Networks and Sustainability

The chapter addresses the urgent scientific and practical challenge of improving environmental monitoring systems in Ukraine, with a particular focus on assessing air quality and radiation levels. The relevance of the research is due to the increasing environmental risks caused by industrial activities, military conflicts, accidents at critical infrastructure sites, and climate change, which threaten public health and sustainable ecosystem functioning. Existing monitoring systems are limited in spatial coverage, have low response efficiency, and lack sufficient transparency for public use, hindering the timely identification of ecological threats. The paper proposes an innovative approach to developing a smart cyber-physical system that integrates stationary and mobile sensors, automated data collection, web and mobile platforms, and Artificial Intelligence (AI) tools, specifically Long Short-Term Memory (LSTM) neural networks, for the prediction of air pollution levels (PM₂.₅) and provide recommendations on radiation hazards. Within the study, simulation models were created to compare the effectiveness of mobile and static sensor systems under various scenarios. The experimental results demonstrate that mobile sensors detected up to 8.5 times more localized contamination points and 4 times more dynamic contamination zones than static systems. In addition, the developed LSTM-based PM₂.₅ forecasting model achieved a forecasting accuracy of approximately 92.2%. The obtained results confirm the effectiveness of the proposed system in enhancing the scalability, adaptability, and analytical capabilities of environmental monitoring, improving early warning mechanisms, and supporting the achievement of key Sustainable Development Goals (SDGs). The study highlights the importance of combining advanced digital technologies and networking solutions with active public engagement to build next-generation environmental protection systems.

IoT-based smart city workflows require energy-efficient network infrastructure but due to the resulting resource constraints are typically unable to handle traffic spikes. Using two Alcatel Lucent OmniSwitch 6465-P12 devices that cooperate, we show how offloading leads to better handling of traffic and associated intelligence computation. One OS6465 serves as a central smart network switch, managing communication and data flow between IoT devices and the other switch. The implementation and integration of three smart city workflow scenarios were tested in a lab environment, focusing on IP camera monitoring for traffic surveillance, environmental monitoring and room occupancy detection. The chapter details the setup, including dynamic QoS policies to manage video streaming effectively and ensure high-quality performance. Moreover, it discusses benchmarks performed on the switches, including packet counters, port monitoring, and the use of the cron-app for scheduled tasks. These experiments demonstrate the system's performance, efficiency, and capability to adapt to evolving smart city requirements while maintaining optimal resource utilization.

Since ChatGPT, the question of the consciousness of artificial intelligence has been around like the famous pink elephant. Philosophy, medicine, psychology, theology, biology, and other disciplines are imperative to rethink intelligence and consciousness. What is certain is that consciousness does not have to be intelligent, and intelligence does not have to be connected to consciousness. They are different qualities. But which ones? Careless statements by some AI pioneers stoked fears of AI's destructive powers and fueled fantasies of a possible takeover by AI. Justified or unfounded? The question has not been decided. And the fears are real. The article deals with different ways of approaching the phenomena of consciousness and intelligence. Philosophers ask about the “being” in consciousness. This is the ontological view. Or they divide into subjects and objects. Subjects perceive, develop will, and express interests. Objects are the phenomena absorbed by subjects in consciousness. Theologians ask about the divine origins of consciousness. Thus, we find in the Bible the image of the tree of knowledge. Awareness and knowledge come from the outside. Medicine asks about the localization of consciousness and about the material carriers of consciousness in the human brain or in the intelligent biological structures that evolution generated on earth. Psychology asks about the qualities of consciousness: thinking, learning, forgetting, remembering. The qualities show intelligence, which has a special consciousness, but are not to be equated with it. With artificial intelligence, people try to map neurological processes, simulate intelligent qualities of organisms, and generate language in a meaningful way. Much has been achieved. But not the decisive step.

New approaches to analyzing the state of rural areas have been proposed, based on a combination of advanced methods for processing geospatial data, machine learning, and artificial intelligence. A methodology for clustering villages, analyzing the development level of rural communities, and identifying directions for infrastructure improvement has been developed. This, in turn, allows for the assessment of infrastructure accessibility, taking into account not only distances but also topological connections between objects. Such an approach simplifies the analysis by enabling the automatic identification of underdeveloped areas. Thanks to open data sources (OSM, HDX), the model is easily scalable and can be applied to different regions. The methodology for assessing the quality of life in rural settlements in Ukraine using geospatial analysis makes it possible to aggregate data on the remoteness of villages from vital facilities such as hospitals, schools, roads, utilities, and more. The proposed approach provides a strong foundation for decision-making regarding the restoration and modernization of rural infrastructure. The assessment of rural development in Ukraine under wartime conditions, using GIS, multicriteria analysis, and AI-based virtual experts, allows the obtained results to be used for shaping policies on the rational allocation of resources for the effective restoration of well-being and development of rural areas. Proposed information technologies for geospatial analysis of rural community development. In particular, an interactive map of rural community development index values has been created, based on multicriteria decision-making (MCDM) and graph models, which enables the evaluation of the current state of rural infrastructure for each selected criterion. For visualization of the results, a geoportal has been implemented in the CREODIAS cloud environment.

Quantum computing is developing quickly but is confronted nowadays with significant technical, economic, and scientific challenges. This contribution summarizes the essential aspects of modern quantum informatics. It serves as a basis for understanding the developments and challenges in the field of quantum computers (QC), quantum communication, and quantum cryptography (QCCC, or Q3C). This subject area will become more relevant in the coming years. This work is intended to help understand the essential concepts, and challenges as well as the state of the art in this pioneering research field. Further advances in scalability, error correction, and software development are required in the mid-term to spread the technology and present its advantages compared to conventional computing clusters. And, although there are currently no suitable algorithms and working with quantum computers is like “shooting sparrows with the cannons”, sustainable practical applications for quantum computers are also to be established.

Staying abreast of the fast-paced development in connected devices and intelligent power networks has recently become a focus of scientific and engineering efforts. In that context, the Replica State Discovery Protocol (RSDP) lays the foundation for operational stability, sustainability, and reliability of clustered systems. RSDP serves as a platform for arbitrary logical extensions and execution of operations on a cluster-wide state. However, the initial model of RSDP interaction relies heavily on the Local Area Network Simulation (SLAN) based on AMQP as well as on frequent state broadcasts for mutual coordination of nodes. This approach is infeasible for networks with constricted resource capacities, such as IoT or Smart Grid. Therefore, the purpose of this article is to improve and augment the existing RSDP model with layered abstractions to separate modules related to communication, coordination, and application logic. The proposed solutions allow one to optimize decisions that cater to the needs of the execution environment. Additionally, this article introduces further separation of “State Reducer” unit into distinct abstract entities, expanding the logical flexibility of the protocol. To address the network congestion complexities that arise in IoT networks, a new RSDP-based interaction method is proposed that significantly reduces communication overhead related to cluster-wide state and operation broadcasts. Finally, this article proposes a new method for coordinating telemetry devices based on RSDP that facilitates granular load balancing, coordination of metric sharing, and operational execution of devices inside IoT.

Targeted at an inclusive society in the EU and Germany, an accessibility law came into effect in 2025 for all web agencies and developing companies. The work is devoted to the analysis of the opportunities for the creation of barrier-free or accessible content, as well as the features and abilities of AI (Artificial Intelligence) powered tools for so-called accessible websites. Such AI-powered tools include LLMs (Large Language Models) like ChatGPT products and corresponding APIs. An example workflow to enhance accessibility for blind and visually impaired individuals was constructed. In the experimental part, ChatGPT Vision APIs and OpenAI's Text-to-Speech (TTS) were used to explore and enhance accessibility for blind and visually impaired individuals. The contribution addresses the ongoing challenges to modern web apps to provide highly accessible content for inclusiveness.

Artificial intelligence (AI) has revolutionized medical research by using machine learning (ML), and deep learning (DL), to improve various aspects of healthcare. In this chapter, the profound impact of artificial intelligence on one area of ​​medical research is considered. The artificial intelligence system “PolliWeather” is created to track and forecast the allergenic danger and the spread of ragweed pollen allergens depending on the weather conditions and districts of Kyiv. This system can inform people about the level of allergen danger in different areas of the city, analyzing and predicting the risk of developing hay fever depending on the time of day, season, and districts of Kyiv. The study predicts a decrease in the risk of pollen-related allergies over the next three years due to global warming, which increases air humidity due to melting glaciers. However, residents of large built-up cities suffer more from pollen allergies due to the urban heat island effect, which leads to higher temperatures in the city. In this way, it is shown that people can take steps in advance to avoid health problems using the AI system “PolliWeather”.

The Internet of Things (IoT) is expanding rapidly, with billions of interconnected devices, ranging from basic sensors to sophisticated actuators, gathering, computing, and sending data. Such systems often provide valuable insights or help control the domain of operation, therefore playing well in the sustainable development paradigm. It is not uncommon for these devices to operate under strict constraints for power usage, bandwidth, and computational capacity. Given these limitations, such IoT systems must communicate reliably and ensure a given quality of service, remain energy efficient, and have the capacity to provide sustainability, even though these modern IoT ecosystems do not always function within an environment with a stable connection.This work examines key IoT communication protocols, focusing on their suitability for constrained environments and analyzing energy efficiency, scalability, security, and latency trade-offs. Special emphasis is placed on lightweight application-layer protocols such as MQTT, CoAP, and AMQP and evaluation of their performance under different conditions.Existing research highlights industry interest in multiprotocol interoperability. However, dynamic switching of application-layer protocols based on evolving network conditions remains an underexplored area. This work identifies key research gaps, particularly in thorough data transfer measurements, minding security overhead, quality of service (QoS) mechanisms, and real-time protocol adaptation on resource-limited static devices. The study concludes by proposing future directions for dynamic multiprotocol frameworks that enhance energy efficiency and reliability in large-scale IoT deployments.

In this chapter the historical review and current state of neural network development is presented . It is pointed out that the modern philosophical concept of linguistic neural networks, which has been in development in the last 60–70 years, is based on both ancient and current history of human knowledge. From a mathematical point of view, the concepts of single-layer and multilayer perceptron’s, corresponding schemes, and corresponding mathematical relations are discussed. Probabilistic models of linguistic neural networks are also considered. Namely, classic recurrent networks, networks with encoders and decoders for translation, networks with attentional mechanisms, and the model of modern Transfer Technology are considered. It is pointed out that modern models of neural networks are based on converting words into vectors and using vector and matrix operations. As examples on using word vectors for text encoding and decoding, context modeling algorithms and length estimation between the symbols are considered. A comparison of these two methods of text coding is also given. Novel approaches and standards in large language models’ neural networks are also considered. Several practical examples are given. Due to immense development, AI-driven apps based on LLMs can be deployed practically everywhere. These areas include healthcare, finances, industry, traffic and logistics, education, science, and customer services. As an example, for further deployment areas, programming and software technology have been considered.

The non-linear and non-stationary processes are frequently observed in economy, finances, production, ecology, especially in economy of transition to sustainable state. However, there is no unified methodology of constructing non-linear and non-stationary models as it has been done for linear processes. Thus, there is a need to develop a methodology for constructing non-linear and non-stationary processes models that would provide for the possibility to increase the model adequacy, and enhance quality of forecasts based upon them. The goal stated is reached using the following methods: system analysis-based methodology to data processing and model building; the method for a model structure estimation and taking into account possible uncertainties of data; the method of identification possible non-linearity and non-stationary features of data and their formal description. The procedure for non-linear on variables and non-stationary model constructing in economy, finances, ecology, production, environmental studies was developed. It can be applied for constructing models and estimation of fore-casts for the processes of transition economy from current state of to the sustain-ability state. The procedure provides for achieving required adequacy of the models constructed and high quality of forecasts based upon the models.The methodology developed for modeling non-linear and non-stationary processes is helpful from the point of view of refining the model constructing procedure, enhancing model adequacy and increasing quality of forecasts. The next step of the study is implementation of the methodology in the frames of specialized decision support system using system analysis principles.

A significant topic is the prediction and sustainability of social systems due to various factors. One of the most serious systems is the human body and the conditions of its existence. Considering this topic more deeply, we can assume that the human body’s resistance is defined as genetic expression changes under the influence of various factors. We offer an approach to analyzing the impact of the environment on genetic expression in humans and further on the resistance and probability of disease under the influence of these factors. It is based on a statistical approach to comparing and sorting known databases of the interaction of chemical compounds and genes.

This paper proposes an innovative model, Adaptive Intelligent Monitoring for Sustainable Development (AIM-SD), which combines modern communication technologies with machine learning (ML) methods to improve the effectiveness of monitoring environmental, social, and economic processes. A key element of this approach is the integration of adaptive algorithms capable of analyzing data in real-time. This allows for increased forecast accuracy, rapid response to changes, and reduced probability of forecast errors. Monitoring automation plays an important role in this process, reducing the impact of the human factor and increasing system efficiency. AIM-SD uses dynamic regression, Bayesian networks, and adaptive models such as recurrent neural networks (RNN) and long-term memory models (LSTM) to quickly respond to environmental changes and improve forecasts of short- and long-term changes. The problem that AIM-SD solves is associated with a high level of uncertainty and the need to quickly respond to changes in real-time. This involves the integration of data from different sources and allows the creation of a single data collection and processing system for more accurate and faster decision-making. Process automation improves the quality of monitoring and reduces the likelihood of errors, ensuring the efficiency of the system. The new AIM-SD model’s comparison with standard methods shows significant advantages: Its accuracy is 0.885, significantly higher than that of the standard model (0.466), which is based on simple linear regression with high noise levels. This basic approach is less effective in handling data variations, leading to lower accuracy. Additionally, AIM-SD demonstrates superior anomaly resilience (0.969 compared to 0.621 for the standard model), indicating better adaptation to data changes. Unlike the standard model, AIM-SD uses advanced adaptive smoothing techniques, uncertainty correction, and improved anomaly isolation, ensuring more robust responses to fluctuations in input data and a higher degree of accuracy and resilience. The data processing time for both models is almost the same, which confirms the effectiveness of AIM-SD in processing large amounts of information. AIM-SD consistently outperforms Isolation Forest and Bayesian Ridge in accuracy and anomaly resilience, demonstrating superior performance across all sample sizes, especially in noisy and unstable data. Thus, the proposed method, which combines adaptive models and automation of the monitoring process, demonstrates clear advantages over traditional methods, especially in terms of forecasting accuracy and resistance to changes.

In this chapter the problem of developing artificial intelligence standards in Ukraine is presented. The significance of the development of artificial intelligence in terms of sustainable development goals is considered. The importance of developing standards for artificial intelligence is emphasized. The state of approved standards for artificial intelligence in the USA and UK is investigated. Special attention is paid to developing artificial intelligence professional standards in Ukraine. The current state of development of these standards in Ukraine is described. This research will be useful for the developers of artificial intelligence professional standards in Ukraine.

This chapter proposes effective methods and algorithms for the interpolation and extrapolation of large numerical data sets using root-polynomial and root-fractional-rational functions. The corresponding analytical expressions for calculating the coefficients of root-fractional-rational functions are given. Also given are corresponding examples of using these functions to solve complex interpolation and extrapolation problems, including problems in electron optics, probability theory, and mathematical statistics, as well as descriptions of membership functions in fuzzy-logic problems. The provided studies have shown that modified analytical expressions for root-polynomial and root-fractional-rational functions, into which a positive deviation is introduced, can effectively be used to reduce the relative error of interpolation and extrapolation. The convergence of proposed interpolation and extrapolation methods in the case of using deviation is also guaranteed. The use of the proposed methods for interpolation and extrapolation for stiff function data sets is also described in this chapter. Therefore, further implementation of proposed methods in the developed computer software will significantly reduce the time for solving complex mathematical simulation problems both on local computers and in cloud and fog computing in local and global computer networks. In any case, the use of proposed methods and their development will provide significant impetus for further development and improvement of both local computer software and network software for cloud and fog computing.

This chapter examines the integration of decentralized blockchain-based systems to automate financial settlements in local electricity markets, aligning with sustainable energy goals. By leveraging blockchain technology, the proposed system ensures transparency in transactions, secure data storage, and rapid data processing. These features enhance resilience against cyber threats, corruption, and fraud while supporting energy efficiency and cost reduction. The decentralized approach fosters sustainable energy distribution by simplifying system access and promoting balanced electricity allocation between suppliers and consumers, contributing to the advancement of smart grids and sustainable infrastructure.

This work introduces a flexible and adaptive approach to URL classification, aiming to develop a tool capable of learning from diverse datasets to accurately identify malicious URLs. The scientific novelty of this approach lies in developing an ensemble of criteria that, with greater accuracy than previously proposed methods, enables the formulation of an optimal criterion for detecting phishing messages. The main idea is to design a system that remains effective across datasets from different sources, ensuring its robustness in real-world cybersecurity applications. To achieve this, a broad selection of modern classification algorithms was evaluated, allowing the identification of the three most efficient models. These top-performing classifiers were then combined into a VotingClassifier, leveraging ensemble learning to enhance predictive accuracy, reduce variance, and improve overall model stability. The study follows a supervised classification approach, where models are trained on labelled URL data (e.g., “malicious” or “benign”) to classify previously unseen URLs. The application of ensemble methods addresses common challenges in URL classification, such as data imbalance, noisy features, and varying dataset structures. By integrating multiple classifiers, the system compensates for individual model weaknesses, leading to more reliable predictions. The experimental results confirm that the strategic use of data mining algorithms, combined with rigorous data preprocessing and feature engineering, provides a powerful and scalable solution for detecting malicious URLs, effectively adapting to evolving cybersecurity threats, and serving as a valuable asset in network security and threat detection systems.

The rapid digitalization of the financial sector has triggered significant structural changes, reshaping economic processes and driving the adoption of cutting-edge technologies. This transformation has not only altered the way traditional financial institutions provide services but has also paved the way for the emergence of fintech companies as new market players. This article analyzes the current stage of fintech industry development within the framework of the digital economy. Key trends shaping the global financial technology sector are identified. The study examines the level of investment activity in the fintech sector, taking into account regional characteristics and industry specifics, and highlights the main factors influencing this process. The dynamics of financial technology development are assessed, and the primary barriers hindering further expansion are outlined. Based on the findings, strategic guidelines and principles necessary for the effective growth of the fintech industry are proposed. While possessing significant economic potential, the fintech industry must also incorporate the principles of sustainable development, including environmental, social, and corporate responsibility (ESG). In the process of implementing innovative technologies, it is crucial to ensure their compliance with sustainability standards, such as minimizing environmental impact, promoting social equity and transparency, and enhancing ethical business practices. The growing role of financial technologies can become a key driver in achieving global sustainable development goals, fostering more efficient resource utilization, and supporting socially responsible initiatives. Based on the research findings, strategic guidelines and principles essential for the effective development of the fintech industry are outlined.

The sustainability of the development of electron beam technologies for production of components of electronic industry and communication devices based on high-voltage glow discharge electron guns since the middle of the XIX century till today is considered in this chapter. The history of development of high-voltage glow-discharge electron guns, including the first experiments in the XIX century, is considered. Pointed out that today the main advantages of applying high voltage glow discharge electron guns for the production of components of the electronic industry are the simplicity of their construction, relative cheapness, possibility of operation with different gases, as well as the possibility of effective control of beam power. As technological processes, which can be successfully realized with the application of such electron guns, capsulation of electronics components by precision welding and deposition of ceramic coatings are considered. Basic schemes of electrode systems of such electron guns are given. Advanced methods of computer simulation of electric field distribution in the discharge gap and calculation of trajectories of ions and electrons are also considered. Pointed out is that sustainability of electron beam technologies based on high-voltage glow discharge electron guns gives the positive impulse to further development of the electronic industry, and, on the contrary, applying new kinds of digital and programmable electronic devices in control systems of electron beam installations is necessary for stable development of such technologies and their application in the production of modern communication electronic devices, including communication devices and devices for smart homes and IoT.

The work is devoted to the problems of collecting automated networking security information and critical event management for small and medium-sized enterprises (SMEs). So-called SIEM as advanced technology was examined. This technology’s potential for SMEs was analyzed and an outlook on further developments has been provided. Several case studies enable us to outline the advantages and disadvantages of the automated way compared to the manual process. Legal requirements for cybersecurity were examined.

The high volatility of renewable energy sources makes their widespread use more difficult. The work discusses how a safe, resilient, and sustainable energy supply to power grids can be guaranteed through the deployment of storage technologies. Sustainable and intelligent energy supply services and Smart Grid enable better achievement of SDGs (Sustainable Development Goals by UNO) which address global challenges, including better access to power grids and energy, education, health care, and IT communication services by the population.

This work aims to develop and investigate a mathematical model and method for designing a cyber-resilient and sustainable network. The proposed approach focuses on optimizing the use of available network equipment by considering cost, configuration, performance (throughput), security level, and the network’s ability to maintain stable operation in a dynamic environment. The developed solution offers an integrated, optimized framework for selecting network topology, equipment characteristics, communication links, and traffic distribution (routing). The design method is based on solving a Mixed-Integer Linear Programming (MILP) optimization problem and assumes that the potential locations for network routers are known in advance. A unified mathematical formulation enables a coordinated approach to selecting the network topology, determining the connection order between access networks and core routers, and specifying equipment characteristics. Cyber resilience is incorporated into the designed solutions by including security indicators directly in the objective function, namely, compromise probabilities, Common Vulnerability Scoring System values, and information security risks associated with network equipment. This approach enables synthesizing networks with specified or forecasted cyber resilience characteristics. A cyber-resilient and sustainable network was designed to validate the proposed method based on initial data, including different types of routers, interface modules, and network load volumes. The calculation results confirmed the adequacy of the mathematical model. As network performance requirements increased, the method dynamically involved additional network resources in the design solution. This process was accompanied by expanding the network topology and deploying higher-performing and more expensive routers and interface modules.

We investigate an optimal control problem for a linear parabolic equation with a non-local in time integral term, modeling heat and mass transfer processes that depend on the system's history. Such models are crucial in sustainable development, as they describe processes with memory effects in energy systems, environmental engineering, and material science.We introduce the concept of weak generalized solutions and establish the well-posedness of the initial-boundary value problem. We concentrate on point control case through a special operator appearing on the right-hand side of the equation. To address the optimal control problem, we consider a regularized formulation and prove the existence of solutions for both the original and regularized optimal control problems. Furthermore, we claim that as the regularization parameter approaches zero, the solutions of the regularized problem converge to the solutions of the original problem.For the regularized problem, we derive the Fréchet derivative of the quality functional, enabling the application of gradient-based numerical methods to compute the optimal control. Finally, we present numerical examples illustrating the efficiency of this approach in approximating optimal control solutions. Our results contribute to the theoretical and computational understanding of pointwise control in systems with memory effects, offering insights for sustainable technological advancements.

The problems of recommender systems (RS) improvement were considered in the context of big data (BD) processing in information and communication systems. The popularity of this topic is due to the need to develop high-performance personalization algorithms among the growing volumes of data used by e-commerce, education, healthcare, and live broadcast platforms. The challenges associated with the scalability and efficiency of big data processing require innovative approaches to ensure sustainable development. The paper proposes two modified versions of the Funk SVD algorithm. The first method is based on optimizing the latent factor matrix using gradient descent (GD), which reduces computation time. The second approach integrates additional functionality by normalizing the input data and applying gradient boosting (GB) to the regressors, which allows the computation of complex dependencies between users and objects. Both approaches aim to increase the speed and scalability of recommendation algorithms. Experimental results show that the first modification of Funk SVD exhibits a slight decrease in accuracy but works 22.9% faster than the standard method. The second modification takes into account complex data, reducing execution time by 50.38%. Experimental and graphical results confirm that the proposed algorithms significantly reduce computation time and maintain stable performance even on large data sets. In particular, the second modification has better scalability due to the normalization of functions and the use of gradient boosting. The developed methods are of important practical significance, as they provide a balance between accuracy, speed, and efficiency. This makes them suitable for scalable recommender systems, especially when speed and processing large amounts of data are critical, as in the case of streaming services and marketing platforms. By addressing the efficiency of resource utilization and enabling scalable solutions, the results contribute to the sustainable development of digital infrastructure across diverse sectors.

In this chapter, the ecosystem of artificial intelligence in Ukraine in terms of Sustainable Development is presented. An overview of the European Commission's ideas and actions to shape Europe's digital future and specific proposals for a European Strategy for Data and the White Paper on Artificial Intelligence are given. 17 Sustainable Development Goals are given. The authors emphasized that the availability of intellectual technologies and the intensity and effectiveness of their implementation become a criterion for developing national economies. Accordingly, the degree of AI implementation will depend on the attractiveness of countries and regions, the concentration of skilled labor, high-tech industries, material and financial resources, educational institutions, infrastructure, and cultural facilities. It is shown how much artificial intelligence is used in various sectors of the economy. The European AI Standards set out in the EU AI Act have become a global benchmark for regulating the development and implementation of AI technologies. They are crucial to making AI technologies more accountable at a level that directly affects the well-being of the population and the economy.The ecosystem of artificial intelligence in Ukraine, consisting of various organizations and technologies working together in artificial intelligence, is considered. The impact of artificial intelligence on various sectors of the economy in Ukraine is considered. The human and scientific potential for the development of AI in Ukraine is presented. The regulatory documents on artificial intelligence and the Strategy for Artificial Intelligence Development in Ukraine are given.

This study presents a dual approach for improving the performance of small-scale Smart Grid systems under varying operating conditions. First, a dynamic reconfiguration method is proposed to address partial shading in solar panel arrays. By employing a switching matrix and adaptive control signals, the array can be reconfigured in real time to mitigate mismatch losses and enhance energy capture. Simulation results in Matlab/Simulink show that, for a four-panel configuration, dynamic switching can yield a power increase of up to 6.33% at specific instances and an overall gain of 1.5% over the entire operating period. Second, a dynamic battery switching system is introduced to optimize the state of charge (SOC), depth of discharge (DOD), and maintenance schedules in parallel-connected battery modules. This approach ensures reliable load supply by matching discharge durations to the forecasted energy shortfall and balancing usage among individual batteries. Modeling demonstrates that incorporating Peukert?s effect into the battery selection process improves the accuracy of power delivery predictions while reducing the number of deep charge-discharge cycles that degrade battery health. Taken together, these findings highlight how dynamic topology switching for both solar panel arrays and battery banks can boost energy efficiency, extend component lifetimes, and provide a more resilient Smart Grid solution.

Integrating Artificial Intelligence (AI) with the Internet of Things (IoT) to so-called AIoT revolutionizes industries by enhancing automation, decision-making, and operational efficiency. This work explores the convergence of AI and IoT, highlighting key benefits such as predictive maintenance, real-time data analysis, improved cybersecurity, and optimized resource management. Despite these advantages, challenges persist, including data privacy risks, security vulnerabilities, and regulatory concerns. Practical applications are examined, including AI-driven autonomous vehicles, cybersecurity solutions, wearable health monitoring, and AI-enhanced imaging for medical diagnostics. The study also contrasts traditional IoT with AI-enabled IoT. This contribution provides insights into the future potential of AI-IoT integration and its transformative impact across sectors.

The fundamental architectural requirements for modern software application development are examined. It has been determined that enabling efficient delivery of dynamic data via Content Delivery Networks (CDNs), along with the capability to offload processing to edge locations, is closely tied to the adoption of microservice-based application architectures. A method for adaptive microservice creation at the edge location of СDN network to ensure effective dynamic data processing is presented. The contribution considers two ways of using the proposed adaptive microservice creation method. The first option involves the use of a centralized data store, which is located on the origin server side. The second one involves distributed data storage between CDN network locations. Several scenarios of the proposed method were simulated. The simulation was carried out on the created prototype of the content delivery network. The simulation results showed that even with the doubled number of requests, the latency for the end user decreased. The historical data of microservice activation on Edge locations can be leveraged to identify patterns in system behavior and enable predictive analytics aimed at forecasting periods of increased load. In addition, such a sustainable approach contributes to a more efficient use of resources, reduces their impact on the distributed environment, and increases the resilience of the system.

The integration of satellite (SAT)-based Geographic Information Systems (GIS) in agriculture is revolutionizing data-driven decision-making across the globe. Ukraine, with its vast arable lands and strategic agricultural position in Europe, stands to gain significantly from this technology. This paper explores the methodologies and applications of SAT-based GIS data processing in Ukraine's agricultural sector. It discusses the current state of data acquisition, processing workflows, analytical techniques, and how these inform crop monitoring, yield prediction, and precision agriculture practices. The integration of remote sensing technologies into precision agriculture has opened new possibilities for efficient crop management. This study explores the use of interpreted satellite imagery to monitor crop development and create variable-rate fertilization (VRF) task maps. Without relying on in-field sensors or ground truth data, vegetation indices derived from satellite images—particularly the Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI)—were used to assess crop health and vigor. Spatial analysis of temporal satellite data enabled the classification of field zones based on vegetative performance, which formed the basis for task map generation. The results demonstrate the feasibility and reliability of satellite-based remote sensing for optimizing fertilizer inputs, reducing environmental impact, and increasing yield efficiency. These technologies enable early assessment of crop productivity, allow tracking of the long-term dynamics of agroecological features of natural and artificial phytocenoses, both by individual indicators and in combination with other indicators, for example, meteorological and climatic ones.

The main steps of the proposed genetic algorithm (GA)-based method, which enables the determination of an optimal route considering dynamically changing operating conditions of sensor networks, are presented. To evaluate the efficiency of the proposed solution, sequential node addition and removal from the network topology were analyzed using the developed software. The simulation results demonstrated that the gradual removal of five nodes from a network topology of 25 nodes confirmed the feasibility of the proposed solution for establishing a data transmission route between two specific nodes, outperforming the greedy algorithm. The addition of new nodes using the GA-based method reduced the route length, highlighting its ability to effectively adapt routes in response to new nodes while optimizing overall distance. The greedy algorithm successfully formed routes in certain cases of node addition, but the route lengths were longer compared to the GA-based results. Thus, it was shown that the proposed method enhances the performance of sensor networks by ensuring data transmission under dynamically changing conditions, compared to the greedy algorithm. Additionally, it aligns with sustainable development principles aimed at creating adaptive information and communication systems.

In this work, we propose a system for the automated deployment of microservice applications in a cloud environment with support for SDN technology. A key element in the architecture of the proposed system is the Service-Oriented Management System, which implements methods for monitoring and managing computing and network resources, software services, and the process of providing services to users. In particular, in this contribution, we developed a method for deploying microservices based on their affinity, which allows making informed decisions on the placement, scaling, or migration of microservices following QoS and SLA requirements. According to the proposed method, the service-oriented management plane evaluates the relationships between services based on interaction logs, request frequency, and delays, forming an affinity matrix, identifying logically related groups of services, and providing clustering taking into account the real network context. The integration of the SDN controller into the management system provides dynamic statistics collection, continuous analysis of the network state, and automatic reconfiguration of network resources under the dynamic deployment of microservices to reduce delays and optimize traffic. The simulation results confirm the ability of the system to accurately identify the most related services and adaptively manage their placement and scaling, maintaining the required level of QoS and QoE. The proposed management system, using the method of deploying microservices based on their affinity, allows for moving from a static to an adaptive model of managing service-oriented systems, which takes into account dynamic network conditions and inter-service connections in real time. The proposed system is highly relevant for the effective management of distributed microservice systems in the conditions of modern Edge/Cloud infrastructures and meets the principles of sustainable development in the field of information and communication technologies.

The possibility of involving artificial intelligence systems in working with researchers in the framework of scientific research is discussed. For this purpose, three stages of artificial intelligence participation in solving these problems are identified. The first is the collection of materials and developments on the identified problem. The second stage of scientific research is the formation of a cycle of auxiliary tasks to clarify the connections between concepts and ideas, identifying qualitative and quantitative dependence. This will require an interactive mode of coordinating the formulations of these problems and clarifying the results. At the third stage, a synthesis of the obtained data and knowledge should be carried out to form a complete scientific theory. In addition to using language models, in particular, at the first stage of data collection, it becomes possible to involve Kolmogorov-Arnold networks to identify the dependence between variables in an analytical-symbolic form at the last stages of scientific research. The procedures for synthesizing an array of data and knowledge require clarification to orient artificial intelligence networks when comparing all knowledge obtained at the second stage of scientific research with known formed theories.

The United Nations General Assembly resolution of September 25, 2015, defined measures aimed at sustainable, harmonious development and ensuring stability and protection of state interests. The UN Sustainable Development Goals are consistent with national and border security issues. This prominent place belongs to optoelectronic surveillance systems in state border protection. These systems are key in modern border protection strategies, providing increased efficiency, security, and reliability. One of the technologies widely used in optoelectronic surveillance systems is image contour detection. The paper proposed and researched an improved approach in gradient methods for detecting contours of images in optical-electronic surveillance systems, which consists of matching convolutional kernels for detecting contours of objects. It is shown that the dynamic adaptation of parameters of convolutional kernels, based on local orientation and normalization, makes it possible to increase the accuracy of detecting contours of objects in images in a wide range of image quality. An algorithm for adapting classical gradient methods, particularly the Prewitt and Sobel methods, is proposed for dynamic convolutional kernels. In the example of test images, it is shown that such a modification of these methods makes it possible to improve the quality of contour detection and reduces the error in the detection of weak or blurred contours with a slight increase in computational costs. The presented results of the experiments demonstrate the effectiveness of the adapted methods in the actual application conditions, confirming the expediency of their use to increase the efficiency of image analysis in optical-electronic surveillance systems.

This chapter presents key innovations that significantly enhance the performance, efficiency, and sustainability of the AI (Artificial Intelligence) -powered bionic robot dog Wavego. We integrated long-range, low-power LoRa (Long Range) communication, enabling stable operation over 600 m. The proposed dual-controller architecture (ESP32 microcontroller for kinematics and an embedded Raspberry Pi computer for high-level AI tasks) optimizes resource utilization and system stability. To improve control, we developed both a web interface and a handheld remote equipped with a joystick and display, alongside real-time face recognition using computer vision. Experimental validation confirmed low latency (approximately 1.2 s), strong signal reception (up to –73 dBm, decibels relative to one milliwatt), and reliable operation in both indoor and outdoor environments. Furthermore, we implemented an intelligent RL (Reinforcement Learning)-based multichannel manager for dynamic selection among LoRa, Wi-Fi, BLE (Bluetooth Low Energy), LTE (Long-Term Evolution), GSM (Global System for Mobile Communications), ensuring adaptive, reliable and energy-efficient communication. Overall, the proposed approach advances sustainable and reliable robotic technologies aligned with the UN SDGs (Sustainable Development Goals).

Energy conservation in daily life is an important way to address climate change. This study is set to investigate how effectively a self-developed quiz app teaches sustainable energy use to young people. The app focuses on energy topics, including energy conservation and everyday applications. The effectiveness of the app was compared to a traditional learning method, specifically text-based learning. Seventh and eighth graders either used the app for 45 min or studied a text covering the same content. A third group received no prior instruction. One week later, all three groups took a knowledge test. Results showed that both learning groups outperformed the control group, with the text group slightly ahead of the app group. The study suggests that quiz apps have potential for increasing motivation, but further research is needed to optimize their effectiveness, to catch up with text-based learning or even reach better results.

The transition to sustainable energy systems requires innovative digital approaches to address growing challenges such as the instability of renewable energy sources (RES), unpredictable demand, high transmission losses, and insufficient grid flexibility. In this chapter, we propose a conceptual model based on Artificial Intelligence (AI), Internet of Things (IoT), and Blockchain technologies for smart energy management with a focus on improving forecasting accuracy, optimizing load balancing, and enabling decentralized, transparent energy transactions. The key contribution of this research is a thorough analysis of electricity production patterns using advanced data science techniques. Using a dataset covering more than 6.5 years of hourly electricity consumption and generation in Romania, we applied exploratory data analysis, time series decomposition, and correlation analysis to characterize the temporal behavior, variability, and relationships between multiple generation sources. In addition, we developed and evaluated deep learning models based on recurrent neural networks (RNNs) and long short-term memory (LSTM) architectures to forecast electricity supply and demand. The developed models achieved high prediction accuracy, demonstrating their effectiveness for real-time energy forecasting and intelligent load management in sustainable smart grid systems. The proposed concept aligns with the United Nations Sustainable Development Goals (SDGs), particularly SDG 7 (affordable and clean energy), SDG 9 (industry, innovation, and infrastructure), and SDG 13 (climate action). This concept is particularly relevant for countries with unstable energy conditions and transitional economies, such as Ukraine, where digital innovation and energy independence are strategic priorities. Moreover, it aligns with the broader objectives of European countries in striving to enhance grid flexibility, decarbonize energy systems, and accelerate the deployment of smart grids. By combining AI forecasting, real-time IoT data, and blockchain-based trust mechanisms, the study contributes to the development of intelligent, sustainable, and decentralized smart grid infrastructures.