parameter tuning—the model is resistant to missing values, interpretable, and does not rely on retraining or imputation. A compensation mechanism for incomplete data and adaptation to changing feature structures is incorporated, ensuring consistent performance in fragmented and asynchronous information contexts. Comparative evaluation with machine learning models and heuristic methods shows that the proposed approach achieves high ranking accuracy (via Spearman correlation), stability under up to 50% missing data, and linear scalability as the number of tasks and criteria increases. Experimental results on synthetic and semi-real datasets confirm its practical effectiveness. The model is applicable in a wide range of digital platforms, including decision support systems, DevOps, logistics, monitoring, and incident management, especially where adaptability and transparency are critical under uncertainty and dynamic change.

The article discusses a distributed system for intelligent video stream analysis designed for automatic detection of destructive behavior in educational institutions. The relevance of the study is due to the need to improve the efficiency of security systems in organizations where existing systems demonstrate significant limitations in response speed and objectivity of assessment. The purpose of this work is to develop the architecture of a distributed system for intelligent video stream analysis based on a three-level video data processing pipeline to identify destructive behavior. The main focus is on methods for processing and annotating video data within a three-level pipeline, including object detection (YOLO), behavior classification (CNN) and contextual event analysis. The system based on the proposed architecture and modules of the three-level pipeline allows for effective detection of destructive behavior, which demonstrates the promise of using neural network technologies to create intelligent security systems in organizational structures. In this article, the authors consider the use of a distributed system for intelligent analysis of video streams in educational institutions, but the proposed solution can be adapted for other organizational structures where prompt detection of aggression, fights and other forms of destructive behavior in crowded conditions is required.

The article examines the wave and corpuscular nature of quantum objects using the example of the Mach–Zehnder interferometer and discusses the possibility of the so-called “supertunnel effect”. It is shown how the behavior of a photon in an interferometer is determined not by switching between a wave and a particle, but by the preservation or loss of coherence of its quantum amplitude. Key mechanisms are analyzed: formation of superposition at the beam splitter, interference from coherent amplitude recombination, decoherence induced by path information leakage, and recovery of interference in quantum-eraser schemes. Analogous phenomena for electrons, neutrons, atoms and large molecules are discussed, with attention to dominant decoherence sources (collisions, thermal radiation, internal degrees of freedom) and the shrinking de Broglie wavelength of massive objects. The influence of mass, momentum and barrier parameters on tunneling probability is treated, and practical strategies to enhance tunneling (barrier engineering, resonant tunneling, collective effects, and reducing effective mass) are outlined. The conclusion is that quantum laws are universal: wave-like and particle-like manifestations depend on the experimental context and coherence preservation rather than an intrinsic conversion of the object. The concept of “supertunneling” is framed as potentially realizable only if decoherence and exponential suppression can be overcome, with suggested routes for experimental pursuit.

The efficiency of photovoltaic (PV) systems is significantly influenced by the tilt angle of solar panels, especially in regions with varying solar insolation across seasons. This study investigates the optimal tilt angle for a 10 kW solar-powered system installed in the Parkent district of Uzbekistan, a region characterized by a continental climate and high solar irradiance. Based on empirical formulas, the research identifies 33° as the fixed optimal tilt angle for year-round operation. Seasonal adjustments offer marginal gains, with two- and four-season tilt configurations improving performance by up to 4%. The findings highlight the importance of site-specific tilt optimization in maximizing solar energy harvesting, which is particularly relevant for autonomous renewable energy systems used in hydrogen production.

This paper presents a meta-analysis of four experimental studies from the Norm! project, aimed at systematically studying the effectiveness of large language models in the legal field. The study includes a comparative analysis of junior and senior models, optimization of system prompts, and testing of multi-agent architectures on tasks in Russian family and civil law. A key discovery was the identification of a nonlinear relationship between architectural complexity and the quality of results: the transition from simple to complex systems provides a slight increase in quality (15–40%) with an exponential increase in resource costs (by a factor of 10–15). The flagship models GPT-4.1 and Gemini 2.5 Pro demonstrate superior quality (9.04 and 8.52 points), but junior LLMs with efficiency coefficients up to 130.3 remain cost-effective. A universal problem area for all architectures is tasks requiring an integrative analysis of multiple legal norms. The results form scientifically sound recommendations for various implementation scenarios: from mass consulting services to specialized legal applications, defining the prospects for the development of hybrid architectures in legal practice.