KPI Science News

METHOD FOR REFINING WEIGHTS IN MULTI-CRITERIA UTILITY FUNCTION IN MAUT

Viktoriia Polutsyhanova, Serhii Smyrnov — Tue, 30 Sep 2025 00:00:00 +0300

Background. In modern multi-criteria decision-making, a critical challenge is the determination of weight coefficients in the utility function. Classical MAUT (Multi-Attribute Utility Theory) methods often rely on subjective expert evaluations, leading to potential errors due to expert fatigue and the limited number of comparisons. Additionally, discrepancies in the total weight sum can violate the axioms of linear convolution.

Objective. To develop a method for refining weight coefficients in the multi-attribute utility function of MAUT, which reduces the influence of subjectivity and ensures analytically consistent values.

Methods. An approach based on the Lagrange method applied to a system of normalized weights is proposed. This method transforms relative (non-normalized) expert assessments into precise weights by solving a system of equations analytically. To minimize errors, only relative weight ratios are used, reducing the number of expert queries from quadratic to linear complexity.

Results. A formula for refining weight coefficients is derived, preserving relative expert evaluations while ensuring accuracy and normalization. An example involving four criteria demonstrates the use of Lagrange multipliers to achieve refined weights with an error margin below 0.001. The method provides stable and analytically sound results without requiring complete pairwise comparisons.

Conclusions. The proposed method enables efficient refinement of weight coefficients in MAUT without overburdening experts. Analytical computation reduces error risks and enhances decision-making objectivity. The method is suitable for tasks with numerous criteria and offers a robust foundation for constructing utility functions in multi-criteria models.

ANALYSIS OF MATHEMATICAL MODELS AND METHODS FOR 3D ENVIRONMENT CHANGE DETECTION FROM IMAGERY

Oleg Chertov, Serhii Sakharov — Tue, 30 Sep 2025 00:00:00 +0300

Background. Automated detection of changes in three-dimensional (3D) environments constitutes a fundamental challenge in the development and maintenance of urban infrastructure digital twins, environmental monitoring systems, and facility security applications. Current methodologies exhibit limitations in uncertainty quantification, computational scalability for large-scale datasets, and semantic information integration. Existing reviews predominantly emphasize technical implementations or domain-specific applications, with insufficient attention to the particularities of the mathematical methods used. This gap impedes the principled selection of algorithms for digital twin actualization.

Objective. To provide a comprehensive systematization and comparative analysis of mathematical models and computational methods for three-dimensional environmental change detection. To establish a taxonomic framework that delineates methodological advantages, constraints, and applicability domains. To formulate evidence-based recommendations for method selection across diverse monitoring scenarios of 3D environment.

Methods. We conducted a systematic analysis encompassing classical geometric approaches for point cloud registration and comparison, statistical frameworks for uncertainty quantification, and contemporary machine learning paradigms for automated change classification. Formal mathematical representations were developed for three-dimensional data structures, change typologies, and corresponding evaluation metrics.

Results. A unified mathematical framework for characterizing diverse change phenomena in three-dimensional environments was established, providing a theoretical foundation for automated digital twin synchronization. Method selection criteria were derived based on data properties and application-specific requirements. Empirical evidence demonstrates that hybrid architectures integrating geometric primitives with machine learning techniques achieve superior accuracy while preserving interpretability in digital environment model updates.

Conclusions. The proposed systematic framework facilitates principled method selection for specific change detection applications. Future research directions include the development of self-adaptive algorithms with autonomous parameter optimization and the incorporation of semantic reasoning through advanced deep learning architectures.

A METHOD FOR ADJUSTING THE SOCIAL TENSION INDEX BASED ON METHODS FOR ASSESSING THE CONSISTENCY OF EXPERT ESTIMATES

Maksym Shchoholiev, Vitaliy Tsyganok — Tue, 30 Sep 2025 00:00:00 +0300

Problem. Social tension is the subject of numerous contemporary sociological studies due to its significant impact on the effectiveness of decision-making both at the governmental level and in business contexts. Social tension can be assessed by analyzing public reactions to news events in social networks. However, obtaining generalized estimates of the impact of news events on increasing levels of social tension is complicated by the fact that audiences of different information channels may hold significantly different opinions about the same event. The applicability of methods for assessing the consistency of expert estimates to problems that do not involve traditional expert input remains insufficiently studied.

Objective. To develop and test a method for adjusting the social tension index using methods for assessing the consistency of expert estimates.

Methodology of implementation. The implementation of the method for adjusting the social tension index involves the interaction between the knowledge base of a decision support system and a method for assessing the consistency of expert estimates, based on minimizing the total distance between all pairwise estimates.

Results. A method for adjusting the social tension index based on methods for assessing the consistency of expert estimates is proposed. A procedure for determining a threshold consistency level for evaluating the impact of news events on social tension is introduced. Applying the proposed method for adjusting the social tension index makes it possible to improve the validity of social tension assessments.

Conclusions. The application of the proposed method for adjusting the social tension index based on methods for assessing the consistency of expert estimates enhances the validity of social tension assessments in cases where the consistency index among estimates of news event impacts falls below the defined threshold.

STRUCTURAL-PHASE TRANSFORMATIONS AND MAGNETIC PROPERTIES OF NANOSCALE Pt/Co-BASED THIN FILM COMPOSITIONS DEPOSITED ON FLEXIBLE POLYMER SUBSTRATES

Tue, 30 Sep 2025 00:00:00 +0300

Background. Ferromagnetic CoPt-based nanofilms continue to attract attention due to their potential applications in spintronics and nanoelectronics. They combine unique magnetic properties, high corrosion resistance, and compatibility with modern technologies for magnetic information storage devices and electromagnetic radiation detection. Co–Pt films are studied and applied as materials for sensors, tunnel junctions, magnetic disks, THz generators, and skyrmion carriers. Investigating the formation of their properties is a key task for the development of nanoelectronic technologies.

Objective. To determine the changes in phase composition, depth distribution of chemical elements, and magnetic properties of Pt/Co-based nanofilm compositions deposited on flexible polyimide substrates and subjected to subsequent heat treatment in high vacuum, with the addition of Au interlayers and the formation of varying numbers of interfacial boundaries.

Methodology. Pt/Co/PI, Pt/Au/Co/PI, and Pt/Co/Au/Pt/Co/PI nanofilms were deposited on flexible polyimide substrates by magnetron sputtering at room temperature. After deposition, the samples were annealed in vacuum at 550 °C for 30 minutes. Structural and phase changes were studied using X-ray diffraction (XRD) and mass spectrometry, while magnetic properties were examined using vibrating sample magnetometry (VSM) in an external magnetic field.

Research Results. It has been established that annealing nanoscale compositions at a temperature of 550 °C for 30 minutes leads to the formation of very small crystallites, most likely associated with the relaxation of stresses that arise in the film compositions during deposition. Mass spectrometry results indicate that annealing results in a uniform distribution of the intensity of secondary ion emission from the constituent metals of the film compositions, as well as from the complex CoPt ion, which points to the homogenization of the film compositions’ structure as a consequence of annealing. Magnetic property studies have shown that annealing enables the formation of a magnetically hard state in the film material, while its coercivity demonstrates a pronounced dependence on the initial structure of the composition.

Conclusions. It has been established that thermal treatment of nanoscale compositions Pt(10 nm)/Co(10 nm), Pt(10 nm)/Au(4 nm)/Co(10 nm), and Pt(5 nm)/Co(5 nm)/Au(4 nm)/Pt(5 nm)/Co(5 nm) deposited on polyimide substrates at 550 °C for 30 minutes allows the achievement of a magnetically hard state of the film material. At the same time, its coercivity exhibits a pronounced dependence on the structure of the initial composition namely, the presence of an additional Au layer and the number of interlayer interfaces.

ANISOTROPY OF MICROSTRUCTURE AND PROPERTIES OF SLM ALLOY INCONEL 718

Tue, 30 Sep 2025 00:00:00 +0300

Background. The use of additive technologies, in particular selective laser melting (SLM), opens up new opportunities for the manufacture of parts with complex geometric shapes. However, the specifics of the 3D printing process cause a number of technological problems. Research into the mechanisms of microstructural evolution during SLM is necessary to optimize technological modes and ensure reliable and long-term operation of printed products in various industries.

Objective. Determination of microstructure features and microhardness distribution in different planes of the Inconel 718 SLM alloy, as well as comparative analysis of phase composition, texture, and macroscopic stress level.

Methods. A complex of techniques was applied - metallography, scanning electron microscopy, microdurometric and X-ray phase analysis.

Results. The anisotropy of the microstructure of the SLM alloy Inconel 718, which is formed during 3D printing as a result of sequential solidification of melt pools, has been identified and characterized. Significant variability of microhardness values in different planes has been shown as a consequence of the anisotropy of the structure and phase composition.

Conclusions. The amount of the main γ-phase and the strengthening γ′- and γ″-phases, as well as NbC carbide, is determined. Factors affecting the anisotropy of residual stresses and texture are discussed, including the direction of heat flow, cooling rate, reheating of layers, and the orientation of the grains formed.

SYSTEMATIZATION OF KEY STANDARDS FOR DIGITAL TWINS IN INDUSTRIAL MANUFACTURING

Oleksandr Pupena, Oleh Klymenko, Volodymyr Polupan — Tue, 30 Sep 2025 00:00:00 +0300

Background. Unified tools, approaches, and above all, a conceptual framework are needed to develop and scale digital twins. Given the young age of the standards, awareness of them is quite low, and their content is not always clear.

Objective.The purpose of the work is to obtain answers to questions about which standards directly relate to digital twins in the field of production, their purpose, content and application, and the possibility of their interoperability.

Methods. To achieve the stated objective, an analysis was conducted of publicly available scientific articles representing the perspectives of various research schools. The focus was placed on works addressing the issues of unification and standardization of digital twins. A review and systematization of the key provisions of existing standards were carried out, the classification of essential terms and structural elements was performed, and the scope of application of each standard in the context of industrial digital twins was determined.

Results. The research conducted a comprehensive analysis of key international standards dedicated to digital twins, including ISO/IEC 30173:2023, ISO 23247:2021 (all four parts), IEC 63278-1:2023, ISO/IEC TR 30172:2023, and ISO/TR 24464:2025. A comparison was made of the main terminology related to the digital twin, the target entity, and their interrelation; structural representations of digital twins; concepts of digital twin life cycles; the scope of digital twin applications; their relationship with other standards; and the feasibility of their joint application.

Conclusions. Although, from a formal point of view, there are no explicit contradictions between the ISO/IEC 30173, ISO 23247, and IEC 63278 standards, the analysis of their approaches to constructing digital twin frameworks revealed fundamental differences in their conceptual apparatus, digital twin structures, and the supporting standards they reference. This leads to the conclusion that their simultaneous use within a single digital twin life cycle is impossible without the implementation of an additional adaptation layer to ensure their compatibility