Applied Problems of Computer Science, Security and Mathematics

Title

2025-09-03T05:52:29+00:00

COMPARISON OF REINFORCEMENT LEARNING IN GAME ENGINES UNREAL ENGINE 5 AND UNITY

2025-04-16T13:59:11+00:00

This paper presents a comparative analysis of reinforcement learning implementation in two leading game engines: Unreal Engine 5 and Unity. The study evaluates the integration simplicity, training effectiveness, and tool support of each platform by examining similar open-source projects where autonomous car agents learn to navigate racetracks using the Proximal Policy Optimization (PPO) algorithm. In Unreal Engine 5, training is achieved through the new Learning Agents plugin, leveraging Blueprint scripting and real-time physics-based simulation, while in Unity, the ML-Agents Toolkit is used with C# scripting and external Python-based training. The analysis includes environment setup, reward structures, agent perception systems, and debugging tools. Additionally, user experience and technical documentation quality are assessed. The results provide insights for developers and researchers aiming to choose an optimal platform for integrating RL into interactive simulations or games.

THE PRECISION FORMATION OF INFORMATIVE SIGNS RESEARCHING IN THE SYMBOLIC REPRESENTATION OF RASTER IMAGES FOR PATTERN RECOGNITION SYSTEMS

2025-05-25T18:13:15+00:00

This paper explores the application of symbolic image representation, as a method of visual data transformation. By converting raster images into symbolic forms using ASCII characters, this approach reduces data dimensionality while preserving key informational features. This makes the method particularly attractive for retaining informative attributes needed to form a reference alphabet in pattern recognition systems. The study assesses the effectiveness of symbolic representation in transmitting visual information by quantitatively evaluating the similarity between original and transformed images. Four metrics—Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), Feature Similarity Index (FSIM), and Signal-to-Reconstruction Error Ratio (SRE)—were employed for comprehensive analysis. A subset of the UTKFace dataset, consisting of 45 facial images, was used for the evaluation. Results indicate that while symbolic representations differ significantly from color images, they retain a high degree of informational similarity compared to binary and thinned images, confirming their potential for efficient visual data analysis and storage for pattern recognition systems.

MATHEMATICAL MODELING OF HEAVY METAL POLLUTION DISTRIBUTION RESULTING FROM EMISSIONS USING INTERLINEATION OF THREE-VARIABLE FUNCTIONS

2025-05-31T20:36:24+00:00

The article builds a mathematical model that describes the distribution of heavy metal contamination of soil due to industrial emissions. The input data are experimental data obtained by non-destructive testing methods. To do this, the paper constructs an interlineation operator for a function of three variables based on its known traces on a system of arbitrarily placed parallel lines. The article proves the interlineation properties of the constructed operator and theorems about the form of the error and its estimation.

PERFORMANCE OPTIMIZATION METHODS FOR SINGLE-PAGE APPLICATIONS USING THE REACT FRAMEWORK

2025-05-29T10:44:34+00:00

The paper explores methods for optimizing the performance of single-page applications (SPAs) built with React. It focuses on key techniques such as lazy loading, code splitting, memoization, and dynamic imports. These approaches aim to reduce the initial bundle size, minimize unnecessary re-rendering, and improve load times. The impact of each method on client-side speed and efficiency is analyzed using practical examples and basic performance metrics.

CLOUD MIGRATION AS A REENGINEERING STRATEGY AND ASSESSMENT OF COSTS AND RISKS

2025-06-05T07:39:29+00:00

This article explores cloud migration as a modern technological strategy for reengineering information systems, enabling the transformation of business processes by moving applications, data, and services to cloud environments. It highlights how this strategy provides increased flexibility, scalability, cost-effectiveness, and simplified infrastructure management. The relevance of the topic is justified by the rapidly growing popularity of cloud technologies in both private and public sectors. A structured approach to assessing costs and risks at different migration stages is proposed. Key cost categories are identified, including auditing, software adaptation, staff training, testing, and licensing. A generalized budget distribution table is provided. The article outlines the main risks associated with migration, including technical, organizational, financial, and security-related challenges. A classification method is introduced, categorizing risks by their sources and their potential impact on IT infrastructure stability. Examples from leading global companies that have adopted cloud solutions are discussed. The paper concludes with a summary of the advantages and drawbacks of cloud migration as a reengineering tool and presents suggestions for future research, including opportunities to automate risk and cost assessment using modern analytical tools. This work is intended for IT professionals, analysts, digital transformation consultants, and researchers focused on optimizing IT infrastructure.