Scientific Papers of Donetsk National Technical University. Series: “Computer Engineering and Automation"

RESEARCH ON THE POSSIBILITY OF REDUCING ENERGY CONSUMPTION BY CONTROLLING THE HEATING SYSTEM BASED ON THERMAL COMFORT

2025-05-30T11:55:32+03:00

Against the backdrop of simultaneous changes in the impact of several negative factors on the energy market, as well as energy prices in the European Union, one of the least expensive ways to increase the energy efficiency of a building’s heating system is to use the latest approaches to control. A promising direction in this matter is to replace traditional indoor temperature control with thermal comfort control using a complex PMV indicator. To conduct a comparative analysis of energy consumption for heating using a traditional approach and using a complex thermal comfort indicator based on an improved room model. To develop a model to study the possibility of reducing energy consumption by controlling the heating system based on thermal comfort. The resulting model allows you to study the process of thermal comfort control. After conducting several experiments, it was found that the thermal comfort control system in the room during operation affects the amount of energy resources consumed, unlike the traditional one. Thus, indoor thermal comfort control can provide energy savings by using the complex PMV indicator as the actual feedback value instead of temperature. The proposed temperature control system, unlike the traditional one, does not have constant energy consumption in the room. Considering the change in thermal comfort parameters, which in some cases leads to the effective use of energy resources and the resources of the system itself. At the same time, under the same conditions, the indoor thermal comfort control system is in standby mode without turning on the heating, which ensures a reduction in energy consumption. Thus, this system, in the absence of restrictions on energy consumption, provides better heating compared to the traditional one.

ANALYSIS SYSTEM OF UKRAINIAN CONSUMER PREFERENCES WITH ARTIFICIAL INTELLIGENCE ELEMENTS AND AN INTEGRATED PERSONAL ASSISTANT

2024-11-29T11:27:26+02:00

The article discusses a system for analyzing the preferences of Ukrainian consumers, which uses elements of artificial intelligence and an integrated personal assistant to improve the personalization process in e-commerce. The main functional and non-functional requirements for such a system are defined, ensuring it provides users with an intuitive interface, fast processing of requests, and effective product recommendation personalization. An important feature of the system is the use of the GPT-4o-mini language model to create an integrated personal assistant capable of recognizing user needs and suggesting relevant products based on their preferences and previous purchases.

The development process of the system included two main stages: creating a data analysis system for buyers and developing a personal assistant. A dataset was used for analyzing consumer preferences, including parameters such as age, gender, product category, purchase quantity, unit price, payment method, and other characteristics. Using this data, several graphs were built to reflect consumer behavior trends, which allows for more accurate predictions of future purchases.

At the same time, the system was equipped with the ability to predict future purchases based on neural networks, enabling the prediction of upcoming purchases using machine learning algorithms. The development of this part of the system ensured prediction accuracy and minimized losses during the model training phase. The created neural network showed effectiveness in predicting user behavior, contributing to an enhanced level of personalization and ease of shopping for each user.

The article also analyzes existing solutions in the field of artificial intelligence and personalized recommendation systems used in e-commerce. This analysis helps to identify the strengths and weaknesses of such systems and find ways to improve them for implementation in the Ukrainian market. Furthermore, the proposed solutions combine advanced technologies and user-friendliness, ensuring the system’s successful adaptation to the needs of Ukrainian consumers.

RESEARCH INTO THE PROCESS OF MODELING RANDOM DELAYS IN NETWORKED CONTROL SYSTEMS

2025-05-30T12:52:32+03:00

This paper proposes to consider the main types of delay models that arise and, importantly, negatively affect the efficiency and productivity of networked control systems, and in some cases lead to failures and even failures in operation. As for all other types of control systems, for networked control systems, stability of operation, accuracy and speed of execution of processes and algorithms, high security indicators, availability and flexibility and a number of other requirements are also very important. And all this is achieved through constant study and improvement of technologies, search for new and non-standard solutions and also constant research in the industry. As we already know, networked control systems (NCS) based on even the name assume the presence of networks in their control circuits, which in turn leads to the emergence of a number of problems that are characteristic of network technologies. Such problems as, for example, random delays or packet loss, or transmission of several packets at a time. It should be noted that usually the most attention when studying this issue is paid to the problems caused by random delays, because if they are present, they can also lead to the emergence of other problems that were mentioned above. Therefore, to reduce the negative impact of these random delays, which lead to a decrease in the efficiency of networked control systems and instability in their operation, there is a need to establish a mathematical model of random delays before creating a compensating element. And modeling of networked control systems with the right approach and the choice of successful solutions and modeling methods makes it possible to solve the issue of the presence of imperfections and problems associated with this technology.

AUTOMATED XSS VULNERABILITY DETECTION IN WEB APPLICATIONS BASED ON A MULTI-AGENT APPROACH

2025-03-08T10:42:44+02:00

With the development of information technologies and the increasing volume of sensitive data processing on the Internet, web applications have become a crucial part of modern business processes. However, with this growth, the number of cyber threats also increases, posing the challenge for organizations to ensure the security of their web resources. One of the primary methods of protection is penetration testing, which allows vulnerabilities to be identified by simulating real attacks. This article explores the use of multi-agent systems (MAS) for automating the penetration testing process, specifically for detecting XSS vulnerabilities. Penetration testing is a vital step in ensuring the security of web applications, encompassing several stages: initial analysis, vulnerability identification, exploitation of vulnerabilities, and evaluating the consequences of an attack. Common vulnerabilities, such as XSS, are key targets in penetration testing because they can be easily exploited to compromise a system. However, traditional penetration testing methods often have limited ability to adapt quickly to new attacks, making them less effective.