This research explores the application of machine learning techniques to enhance real-time transport system optimization and predictive maintenance. By leveraging advanced algorithms and data analysis methods, we aim to create high-performance applications capable of efficiently processing large volumes of traffic data. Our approach involves developing intelligent systems that can automatically adapt to changing conditions and predict potential issues on transport routes. We implement clustering methods, such as k-means, for effective data grouping and pattern recognition. Sophisticated data reduction techniques are employed to optimize dataset management and improve system performance. Specialized technologies like GPUs and cloud platforms are integrated to enhance computational capabilities. We focus on implementing efficient data cleaning processes to filter noisy data tracks and anomalies. The Davis-Boldin index is utilized to assess the quality of clustering in transport nodes. Application-level data transfer protocols are developed for distributed computing environments. TCP stream control algorithms are optimized for improved web service performance. Our methodology addresses challenges in real-time data processing, traffic flow prediction, and system stability. These approaches contribute to the development of more efficient and responsive transport infrastructure management solutions. The proposed methods demonstrate significant potential for improving transportation network performance and reducing operational costs through enhanced predictive maintenance capabilities.

This paper investigates the application of genetic algorithms (GAs) in the multi-criteria optimization of transportation networks, with a focus on their role in logistics management. Key considerations include fuel expenses, trip duration, environmental regulations, and vehicle maintenance. The paper provides a detailed examination of the objective function, which includes factors such as loading and unloading queues, traffic congestion, unexpected events, and other variables. Additionally, the paper presents a comparative analysis of various optimization methods, highlighting the benefits of GAs in reducing costs and improving delivery efficiency. The research delves into how genetic algorithms can be utilized for optimizing transportation systems by addressing cost control, delivery timelines, and environmental impact. The objective function discussed covers aspects such as fuel consumption, vehicle upkeep, waiting times, and other critical elements. The paper also offers a succinct comparative analysis of alternative optimization techniques, supported by graphical representations of the results. This comparison aims to showcase the effectiveness of genetic algorithms in solving complex multi-criteria problems and to emphasize their advantages over conventional methods.

This article discusses the issue of developing transport systems based on metaverse technology for planning decisions on managing transport suburban in an augmented reality and virtual reality environment when collecting and analyzing digital data to understand the forecast of the future display of the virtual world based on the rapid response of various transport units based on rapid training on real simulated objects of the transport and road network. Thus, the use of virtual reality in planning a transport suburban helps to increase the efficiency and safety of the transport system, as well as improve the quality of life of citizens. The mathematical model of the transport network under study considered, based on the technology of the metaverse, augmented reality and data from the digital twin of the urban suburban.

This research explores the application of deep learning techniques as a prominent Data Mining methodology for constructing a predictive model of transport demand. Beyond access to high-quality data, analysts face challenges in consolidating computational resources and integrating foreign software with domestic systems. Key aspects in model development involve utilizing existing algorithms for data analysis through a classifier object. The classifier is then trained on various datasets, including those generated and collected from road transport infrastructure devices. Simultaneously, patterns of respondents' movements along randomly selected routes or digitally designed paths are investigated. Predicting transport demand through probability functions for collecting and analyzing passenger transaction datasets is crucial for transport infrastructure and service planning. In modern multimodal transport systems, transport demand distribution between modes is the result of individual users' random choices from available alternatives. Deep data analysis in transport networks optimizes system efficiency and safety. Advanced data collection technologies (sensors, GPS, monitoring systems) provide extensive information for improved route planning, traffic forecasting, enhanced resource utilization, and innovative transport infrastructure management strategies. Data mining plays a pivotal role in predicting transport demand in behavioral models, enabling extraction of valuable insights from large passenger and traveler behavior datasets. Various data mining methods and models facilitate identification of hidden patterns, trends, and correlations in transport demand, ultimately enhancing forecasting accuracy and system optimization. By employing clustering analysis, researchers can identify passenger groups with similar preferences, facilitating targeted service strategies. Classification models enable prediction of transport service demands for different demographic groups or time periods. Associative rule analysis reveals connections between factors like ticket prices, schedules, and seasonal fluctuations, aiding in optimal pricing and resource management. Anomaly detection methods identify unusual transport demand patterns, allowing for swift responses to changes and risk minimization. In conclusion, data mining significantly contributes to transport demand prediction in behavioral models, offering more accurate and efficient solutions for managing transport systems and improving passenger service quality.

This article explores the new reality of using Large Language Models (LLM) to manage digital data of transport infrastructure based on the concept of artificial intelligence and deep learning of the transport network model as a digital twin of agglomeration. Innovations in the field of artificial intelligence based on data make it possible to make reliable predictions and make optimal decisions to solve scientific problems in this study, but it faces a number of critical problems, including high system complexity, large search space, incomplete knowledge and small amounts of data, and all this requires new strategies to effectively solve these problems. In combination with Artificial General Intelligence (AGI), vehicles can use this data to make more complex decisions, for example, choosing the optimal route depending on the current traffic situation and predicting its changes. LLM's artificial intelligence technology can also be used to create safer vehicles. For example, the system can automatically respond to changes in the traffic situation, preventing traffic accidents and minimizing risks to passengers and others.

In the rapidly evolving landscape of digital communications, Virtual Private Networks (VPN) have become indispensable for ensuring secure and private internet connectivity. This paper delves into the technical advancements and efficiency of modern VPN technologies, with a focus on WireGuard, a relatively new entrant that has garnered significant attention for its streamlined architecture and enhanced performance. Through comparative analysis, WireGuard is evaluated against traditional VPN solutions such as OpenVPN and L2TP/IPsec, across multiple dimensions including security, performance, ease of configuration, and codebase efficiency. Our research methodology employs empirical data obtained from performance tests using standardized tools, analyzing metrics such as data transfer speed, encryption security levels, and network latency. The findings reveal that WireGuard offers substantial improvements in terms of speed and reliability, attributed to its use of state-of-the-art cryptographic protocols and a more efficient codebase. Moreover, WireGuard's integration into the Linux kernel signifies a leap towards broader adoption and compatibility across different platforms. The paper aims to provide a comprehensive overview of VPN technologies' current state, spotlighting WireGuard as a potent solution that balances security with performance. This study contributes to the ongoing discourse on enhancing digital security and network efficiency, offering insights for both academia and industry professionals looking to navigate the complexities of VPN implementation in corporate networks and beyond.

In this paper, an approach to training and evaluating an adapter model for the popular language model "zephyr-7b-beta" is described. The adapter was developed to improve the performance of the base model in tasks related to programming and understanding the Russian language. Considering the high quality of the original model in tasks in the English language, the goal of the research was to expand its linguistic and technical spectrum. The proposed adapter was trained using a large and diverse dataset, including question-answer pairs related to programming, as well code-related texts in Russian language. The applied training methodology ensures an improvement in the model's quality of answers in understanding and generating Python code based on Russian instructions. We evaluated the performance of the base model with the installed adapter using various metrics, comparing it to the base model as well as other state-of-the-art models in this field. The obtained results showed significant improvement, both in tasks related to writing Python code and in processing the Russian language, confirming the effectiveness of the proposed adapter.

This article discusses the problem of implementing the Data Fabric architecture as a technology for sustainable data processing and storage for the modernization of industrial platforms in transport and road systems. To perform complex tasks like monitoring road transport facilities entails huge resources for storing interactive data online. The concept of transport enterprises moving to cloud systems opens up gaps in data management in hybrid and multi-cloud systems. To solve such problems, the implementation of the Data Fabric structure allows you to integrate data and metadata for use on various platforms. The sustainable development of such platforms is facilitated by the use of the Data Fabric mechanism as a built-in system for standardization and coding, the implementation of semantic compatibility, proprietary knowledge graphs and the domain of data values of the transport industry. Data Fabric helps organizations solve complex data problems and use cases by managing their data regardless of application types, platforms, and storage locations, while ensuring seamless data access and sharing in a distributed data environment. The article analyzes the stability of the Data Fabric platform as a domain-oriented decomposition model based on a distributed data architecture with centralized management, but not as a fragmented data warehouse. Using the example of such a decomposition, it is necessary to model an architecture that organizes analytical and other data by areas. In this architecture, the domain's interface with the rest of the organization includes not only operational capabilities, but also access to the analytical data that the domain serves to scale the system.

This article discusses the issue of digital maturity and the use of IT trends, as well as how to use the Environmental, Social and Governance (ESG) methodology for the digital transformation of transport and road complex enterprises. In recent years, more and more attention has been paid to sustainable development, especially in the field of technology. One of the key concepts in this area is ESG as a methodology for the sustainable development of the digital maturity of the enterprise. In the modern world, companies that adhere to the principles of ESG are increasingly recognized and supported by investors and consumers. This is because such companies not only care about their profits, but also about the impact of their activities on the environment and society as a whole. Technology also plays an important role in achieving the Sustainable Development Goals. For example, the use of technologies to reduce greenhouse gas emissions or to improve the quality of life of people in different regions of the world. In addition, technology can help companies become more transparent and accountable to their stakeholders. However, in order for technology to truly contribute to sustainable development, it is necessary to take into account not only economic, but also social and environmental aspects. For example, the development of new technologies should take into account their impact on the environment and human health, as well as the social consequences of their use. In general, the direction of ESG in the field of technology is an important step towards sustainable development. Companies that pay attention to these aspects can receive significant support from investors and consumers, as well as contribute to solving global problems related to improving the quality of life and implementing ESG services. Based on an integrated approach, which includes the use of various tools and methods to achieve sustainable development goals in the road complex.

Large language models have a big influence on spheres that require data analysis and information retrieval, one of those fields is Environmental, social, and corporate governance (ESG), a field that depends on analyzing and finding correlations in big amounts of varied data from multiple sources. This paper introduces a quantum computing approach using Grover’s algorithm to create a vector database that stores embeddings based on Controlled-S gates, with each embedding represented by a binary numerical value. The creation of these embeddings, is conducted by a classical computer, while the quantum computer handles the search process through the database. This method is very efficient in terms of qubits, enabling large amounts of data to be stored within a single quantum register. And the use of quantum computers has the potential to make querying this database quick and power-efficient. By making use of this approach, managers, investors and policymakers can make informed decisions based on a wider range of ESG data, ultimately promoting environmental sustainability, social responsibility, and good corporate governance

This article discusses the methodology of the Metaverse in the context of the use of digital twins for data analysis in two environments based on a mathematical model of the agglomeration transport network using breakthrough technologies. By connecting physical elements related to transport with virtual elements, the Metaverse allows you to simulate a transport network taking into account different performance criteria with alternative changes in the movements of people and all objects of the road network. The article analyzes the completeness of data on datasets from the digital twin of the agglomeration transport network using the algorithm for finding the shortest path between graph points in the classical sense and testing this hypothesis based on the Student’s criterion to obtain the missing complete data through the integration of the digital twin and the Metaverse. One of the limitations of the proposed methodology is computational complexity. This limitation can be eliminated by choosing the R language and the Metaverse implementation software that would have the modules presented in this article. Another limitation of the proposed methodology is the inability to adequately take into account the completeness of the information. Thanks to digital technologies and the adoption of new data concepts based on digital tools, the future of mobility will be hyper automatized, more efficient, reliable and sustainable. Well-known manufacturers and developers use digital doubles to optimize the industry with the help of virtual productions, virtual prototypes for car designs, as well as modeling and testing of technical objects of the transport industry. All vehicles will be hyper-connected to digital platforms and smart devices to configure and exchange data based on transit operations to ensure uninterrupted and continuous operation in the agglomeration. Transport system planning can be carried out virtually in a hyper transport system and quickly respond to changing transport needs or problems. The Metaverse platform embedded in the digital twin should be used to plan the intelligent infrastructure of the system, which will solve the problems of the physical world with elements of a ready-made solution for future design work in the field of traffic management based on the data presented, without taking into account time or other costs. Today it is important to focus on the long-term, most sustainable and most significant areas that can be affected by Metaverse technologies.

Existing telematic monitoring systems for the state of functioning parameters and characteristics of cars and other mobile equipment make it possible to obtain a large amount of digital data on the state of objects of observation in real time. At the same time, it is possible to evaluate the features and current characteristics of their control parameters by the driver or operator. The paper proposes a fundamental possibility of obtaining predictive information about the features of changes in the characteristics of driving a vehicle by a driver based on the analysis of information received from telematic control of transport information, a block diagram of an information-mathematical system and the principles of mathematical processing of incoming information are proposed. The paper presents some results of the analysis of the time series of the selected vehicle monitoring parameters, which allow applying the Multivariate State Estimation Technique (MSET) method to their assessment in order to obtain the possibility of early detection of a potential emergency situation in driving a car (falling asleep, driver fatigue, stressful situations, inappropriate behavior). The results of the study can become prerequisites for a larger-scale application of MSET methods in monitoring the operation of mobile equipment and assessing the characteristics of its control.

Low-code technology for software development is investigated. That technology permits the creation of working applications more quickly and with minimal manual coding. Low-code platforms include visual development tools permitting the assembly of applied logic from preexisting components. By that means, problems may potentially be solved by a wide range of specialists in particular applied regions without deep knowledge of programming. That provides competitive benefits to enterprises utilizing low-code platforms. It is found that low-code technology is not confined to the corresponding specialized platforms but may utilize any tools that decrease the volume and complexity of manual operations in solving IT and business problems and minimize the departments and specialists involved in the path from an idea to a working application. It is shown that the servers and services at the core of such systems support a set of characteristics important for high-load applications and permit the rapid solution of business problems on the basis of a ready-made set of entities and standard processes.

This paper highlights a practical research of the possibility of forming quantum circuits for training neural networks. The demonstrated quantum circuits were based on the principles of Grover's Search Algorithm. The perceptron was chosen as the architecture for the example neural network. The multilayer perceptron is a popular neural network architecture due to its scalability and applicability for solving a wide range of problems. Large scale neural networks can require significant computing resources for training purposes, therefore Grover’s quantum search algorithm could be used as a way to speed up this process, thanks to its ability of simultaneously processing multiple input values of a given function.

Digitization of industrial enterprises is proposed, permitting the collection, analysis, and visualization of data regarding products, systems, machines, and facilities. The cloud platform of the Industrial Internet of Things permits the connection of any physical devices and sensors to the digital information space. Algorithms and functions included in the platform permit rapid adjustment of the system for equipment monitoring and its adaptation to the needs of the enterprise. By integrating data from physical devices and corporate systems, the enterprise may attain an unprecedented level of transparency and control over all assets and processes.

Data Fabric technology is analyzed in relation to traditional data management methods. It is shown that traditional methods of data collection and storage do not provide the speed necessary in the digital economy. Accordingly, conversion from data to knowledge is proposed, with corresponding modification of the paradigm for working with corporate information. By this means, data enriched with semantics by means of knowledge graphs become available for automatic analysis. That may be regarded as the first step toward smart information systems based on IT Architectures and Data Fabric. Formal description of the classes of real-world entities is proposed, with a corresponding hierarchy and constraints, including rules of logical inference, according to the most relevant real-time information. Data Fabric platforms are analyzed. A physical model of data fragmentation and data structuring is considered, with the creation of virtual access levels where data are logically combined. This approach permits central data management in the classical sense; the IT architectures present data in accordance with corporate standards.

The article discusses the basic principles of Data Fabric technology using the example of IBM, Microsoft and Cloudera platforms, which opened a new perspective in the field of data management, where it became possible to implement big data in the cloud, which supports the processes of organizing and preparing for use and analysis of huge amounts of information. The presented concept of Data Fabric technology can look like a system for collecting data on transport events and can be created on the principle of a blockchain network using a transport environment modeling tool. The data structures will contain data collected about accidents involving vehicles, reports of vehicle owners, roadside equipment of the transport and road complex systems. Blockchain can be used to create a secure, trusted, and decentralized autonomous Intelligent Transport Systems (ITS) ecosystem that makes better use of the legacy ITS infrastructure and resources, which is especially effective for crowdsourcing technology. While it’s not enough to just collect and store large amounts of data, businesses need to seamlessly and securely access, manage and use this data to drive digital transformation and successfully implement artificial intelligence. The article is aimed at stimulating further developments and providing useful materials for future research in this area.

The article discusses the technology of processing data arrays in Data Fabric using the QML language, which is a powerful open source data management tool. Data is an integral element of the digital transformation of enterprises. But as organizations seek to leverage their data, they face the challenge of handling data from multiple environments and platforms. This predicament with multi-dimensional data becomes even more difficult when hybrid and multi-cloud environments and architectures are implemented in the organization. Today, for many businesses, operational data has largely remained siled and hidden, resulting in massive amounts of big unstructured data. While it’s not enough to simply collect and store large amounts of data, companies need to seamlessly and securely access, manage, and use that data to drive digital transformation and successfully implement artificial intelligence. Data must not only be collected and stored, but also structured and analyzed in order to further use it to make business management decisions. Such decisions can be made not only by shareholders and top managers, but also by middle managers, as well as all employees who need it. In other words, data and the results of their analysis in a modern enterprise should become a service. Such a service should be as flexible and scalable as possible, easily adapting to the needs and size of the organization. Data Fabric solves these serious and complex tasks. It is important to understand that this is not a "boxed" solution, which is enough to buy, install, configure integration with existing systems, and then successfully operate. We are talking about methodologies, so you need to understand how they work, how to apply them correctly. The article is aimed at stimulating further developments and providing useful materials for future research in this area.

The basic principles and stages employed in the creation of digital twins are investigated. At manufacturing enterprises, digital twins permit increase in productivity, efficient resource use, decrease in costs at all stages of the product life cycle, the creation of new products, and reworking of the organization’s business model. Management of industrial systems on the basis of digital twins is associated with the development of high-level models, taking account of the available experience regarding the introduction and creation of digital technologies.

The article is devoted to research related to the processing of uncertainty by means of tensor algebra in the conditions of the road traffic control based on wireless sensor networks determine complex controlling road infrastructure. The organization of traffic on difficult roads is one of the key challenges of the transportation industry. The solution of this problem occurs under conditions of uncertainty, which may appear at different levels of the traffic control process. The article offers a classification of uncertainty in a complex dynamic system. By the example of organizing the interaction of smart controllers in an electronic coupling, the author shows the results of applying the methods of tensor network analysis to obtain the computational base of electronic coupling network interaction of intelligent (smart) vehicle controllers. Examples of processing uncertainty of different types using the apparatus of fuzzy sets and tensor basis are considered. Tensor equations provide efficient processing of big data, obtaining information in-real-time mode, the stability of “Intelligent electronic hitch” system to changes in the topology of the connection of controllers and changes in soft- and hard- components of the connection. The use of tensors in the computational basis of the intelligent electronic hitch makes it possible successful implementation of embedded AI at the level of an intelligent (smart) controller. The results obtained in the article show that the computational basis on the example of the electronic hitch algorithm can be used in smart controller systems of any level of complexity based on wireless sensor networks.