В статье рассматривается концепция применения роевого интеллекта и эмерджентных свойств распределённых систем автономных необитаемых подводных аппаратов в задачах подводной логистики. Показано, что переход от одиночных аппаратов и централизованных схем управления к децентрализованным роевым системам позволяет сформировать качественно новые свойства транспортных систем, включая повышенную энергетическую эффективность, отказоустойчивость и адаптивность к изменяющимся условиям подводной среды. На основе положений теории сложных систем и синергетики проанализированы механизмы самоорганизации и коллективного поведения АНПА, приводящие к возникновению эмерджентных эффектов на уровне транспортной системы. Предложена классификация эмерджентных эффектов, имеющих практическое значение для подводной транспортировки грузов, а также рассмотрены синергетические механизмы, обеспечивающие снижение удельных энергозатрат при коллективном движении аппаратов. Обсуждаются перспективы практического применения роевых систем АНПА в подводной логистике, включая контейнерные перевозки и транспортировку экологически чувствительных грузов, а также ограничения и направления дальнейших исследований. The paper examines the application of swarm intelligence and emergent properties of distributed autonomous underwater vehicle (AUV) systems to underwater logistics tasks. It is shown that the transition from single-vehicle and centralized control architectures to decentralized swarm-based systems enables the formation of qualitatively new transport system properties, including increased energy efficiency, fault tolerance, and adaptability to changing underwater environmental conditions. Based on concepts from complex systems theory and synergetic, the mechanisms of self-organization and collective behavior of AUV swarms leading to the emergence of system-level effects are analyzed. A classification of emergent effects that are practically significant for underwater cargo transportation is proposed, and synergistic mechanisms responsible for the reduction of specific energy consumption during collective vehicle motion are discussed. The prospects for practical implementation of swarm-based AUV systems in underwater logistics, including container transportation and the handling of environmentally sensitive cargoes, are considered, along with the limitations of the proposed approach and directions for further research.
Стаття присвячена створенню мультимодальної системи прогнозування Bitcoin, яка об’єднує традицiйнi ринковi показники з аналiзом новин через нейромережi LSTM та GRU. Завдяки використанню GDELT та моделi FinBERT авторам вдалося видiлити вплив геополiтики й фiнансiв на крипторинок, що пiдняло точнiсть прогнозiв на 15-хвилинних iнтервалах з 53,2% до вражаючих 77,8%. Головна особливiсть пiдходу — механiзм щотижневого адаптивного донавчання, який рятує модель вiд застарiвання, та виявлення 30-хвилинної затримки, з якою макроекономiчнi новини реально вiдображаються на цiнi. Наукова новизна зосереджена на алгоритмi автоматичного коригування ваг мережi, що дозволяє системi самостiйно пiдтримувати актуальнiсть в умовах хаотичного ринку.
Запропоновано середовище імітаційного моделювання явища максимально екстрактованої вигоди MEV (англ. Maximal Extractable Value), реалізоване мовою програмування Python із використанням бібліотеки Gymnasium, яке відтворює взаємодію сховища-мемпулу, конструювальника блоків, агента MEV-екстрактора та AMM-пулу децентралізованої біржі. Формально середовище описано як розширений та частково спостережуваний процес прийняття рішень, у межах якого агент взаємодіє з дискретно-часовою моделлю епізодів, що відображає послідовність надходження транзакцій, побудови блоків і виконання swap-операцій обміну на децентралізованій крипто-біржі. Для моделювання адаптивної поведінки агента використано методи навчання з підкріпленням, а для кількісного аналізу втрат користувачів застосовано контрфактичний підхід до оцінювання, що дає змогу порівнювати результати виконання транзакцій у різних режимах впорядкування за однакових вхідних умов. У дослідженні використано раніше описаний авторами метод зменшення негативних ефектів MEV-екстракції на основі логічних часових міток Лампорта, який реалізує локальне причинно-наслідкове впорядкування транзакцій у межах окремого смарт-контракту без модифікації глобального механізму консенсусу мережі блокчейн Ethereum. Для оцінювання практичної ефективності цього підходу сформовано три сценарії моделювання: базовий сценарій без систематичної MEV-атаки для визначення накладних витрат застосування механізму захисту, сценарій систематичної sandwich-атаки для аналізу та здатності методу зменшувати втрати користувачів та обмежувати можливості MEV-екстрактора, а також сценарій параметричного аналізу, спрямований на дослідження компромісу між рівнем захисту та "вартістю" його застосування. Отримані результати показали, що запропонований метод MEV-захищеного впорядкування може зменшувати цінові втрати користувачів від sandwich-атак і, водночас, впливати на частоту відхилення транзакцій та пов'язані комісійні витрати, що вказує на наявність керованого компромісу між ефективністю захисту та накладними витратами його використання. Практична цінність роботи полягає у створенні відтворюваного середовища імітаційного моделювання для дослідження стратегічної поведінки MEV-агентів і перевірки механізмів зменшення негативних наслідків MEV у контрольованих умовах, що може бути використано для подальшого аналізу безпеки протоколів децентралізованих фінансів та проєктування нових методів впорядкування транзакцій.
У статті досліджується проблематика надмірного енергоспоживання класичних блокчейн-мереж та розробка екологічно стійких архітектур для промислової Web3-інфраструктури. На тлі глобальних кліматичних ініціатив (таких як Європейський зелений курс) та жорстких нормативних вимог (регламент MiCA) обґрунтовано необхідність системного підходу до технологічної оптимізації децентралізованих систем. Проаналізовано еволюцію протоколів консенсусу з акцентом на застосуванні оптимізованих модифікацій алгоритму PBFT (зокрема ієрархічних, репутаційних та багатолідерних моделей) як найефективнішого стандарту для корпоративних консорціумних мереж. Розглянуто переваги диверсифікації мікроархітектур, зокрема стратегічний перехід від традиційних процесорів x86 до спеціалізованих енергоефективних ARM-рішень, що здатні знизити споживання енергії вузлами на 60%. Окрему увагу приділено подоланню термодинамічних обмежень центрів обробки даних завдяки впровадженню технології двофазного занурювального охолодження (2-PIC), яка дозволяє досягти безпрецедентного показника енергоефективності PUE на рівні 1.02 у прохолодному кліматі. Визначено критичну роль рішень другого рівня (Layer 2, зокрема ZK-Rollups) та горизонтального масштабування через шардинг у радикальному розвантаженні базового обладнання та зниженні сукупного енергоспоживання. Доведено, що інтеграція алгоритмів глибокого навчання з підкріпленням (DRL) для динамічного та автономного розподілу ресурсів дозволяє підвищити пропускну здатність мереж і зменшити споживання обчислювальних потужностей на 30%. Робиться висновок, що комплексне поєднання наведених технологій гарантує оптимізацію сукупної вартості володіння (TCO) та відповідність індустріальних блокчейн-рішень сучасним міжнародним ESG-стандартам екологічної стійкості.
The subject of research covers the theoretical, methodological, and applied aspects of implementing blockchain technology and smart contracts into microgrid management systems, as well as the automation processes of energy resource exchange between participants of a distributed energy system. The purpose of this work is to investigate the methodological foundations for the application of blockchain and smart contracts in microgrids through the analysis of contemporary scientific research, systematization of approaches to consensus algorithm implementation, classification of smart contracts by application areas, and experimental verification of the proposed solutions. To achieve this goal, the following tasks were addressed: analyzing existing microgrid architectures and management methods; conducting a comparative analysis of consensus algorithms (PoW, PoS, PoA, PBFT, RAFT, etc.) regarding their applicability in private and public energy grids; developing a classification of smart contracts based on their application areas; and investigating software tools for implementing decentralized applications. Research Methods. The study employs system analysis methods to investigate microgrid architecture, comparative analysis to evaluate the efficiency of consensus algorithms, and classification methods for grouping smart contracts. For the practical part, computer modeling and experimental verification methods were used: smart contract development in Solidity, testing in the Remix IDE environment, and simulation of a local blockchain network using the Hardhat toolkit. Research results. The research systematized the methodological foundations for integrating blockchain into microgrids. It was determined that hybrid or private consensus models are most effective for energy trading within local communities. A classification of smart contracts was developed and justified, covering four levels: energy trading, monitoring, distributed management, and cybersecurity. The practical result is the implementation of the EnergyTrading smart contract, which successfully automates the process of listing offers and purchasing electricity, as confirmed by experiments in a local environment. The implementation of smart contracts allows for the creation of a reliable P2P platform for electricity trading without intermediaries, increasing economic efficiency for households. The functionality of the automated settlement mechanism was experimentally confirmed. At the same time, key challenges were identified: the limited scalability of existing blockchain solutions and the need to improve cyber defense against vulnerabilities in contract code. Further development requires adaptation of the legislative framework and modernization of the hardware components of energy grids.
У статті запропоновано модель зберігання та верифікації персональних даних на основі технології розподіленого реєстру (блокчейну), орієнтовану на підвищення довіри до цифрових сервісів. Розглянуто архітектуру системи, що включає модулі збору, шифрування, запису метаданих у блокчейн, контроль доступу за допомогою смарт-контрактів і алгоритми перевірки цілісності даних без їх розкриття. Описано формат блоку для запису, модель управління правами доступу на основі мультипідпису та реалізацію політик доступу у вигляді смарт-контрактів. Проведено експериментальне тестування продуктивності моделі в середовищі Hyperledger Fabric із використанням типових сценаріїв, зокрема перевірки освітніх і медичних записів, електронної ідентифікації тощо. Отримані результати свідчать про високу швидкість верифікації, низьке ресурсне навантаження та масштабованість. Запропоноване рішення демонструє наукову новизну завдяки поєднанню механізмів zero-knowledge proof, гнучких політик доступу й інтеграції з зовнішніми цифровими платформами через API. Розроблена модель може бути основою для створення довірених цифрових інфраструктур у сфері електронного врядування, охорони здоров’я та фінансів.
O. Kravets, B. Martynenkov, A. Tcvetkov, E. Puzhanova · 7 authors
The article discussed an algorithm for achieving mutual information coordination for a system with distributed ledger technology based on a blockchain. The goal is to develop a generalized approach to formalizing the operation of the distributed ledger technology blockchain system in the course of achieving mutual coordination, including taking into account the possibilities of implementing abnormal functions by the distributed ledger technology blockchain node of the system and grouping nodes. The rules of block chain formation in algorithms for achieving mutual information coordination are proposed. The process of achieving mutual information coordination is described. A mathematical model of the process of achieving mutual information coordination between the nodes of the distributed ledger technology blockchain system is proposed, which differs in the representation of the system by a team of finite automata with the possibility of creating associations (pools) and providing an assessment of the centralization of the system in the conditions of choosing different variants of behaviour strategy by automata.
Open access
Cybersecurity and Information Systems
Advanced Research in Systems and Signal Processing
In modern distributed information systems, the need to ensure a high level of cybersecurity, data integrity, and confidentiality under conditions of interorganizational interaction is steadily increasing.Blockchain technologies enhance transparency and trust among participants; however, traditional consensus mechanisms are accompanied by significant computational overhead, risks of centralization, and limited capabilities for protecting sensitive information.These issues are particularly acute in corporate environments of small and medium-sized enterprises, where the computational resources of network nodes are constrained while the requirements for business data confidentiality remain high.A promising direction is the integration of Zero-Knowledge Proof (ZKP) mechanisms, which enable verification of operation correctness without disclosing the underlying data.Nevertheless, their practical adoption is hindered by the high cost of proof construction for classical cryptographic primitives.In particular, for the SM3 hash function there are no efficient optimized implementations of preimage proofs, and its bit-oriented structure leads to a substantial increase in circuit size and proof generation time, making its use infeasible in resource-constrained environments.This paper proposes a dockerized private blockchain architecture oriented toward corporate environments with limited resources, combining the trust-oriented Proof of Friendship consensus with Zero-Knowledge Proof mechanisms.The key result is the development of an approach for optimizing SM3 hash preimage proofs in ZKP systems.The paper introduces principles of manual optimization of the SM3 circuit representation, including reduction of bitwise operations, aggregation of 1965 constraints, optimization of message expansion, and reduction of round depth.It is shown that these transformations significantly decrease the size of arithmetic circuits and proof generation time compared to naive algorithm translation, enabling practical use of SM3 in zero-knowledge systems and corporate blockchain solutions.The proposed approach provides a balance between blockchain transparency and business data confidentiality, forming a "trust but do not disclose" model.The obtained results establish a scientific and practical foundation for deploying privacypreserving computation in distributed information systems and for developing nextgeneration secure blockchain platforms.
The paper investigates the issues of secure functioning of a two-level blockchain with a complex mixed consensus protocol — Proof-of-Work in the main blockchain (mainchain) and Proof-of-Stake in the secondary (sidechain). The principle of building such a blockchain is based on the Proof-of-Proof protocol, where a stable blockchain (mainchain) is used to ensure the stability of the sidechain, by referring the mainchain blocks to the sidechain blocks using special transactions. Such a structure allows for faster block generation in the sidechain and, accordingly, faster processing of transactions without reducing stability and without increasing the block size. In turn, such a two-level blockchain is of the greatest interest for the creation of a cascade system of state registers, which will be guaranteed to be protected against the substitution and forgery of documents. The main results of the work areexplicit analytical expressions for estimates of the probability of double spend attack on such a two-level blockchain, under the condition of an adversary in the sidechain and in the mainchain. The expressions obtained allow finding the number of confirmation blocks in the sidechain, which guarantees security against the attack with a probability no less than a preset value. Keywords: blockchain, mainchain, sidechain, cryptocurrencies, mining, Proof-of-Proof consensus protocol, double spend attack.
The article examines the theoretical foundations for selecting algorithms and data structures to ensure secure storage and processing of metadata in IoT systems using the Ethereum blockchain. A classification of metadata types specific to heterogeneous IoT environments is presented, taking into account semantic significance, update frequency, and data criticality. Formal requirements for algorithms are formulated, covering resistance to forgery, computational complexity, scalability under high-intensity request loads, and resource efficiency in terms of gas costs and network throughput. A comparative analysis of data structures employed in the Ethereum infrastructure, including Merkle Tree, Merkle-Patricia Trie (MPT), Multi-State MPT, and GPU-accelerated modifications, is performed according to criteria such as asymptotic complexity, memory efficiency, and suitability for incremental updates. A conceptual model for organizing metadata exchange between IoT nodes and smart contracts is proposed, incorporating modules for encoding, verification, gas cost optimization, and standardized interaction interfaces. The presented results provide a theoretical basis for developing formally verified and energy-efficient solutions in the field of secure Ethereum blockchain integration with the Internet of Things.
In the modern context of information technology development, the management of labor processes in complex IT projects acquires the features of self-organization and dynamic adaptation. The article examines the principles of configuring agent interactions within the labor environment of IT projects as a tool for enhancing the efficiency of team management. The agent-based interaction model makes it possible to consider each team member as an autonomous agent capable of making decisions, adapting behavior to the task context, and interacting with other elements of the system within a distributed environment. Conceptual foundations have been developed for constructing the architecture of agent interactions, based on the principles of cognitive exchange, communicative coherence, flexible role distribution, and multilevel task management. It is determined that the key factor in the effectiveness of such interactions is the balance between agent autonomy and centralized process coordination. A systematic classification of agent configuration types is proposed: hierarchical, decentralized, hybrid, and cognitively adaptive, which differ in the level of information connectivity and the system’s response speed. The study also investigates the impact of cognitive factors on the dynamics of interactions between agents, such as trust, intellectual compatibility, role specialization, and the ability for collective learning. A model for assessing the effectiveness of agent interaction is proposed, using indicators of performance, informational transparency, decision synchronization level, and team adaptability index. It is established that the configuration of agent connections directly determines the speed of decision-making, the coherence of actions, and the level of project innovation activity. The results of the study have practical significance for building multi-agent IT team management systems, developing algorithms for adaptive resource allocation, and creating cognitive project management dashboards. The proposed principles can be used to optimize communication processes, reduce the risk of conflicts, and enhance the resilience of organizational structures under conditions of high labor environment complexity.
Open access
Information Systems and Technology Applications
Mathematical Control Systems and Analysis
Technology and Human Factors in Education and Health
In the modern context of information technology development, the management of labor processes in complex IT projects acquires the features of self-organization and dynamic adaptation. The article examines the principles of configuring agent interactions within the labor environment of IT projects as a tool for enhancing the efficiency of team management. The agent-based interaction model makes it possible to consider each team member as an autonomous agent capable of making decisions, adapting behavior to the task context, and interacting with other elements of the system within a distributed environment. Conceptual foundations have been developed for constructing the architecture of agent interactions, based on the principles of cognitive exchange, communicative coherence, flexible role distribution, and multilevel task management. It is determined that the key factor in the effectiveness of such interactions is the balance between agent autonomy and centralized process coordination. A systematic classification of agent configuration types is proposed: hierarchical, decentralized, hybrid, and cognitively adaptive, which differ in the level of information connectivity and the system’s response speed. The study also investigates the impact of cognitive factors on the dynamics of interactions between agents, such as trust, intellectual compatibility, role specialization, and the ability for collective learning. A model for assessing the effectiveness of agent interaction is proposed, using indicators of performance, informational transparency, decision synchronization level, and team adaptability index. It is established that the configuration of agent connections directly determines the speed of decision-making, the coherence of actions, and the level of project innovation activity. The results of the study have practical significance for building multi-agent IT team management systems, developing algorithms for adaptive resource allocation, and creating cognitive project management dashboards. The proposed principles can be used to optimize communication processes, reduce the risk of conflicts, and enhance the resilience of organizational structures under conditions of high labor environment complexity.
Oleg Kravets, Ali Husein, Diana Getmanskaia, Mustafa Al-Imari · 7 authors
The article discussed an automata model of the functioning process of a system with distributed ledger technology based on a blockchain. The goal is to develop a special mathematical system with a distributed ledger based on a blockchain, taking into account the specifics of implementing algorithms for mutual information coordination, the possibility of combining individual nodes into groups, and implementing alternative strategies based on the development of appropriate models and algorithms that ensure increased stability in their operation. Automata theory was used, theory and methods of mutual coordination were applied, and the mechanism of the functioning process of a system with a distributed register was implemented. An automata model has been obtained and investigated. Thus, an automata model of the distributed ledger technology blockchain system node functioning process has been developed, which differs from the node representation by a finite state machine with a variable structure and a linear tactic with the possibility of implementing non-standard functions: the formation of a branch of processed data and a temporary blockage attack, and provides for obtaining the dependence of the sequence of changing the node's behaviour strategy options on the conditions of the environment it interacts with.
Просолов, Владислав Валерійович, Халімов, Геннадій Зайдулович, Шулік, Павло Вікторович, Смірнов, Антон Олександрович · 5 authors
The subject of the research is methods for detecting attacks in networks with the Proof-of-Stake (PoS) consensus mechanism. The purpose of this experimental investigation and analysis is to evaluate the effectiveness of classical machine learning algorithms for detecting malicious nodes in blockchain systems. The tasks include the analysis of blockchain technology vulnerabilities, the creation and use of a specialized dataset for PoS networks, as well as the construction and testing of machine learning models. The main focus is placed on comparing three algorithms – Random Forest, Support Vector Machine, and k-Nearest Neighbors – in order to determine their suitability for monitoring node activity and detecting anomalies. To solve the tasks set, the following methods were implemented: modeling, empirical, and mathematical approaches were applied. Modeling consisted of software implementation of the selected algorithms and subsequent analysis of their performance using accuracy, recall, F1-score metrics, and confusion matrices. Empirical methods were realized through testing the models on a partially synthetic dataset containing more than 10,000 records of blockchain nodes and transactions. Mathematical methods involved the calculation of statistical performance indicators and the analysis of feature importance that characterizes node behavior. The achieved results include the validation of a dataset for PoS blockchains that incorporates key operational parameters of transactions and nodes, the development of recommendations for further use of machine learning models, and the testing of selected models. Conclusions. The study demonstrated that machine learning is an effective tool for identifying anomalies and malicious activity in blockchain systems. The obtained results lay the foundation for further research, which may focus on expanding the feature space, integrating deep neural networks, developing ensemble approaches, and adapting methods to different types of blockchains.
Деменко, Євгеній Євгенович, Гребеннік, Ігор Валерійович, Колмиков, Максим Миколайович
The problem of reproducibility of experiments in optimizing validator allocation in blockchain networks with Proof of Stake consensus was investigated, in particular due to the absence of standardized datasets and unified testing methods, which complicates the objective comparison of algorithms. To tackle this issue, we propose a method for building test datasets that rely on deterministic pseudorandom sequence generators and validator profiles calibrated against Ethereum network statistics. Each validator is described by a set of parameters that includes the stake size with the minimum requirement according to Ethereum standards, performance with a uniform distribution, reliability in a high range, network delays depending on the geographical proximity of participants, geographical location according to the actual statistics of validator distribution by regions, quality of network connection, and slashing history according to the violation statistics in the Beacon Chain. Three datasets of different scales were created for small, medium, and large network configurations with fixed initial values of the generators to ensure full reproducibility of experiments. A multi-criteria evaluation system was developed based on a generalized quality indicator that maximizes system throughput and minimizes load imbalance and network delays with scientifically grounded weighting coefficients. The tenfold testing protocol ensures the statistical reliability of results and reduces the impact of randomness on conclusions. The experiments conducted a comparative analysis of four allocation algorithms: a hybrid metaheuristic method based on particle swarm optimization with local search, random allocation with correction, an adapted Ethereum shuffling mechanism, and a greedy algorithm. The experimental results revealed scale-dependent efficiency of the algorithms: the hybrid method provides high optimization quality at all investigated scales, but quadratic growth of execution time limits its application to periodic offline planning of network configuration; the shuffling mechanism demonstrates stable medium-quality results with fast execution; the random method is characterized by moderate speed with variable results; the greedy algorithm shows maximum speed with deterministic results but variable efficiency depending on the network scale. The proposed method forms a basis for standardizing experimental research in Proof of Stake consensus systems. It ensures the objective comparison of new algorithmic solutions for validator allocation in decentralized blockchain networks.
Андрій Олександрович Гашко, Андрій Петрович Бондарчук, Максим Петрович Трембовецький, Олександр Ілліч Чумак
The article examines an automated method for verifying the correctness of smart contracts in the Solana blockchain network. The relevance of the research is driven by the growing popularity of Web3 applications and the need to ensure their security, as even minor errors in smart contract code can lead to significant financial losses. The primary goal is to develop an automated verification methodology for smart contracts that can detect vulnerabilities such as the absence of founder rights verification, arithmetic operation errors, and missing transaction check signatures. Using static analysis techniques in the Rust programming language, the authors propose an approach that enables rapid analysis-taking less than three minutes per contract-and automatic generation of reports on identified vulnerabilities. The methodology is based on analyzing external data flows through smart contracts, allowing for the early detection of potential threats. To automate the process, Python and Bash scripts are employed, integrating with cloud services such as Amazon Web Services to scale the analysis. Testing results on real Web3 applications demonstrate the effectiveness of the methodology, particularly in reducing analysis time and improving the accuracy of error detection. An important aspect of the research is the continuous updating of knowledge bases and analysis tools, enabling the consideration of new types of attacks and vulnerabilities. The article also highlights the importance of interoperability between different blockchain networks, which remains a challenging task but is a key element for the future development of Web3. The research results show that the proposed methodology is promising for scaling and adapting to new challenges in blockchain ecosystems such as Solana. Thus, the developed approach to automated smart contract verification not only enhances the security of Web3 applications but also contributes to their further development, ensuring stability and reliability in the dynamic evolution of blockchain technologies.
The prevalence of financial fraud poses significant challenges to global financial stability, resulting in billions of dollars in losses annually and undermining consumer trust in financial institutions. With the increasing complexity and volume of financial transactions driven by the rapid growth of digital banking and e-commerce, traditional fraud detection methodologies have proven inadequate in addressing the scale and sophistication of modern fraudulent activities. This paper seeks to investigate and delineate the development of advanced data science and artificial intelligence (AI) methodologies aimed at detecting, mitigating, and preventing financial fraud in real-time systems. By exploring a range of state-of-the-art models, algorithms, and technologies, this research aims to provide comprehensive insights into how these systems can be deployed effectively to safeguard financial operations and maintain systemic integrity. Financial fraud detection is inherently challenging due to the dynamic and evolving nature of fraudulent tactics. The emergence of techniques such as machine learning (ML) and deep learning (DL) has significantly enhanced the ability to identify complex, non-linear patterns within large datasets that were previously undetectable by conventional rule-based systems. This paper focuses on the integration of supervised, unsupervised, and semi-supervised learning methods, as well as hybrid approaches that combine different algorithmic strategies for greater detection accuracy. In the context of financial fraud, algorithms such as decision trees, support vector machines (SVM), random forests, and neural network architectures have been adapted and fine-tuned to operate under stringent latency constraints inherent in real-time processing systems. Moreover, the adaptation of generative adversarial networks (GANs) for synthetic data generation and anomaly detection is examined to bolster the robustness and adaptability of fraud detection models. A critical aspect of this research lies in the exploration of feature engineering and data pre-processing techniques to optimize the input datasets for AI models. Given that the quality of data directly influences the efficacy of predictive algorithms, innovative feature extraction, dimensionality reduction, and data augmentation methods are discussed in detail. The use of time-series analysis and sequence modeling, especially through recurrent neural networks (RNNs) and long short-term memory (LSTM) networks, is emphasized for fraud detection in transactions that require contextual and sequential understanding. Such methodologies enable the capture of temporal dependencies that are essential for detecting anomalous behaviors indicative of fraudulent activities. Additionally, the paper addresses the significance of explainable AI (XAI) in the realm of financial fraud prevention. Trust in AI-driven fraud detection systems can be undermined by their "black-box" nature, where decision-making processes remain opaque to users and regulators. As such, incorporating interpretable models and explainability tools is essential for meeting regulatory requirements and fostering confidence in automated systems. This research evaluates various XAI techniques, such as SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-agnostic Explanations), and their integration with AI models to ensure that the decision-making process can be audited and understood by human analysts. The paper also explores the real-world applicability of AI and data science-based fraud detection through case studies of financial institutions and tech firms that have implemented such systems. These case studies illustrate the challenges faced, such as the need for real-time processing, false positive management, and system scalability. Furthermore, it provides an analysis of the trade-offs between model accuracy, computational resources, and real-time performance requirements. The dynamic nature of fraud tactics demands adaptive learning mechanisms that can update models in response to new data, which brings attention to the necessity of continuous learning and model retraining protocols. Techniques such as online learning and active learning are discussed as viable solutions to ensure that models remain effective against emerging fraud patterns. The challenges of data privacy and security are also examined, given the sensitive nature of financial data. AI and ML models, particularly those deployed in real-time environments, must comply with stringent data protection laws such as the General Data Protection Regulation (GDPR) and regional financial regulations. The implications of privacy-preserving machine learning, differential privacy, and federated learning as methods to process data without compromising individual user privacy are evaluated. This aspect is critical for building trust between financial institutions and customers, ensuring that fraud detection efforts do not come at the expense of user data confidentiality. Lastly, the research covers future directions and emerging trends that could shape the landscape of financial fraud detection and prevention. The integration of blockchain technology and distributed ledger systems is considered for enhancing transparency and reducing opportunities for fraudulent activities. Advanced threat intelligence platforms that leverage cross-industry data sharing and the collective insights of AI models trained on diverse datasets are also discussed as potential avenues for mitigating fraud in a proactive manner. The role of collaborative networks and the potential for AI-driven fraud detection to be part of a larger cybersecurity framework are posited as next-generation solutions to create a more secure financial ecosystem. The findings of this research underline the significance of continuous advancements in data science and AI to stay ahead of increasingly sophisticated financial fraud tactics. While AI models have shown promising capabilities in detecting fraudulent activities in real-time, challenges such as model interpretability, scalability, and adaptability remain prominent. This paper concludes with a strategic roadmap for financial institutions, policymakers, and technology developers to enhance the efficacy of fraud prevention strategies, which include fostering innovation in AI-driven solutions, promoting the development of robust real-time processing infrastructures, and encouraging collaborative research efforts that leverage cross-sector knowledge and resources.
Розвинені демократичні країни стрімко удосконалюють інфраструктуру систем електораль-ного волевиявлення. Технологія блокчейн швидко заполонила дефіцит інновацій в різноманітних сферах людської діяльності. У системи підтримки виборчого процесу також поступово впроваджуються концепції децентралізованого реєстру зберігання голосів та виключення із парадигми голосування третіх зацікавлених осіб. Від третіх осіб, які зазвичай є фальсифікаторами голосів, ніяк не можна було позбутись. Із появою блокчейнів така можливість стає реальністю. Об’єктом дослідження є процес електронного голосування. Предметом дослідження є системи електронного голосування на децентралізованих реєстрах типу блокчейн. Метою роботи є проведення оглядового дослідження існуючих систем електронного голосування на найбільш вживаних, поширених і надійних блокчейнах Bitcoin та Ethereum. Серед досліджених блокчейн-рішень електронного голосування, нажаль, жодне не впроваджено на загальнонаціональному рівні. У майбутніх дослідженнях планується пошук систем електронного голосування на новітніх блокчейнах, зокрема на блокчейні Near Protocol.
Vyacheslav Petrenko, Фариза Тебуева, Igor Struchkov, Sergey Ryabtsev
Objective . The purpose of the work is to increase the efficiency of the functioning of agents of a cyber-physical system in the presence of agents with incorrect or malicious behavior. The goal is achieved by establishing a trusted interaction between agents and quick detection of malicious impact. Method . Trusted interaction is carried out using distributed ledger technology and agent confidence indicators. The novelty of the proposed solution lies in the fact that information about the actions of agents is stored and aggregated using smart contracts at specified time intervals. Each agent keeps a local copy of the agent interaction chain. If several agents interact with each other, then they exchange information stored in their copies of the ledger. Result . To test the proposed method, we implemented it in C++. For the experiments, the scenario of collective perception by agents of a decentralized cyber-physical environment in a specialized simulation program Contiki Cooja was used. Conclusion . The method implemented using the solutions proposed in this work showed higher efficiency than the method based on the dynamic calculation of the confidence index. The proposed solutions can be applied not only in cyber-physical systems, but also in any other systems with decentralized control.
В роботі виконано порівняльний аналіз програмних реалізацій спеціальних мереж смартфонів (SPAN), а саме FireChat, Bridgefy, Serval Project та Briar та проаналізовано їхні підходи до безпеки та автентифікації користувачів. В роботі виявлено, що існуючі механізми автентифікації користувачів недостатні та вразливі, тому запропоновано рішення, що використовує технологію розподілених застосунків Blockchain для автентифікації користувачів. Рішення написане на мові програмування Java та NextJS та використовує Ngrok для публічного доступу до застосунку з використанням SSL. Таке рішення використовує Web3 гаманець MetaMask для виконання процедури автентифікації та спеціальний API від Moralis. Результуюче рішення є більш безпечне завдяки особливостям роботи технології Blockchain та окрім автентифікації, дозволяє перевірити користувача в мережі Blockchain. Воно може бути застосоване для програмних реалізацій SPAN, а також для будь-яких інших мобільних застосунків, що потребують механізму автентифікації користувачів. Ключові слова: БЛОКЧЕЙН, СПЕЦІАЛЬНІ МЕРЕЖІ СМАРТФОНІВ, METAMASK, MORALIS, WEB3, SPAN.
Метою статті є аналіз тлумачення словосполучення Distributed Ledger Technology (далі DLT) таблокчейн, визначення спільного, та відмінного, переваги та недоліки технологій. З'ясовано, що блокчейн єпідмножиною DLT, і використовувати термін блокчейн не завжди доречно та тлумачно правильно,краще використовувати термін DLT. В патентах та статтях було виявлено, що термін DLT все жтаки був використаний раніше ніж термін блокчейн, хоча багато авторів статей, чомусь помилкововважають, що саме блокчейн породив поняття DLT.Окремо у статті розкрито, позитивні та негативні аспекти блокчейну, а також проведенакатегоризація DLT. Звернуто увагу на те, що вузли, які проводять та контролюють операції вдецентралізованих мережах, залежно від типу систем можуть по різному називатись, від валідаторівдо нод.Визначено, що DLT є більш грамотним та тлумачно вірним терміном ніж блокчейн, і використовуючийого, автор ніколи не помилиться в суті
Vladislav Gordeev, O. V. Gromov, V. K. Gromov, G. I. Litinsky · 5 authors
In the process of air transportation, a large amount of information exchange plays an important role in the timely management of aircraft flights. The work of any airline and airport consists of many processes that involve a big number of participants. One of these issues is timely aircraft refueling. The use of the blockchain technology makes it possible to process an airline request for aircraft refueling in a timely manner, make payment and exchange of accounting documents between the airline and the refueling complex. The paper gives the main definitions for the elements of the smart contracts and their interrelationships based on the blockchain technology when performing accounting operations and payment transactions for aircraft refueling. The article is devoted to a comprehensive study of the smart contract technology application in the aircraft refueling system, in particular, the exchange of accounting and payment documentation between the airline, the refueling complexes of civil aviation airports and banks. The aim of the research work is to study the application of the blockchain technology in the aircraft refueling operations. Based on the analysis it is necessary to develop a scheme for the use of the smart contract technology when aircraft refueling, which allows the parties concerned to reduce the volume of accounting and payment operations and increase the operating efficiency of the objects and subjects of the refueling process. The paper presents the chain of information passing and blockchain transformation varying from the execution of refueling operations to the execution of banking operations, payment for jet fuel and related services for aircraft refueling. Special attention is paid to the role and location of aircraft refueling facilities as a key element of the module for automatic reconciliation of accounting and payment documents in the formation of a smart contract. Based on the analysis of the blockchain technology application, a scheme of interaction among an airline, a refueling complex and a bank is proposed. The application of the proposed scheme allows the airline to pay for refueling at the time of refueling without time-consuming accounting operations and prepayment for jet fuel, thereby reducing the accounting time.