Articles in "Computer Science"

Summary: In this paper, a model of single server queueing-inventory system (QIS) with Markovian Arrival Process (MAP) and phase-type distribution (PH-distribution) of the service time of consumer customers (????-customers) is considered. After completing the service of ????-customer, he (she) can make one of the following decisions: (1) eventually leave the system with probability (w.p.) ????ℓ; (2) after a random “thinking” time returns the purchased item w.p. ????????; (3) after a random “thinking” he (she) feedback to buy a new item w.p. ???????? . It is assumed that ????ℓ+????????+???????? = 1. If upon arrival of the ????-customer the system main warehouse (SMW) is empty, then the incoming customer, according to the Bernoulli scheme, is either joins the infinite queue or leaves the system. A virtual finite orbit can be considered as a waiting room for feedback customers (????-customers). Returned items are considered new and are sent directly to SMW if there is at least one free space; otherwise, this item is sent to a special warehouse for returned items (WRI). After completing the service of each customer, one item is instantly sent from the WRI (if any) to the SMW. In SMW, the (????, ????) replenishment policy is used and it is assumed that the lead time follows exponential distribution with finite parameter. When the stock level reaches its maximum value due to items returns, the system immediately cancels the regular order. Along with classical performance measures of QIS new specific measures are defined and numerical method for their calculation as well as maximization of the revenue function are developed. Results of numerical examples to illustrate the effect of different parameters on the system’s performance measures are provided and analyzed. We also provide a detailed analysis of an important special case of the Poisson process/exponential service time model.

Summary: Invisible watermarking has become a vital technique in digital image security, allowing hidden data to be embedded without affecting visual quality. This study addresses the challenge of maintaining watermark imperceptibility while resisting common image distortions. The goal of this research is to assess the effectiveness of a hybrid watermarking method that combines Discrete Cosine Transform (DCT) and Discrete Wavelet Transform (DWT). In the proposed approach, the watermark is embedded in the DCT coefficients of the low-frequ-ency DWT sub-band. Experiments were conducted in both RGB and YCbCr color spaces, with varying strength factors (alpha). The method demonstrated high imperceptibility, with SSIM values above 0.998, but limited robustness against JPEG compression, Gaussian blur, and noise. These findings highlight the need for more resilient hybrid methods

Summary: The importance of forming and bringing the organization of databases of medical institution branches into a more suitable form in terms of their structure, accessibility and flexibility of applications, and the development of methods in this regard are presented in the article. Thus, having each branch's own database increases the flexibility of local queries, minimizes duplication of general data, and provides a structure for distributing certain data in a way that is specific to the functions and medical aspects of the branches. Many existing methods for distributed databases have been considered and the suitability of some of them for medical documents has been noted.

Summary: The thesis discusses interactive technologies for creating easy-to-understand programs by minimizing latency. What interactive techno-logy is and where it is used have been investigated. Which technologies can be interactive, how can we create interactive programs with basic elements, the design of interactive applications, its negative and positive aspects, and such issues are analyzed. Its primary benefit in the manage-ment field is explained, and its applications to increase efficacy in various sectors are also discussed. In this thesis, we can see the impor-tance of the interactive approach and its positive impact on other areas. At the same time, we will witness that future life will be mostly interac-tive, and people will reach their goals more easily thanks to applications. It is emphasized that the main starting point for the interaction of developed technologies is good design, and what are the necessary conditions for this design.

Summary: This paper considers the issue of assessing the security of a key exchange protocol based on matrix algebra. To protect the confidentiality of information, it is important that the keys are exchanged securely. This can be done by using methods such as encrypted key transmission, creation of dedicated secure channels, formation of a common key between parties without exchanging secret information using public key algorithms, etc. The paper analyzes the security of the protocol implemented using the approach based on non-invertible matrices against linear algebra and brute force attacks, shows that the use of non-invertible matrices increases the security of the system, and also evaluates the consumption of computing resources. The results confirm that this approach can be used as a secure and practical method of key exchange.

Summary: On the basis of machine-building industry area analysis there were defined actuality and a goal of this material. As the applied object of this study, flexible manufacturing system is chosen and also it was given a task of designing an algorithm for investigation of its control system. When creating a control model for a flexible manufacturing system a finite automata was used to describe the operation of a crane manipulator. The capabilities of the finite automata are not enough for a general system analysis of the control object. For a detailed functional analysis with the definition of transfers and productions of this control object, it is necessary to use a more powerful mathematical apparatus, such as Petri net. The study of the control algorithm was held on the basis of the round composed flexible manufacturing module scheme of flexible manufactu-ring system. The resulting algorithm with using the transformation of a finite automata of Petri net allows for the creation of a more correct, efficient and reliable control system for the production module.

Summary: Artificial Intelligence (AI) models are increasingly pivotal in enabling face recognition across various fields, from educational and research settings to public spaces. Effective deployment of these models requires high-performance hardware, such as RTX graphics cards or embedded edge devices like Nvidia's AGX Orin and Jetson Nano. This paper pre-sents a comprehensive benchmarking study comparing the performance of these two devices, representing high and low-power edge computing options, using two face recognition models: ResNet and MobileNet. The benchmarking process assesses each model across two different input sizes deployed on both devices with varied configurations, inclu-ding CPU thread allocation and GPU power distribution within contai-nerized environments. Performance metrics such as inference time, GPU utilization, memory usage, and CPU load are analyzed to determine each device's suitability and efficiency. Additionally, model-specific parameters, including FLOPS, parameter count, and memory footprint, are examined to provide for an in-depth comparison. This paper pre-sents detailed results and analyses of these performance indicators.

Summary: This article presents a comprehensive study of algorithmic bargaining within dynamic pricing systems, focusing on its usability, challenges, and broader market implications. Algorithmic bargaining, increasingly used in industries such as e-commerce, retail, and transportation, allows companies to dynamically adjust prices based on real-time data, consumer behavior, and market conditions. However, its widespread adoption raises significant questions regarding fairness, transparency, and regulatory oversight. Drawing on studies from fields like game theory, behavioral economics, and data science, this research explores the impacts of algorithmic bargaining on both businesses and consumers. While machine learning and deep reinforcement learning technologies enhance pricing efficiency, they also present risks of consumer harm and market manipulation. This paper critically examines the ethical implications, regulatory responses, and potential consequences for competition and consumer welfare. Through an analysis of existing literature, including case studies from various industries, this study provides a balanced evaluation of the benefits and drawbacks of algorithmic bargaining, encouraging a deeper understanding of its multifaceted role in modern pricing strategies.

Summary: A mathematical model of a queuing-inventory system (QIS) with catastrophes is built. Incoming customers form a Poisson flow with rate λ. The customer servicing time in the considered QIS is zero. The (S,Q) replenishment policy is used to increase the inventory level in the system. Here, S is the maximum storage size of the QIS, and Q=S-s indicates the fixed size of the proposed order.

Summary: Pneumonia is a prevalent illness that has a global impact, primarily affecting children and older individuals. Timely identification is essential for immediate intervention, particularly in regions with restricted healthcare availability. The purpose of this paper is to compare how well two methods work for finding pneumonia on chest X-rays creating a custom Convolutional Neural Network (CNN) and using a model that has been pre-trained, such as ResNet-50. The results indicated that although the customized CNN had difficulties achieving satisfactory performance, the ResNet-50 model demonstrated encouraging outcomes following the process of fine-tuning. This paper seeks to improve the detection of pneumonia, especially in disadvantaged places with limited medical resources, by utilizing modern technologies such as deep learning and pre-trained models. Considering the results gained to enhance patient outcomes and decrease mortality rates associated with pneumonia by enabling more precise and prompt detection, this has a beneficial effect on global public health