All issues

Large Language Models (LLMs) are transforming software engineering by bridging the gap between natural language and programming languages. These models have revolutionized communication within development teams and the Software Development Life Cycle (SDLC) by enabling developers to interact with code using natural language, thereby improving workflow efficiency. This survey examines the impact of LLMs across various stages of the SDLC, including requirement gathering, system design, coding, debugging, testing, and documentation. LLMs have proven to be particularly useful in automating repetitive tasks such as code generation, refactoring, and bug detection, thus reducing manual effort and accelerating the development process. The integration of LLMs into the development process offers several advantages, including the automation of error correction, enhanced collaboration, and the ability to generate high-quality, functional code based on natural language input. Additionally, LLMs assist developers in understanding and implementing complex software requirements and design patterns. This paper also discusses the evolution of LLMs from simple code completion tools to sophisticated models capable of performing high-level software engineering tasks. However, despite their benefits, there are challenges associated with LLM adoption, such as issues related to model accuracy, interpretability, and potential biases. These limitations must be addressed to ensure the reliable deployment of LLMs in production environments. The paper concludes by identifying key areas for future research, including improving the adaptability of LLMs to specific software domains, enhancing their contextual understanding, and refining their capabilities to generate semantically accurate and efficient code. This survey provides valuable insights into the evolving role of LLMs in software engineering, offering a foundation for further exploration and practical implementation.

Modern power transportation systems are the complex engineering systems. Such systems include both point facilities (power producers, consumers, transformer substations, etc.) and the distributed elements (f.e. power lines). Such structures are presented in the form of the graphs with different types of nodes under creating the mathematical models. It is necessary to solve the system of partial differential equations of the hyperbolic type to study the dynamic effects in such systems.

An approach similar to one already applied in modeling similar problems earlier used in the work. New variant of the splitting method was used proposed by the authors. Unlike most known works, the splitting is not carried out according to physical processes (energy transport without dissipation, separately dissipative processes). We used splitting to the transport equations with the drain and the exchange between Reimann’s invariants. This splitting makes possible to construct the hybrid schemes for Riemann invariants with a high order of approximation and minimal dissipation error. An example of constructing such a hybrid differential scheme is described for a single-phase power line. The difference scheme proposed is based on the analysis of the properties of the schemes in the space of insufficient coefficients.

Examples of the model problem numerical solutions using the proposed splitting and the difference scheme are given. The results of the numerical calculations shows that the difference scheme allows to reproduce the arising regions of large gradients. It is shown that the difference schemes also allow detecting resonances in such the systems.

Data hiding in digital images is a promising direction of cybersecurity. Digital steganography methods provide imperceptible transmission of secret data over an open communication channel. The information embedding efficiency depends on the embedding imperceptibility, capacity, and robustness. These quality criteria are mutually inverse, and the improvement of one indicator usually leads to the deterioration of the others. A balance between them can be achieved using metaheuristic optimization. Metaheuristics are a class of optimization algorithms that find an optimal, or close to an optimal solution for a variety of problems, including those that are difficult to formalize, by simulating various natural processes, for example, the evolution of species or the behavior of animals. In this study, we propose an approach to data hiding in the hybrid spatial-frequency domain of digital images based on metaheuristic optimization. Changing a block of image pixels according to some change matrix is considered as an embedding operation. We select the change matrix adaptively for each block using metaheuristic optimization algorithms. In this study, we compare the performance of three metaheuristics such as genetic algorithm, particle swarm optimization, and differential evolution to find the best change matrix. Experimental results showed that the proposed approach provides high imperceptibility of embedding, high capacity, and error-free extraction of embedded information. At the same time, storage of change matrices for each block is not required for further data extraction. This improves user experience and reduces the chance of an attacker discovering the steganographic attachment. Metaheuristics provided an increase in imperceptibility indicator, estimated by the PSNR metric, and the capacity of the previous algorithm for embedding information into the coefficients of the discrete cosine transform using the QIM method [Evsutin, Melman, Meshcheryakov, 2021] by 26.02% and 30.18%, respectively, for the genetic algorithm, 26.01% and 19.39% for particle swarm optimization, 27.30% and 28.73% for differential evolution.

The method for mapping of intermediate representations (IR) set of C, C++ programs to vector embedding space is considered to create an empirical estimation framework for static performance prediction using LLVM compiler infrastructure. The usage of embeddings makes programs easier to compare due to avoiding Control Flow Graphs (CFG) and Data Flow Graphs (DFG) direct comparison. This method is based on transformation series of the initial IR such as: instrumentation — injection of artificial instructions in an instrumentation compiler’s pass depending on load offset delta in the current instruction compared to the previous one, mapping of instrumented IR into multidimensional vector with IR2Vec and dimension reduction with t-SNE (t-distributed stochastic neighbor embedding) method. The D1 cache miss ratio measured with perf stat tool is considered as performance metric. A heuristic criterion of programs having more or less cache miss ratio is given. This criterion is based on embeddings of programs in 2D-space. The instrumentation compiler’s pass developed in this work is described: how it generates and injects artificial instructions into IR within the used memory model. The software pipeline that implements the performance estimation based on LLVM compiler infrastructure is given. Computational experiments are performed on synthetic tests which are the sets of programs with the same CFGs but with different sequences of offsets used when accessing the one-dimensional array of a given size. The correlation coefficient between performance metric and distance to the worst program’s embedding is measured and proved to be negative regardless of t-SNE initialization. This fact proves the heuristic criterion to be true. The process of such synthetic tests generation is also considered. Moreover, the variety of performance metric in programs set in such a test is proposed as a metric to be improved with exploration of more tests generators.

Mechanical properties of eukaryotic cells play an important role in life cycle conditions and in the development of pathological processes. In this paper we discuss the problem of parameters identification and verification of viscoelastic constitutive models based on force spectroscopy data of living cells. It is proposed to use one-dimensional continuous wavelet transform to calculate the relaxation function. Analytical calculations and the results of numerical simulation are given, which allow to obtain relaxation functions similar to each other on the basis of experimentally determined force curves and theoretical stress-strain relationships using wavelet differentiation algorithms. Test examples demonstrating correctness of software implementation of the proposed algorithms are analyzed. The cell models are considered, on the example of which the application of the proposed procedure of identification and verification of their parameters is demonstrated. Among them are a structural-mechanical model with parallel connected fractional elements, which is currently the most adequate in terms of compliance with atomic force microscopy data of a wide class of cells, and a new statistical-thermodynamic model, which is not inferior in descriptive capabilities to models with fractional derivatives, but has a clearer physical meaning. For the statistical-thermodynamic model, the procedure of its construction is described in detail, which includes the following. Introduction of a structural variable, the order parameter, to describe the orientation properties of the cell cytoskeleton. Setting and solving the statistical problem for the ensemble of actin filaments of a representative cell volume with respect to this variable. Establishment of the type of free energy depending on the order parameter, temperature and external load. It is also proposed to use an oriented-viscous-elastic body as a model of a representative element of the cell. Following the theory of linear thermodynamics, evolutionary equations describing the mechanical behavior of the representative volume of the cell are obtained, which satisfy the basic thermodynamic laws. The problem of optimizing the parameters of the statisticalthermodynamic model of the cell, which can be compared both with experimental data and with the results of simulations based on other mathematical models, is also posed and solved. The viscoelastic characteristics of cells are determined on the basis of comparison with literature data.

An approach to computation of open-loop optimal Nash–Cournot strategies in dynamical games which is based on the Z-transform method and factorization is proposed. The main advantage of the proposed approach is that it permits to overcome the problems of instability of economic indicators of oligopolies arising when generalized Riccati equations are used.

A new approach to the identification of hardly perceptible diagnostically significant changes in electrocardiograms is suggested. The approach is based on the analysis of topological transformations in wavelet spectra associated with electrocardiograms. Possible practical application of the approach developed is discussed.

When making decisions in various fields of human activity, it is often required to create text documents. Traditionally, the study of texts is engaged in linguistics, which in a broad sense can be understood as a part of semiotics — the science of signs and sign systems, while semiotic objects are of different types. The method of rank distributions is widely used for the quantitative study of sign systems. Rank distribution is a set of item names sorted in descending order by frequency of occurrence. For frequency-rank distributions, researchers often use the term «power-law distributions».

In this paper, the rank distribution method is used to analyze the Criminal Code of various countries. The general idea of the approach to solving this problem is to consider the code as a text document, in which the sign is the measure of punishment for certain crimes. The document is presented as a list of occurrences of a specific word (character) and its derivatives (word forms). The combination of all these signs characters forms a punishment dictionary, for which the occurrence frequency of each punishment in the code text is calculated. This allows us to transform the constructed dictionary into a frequency dictionary of punishments and conduct its further research using the V. P. Maslov approach, proposed to analyze the linguistics problems. This approach introduces the concept of the virtual frequency of crime occurrence, which is an assessment measure of the real harm to society and the consequences of the crime committed in various spheres of human life. On this path, the paper proposes a parametrization of the rank distribution to analyze the punishment dictionary of the Special Part of the Criminal Code of the Russian Federation concerning punishments for economic crimes. Various versions of the code are considered, and the constructed model was shown to reflect objectively undertaken over time by legislators its changes for the better. For the Criminal Codes in force in the Federal Republic of Germany and the People’s Republic of China, the texts including similar offenses and analogous to the Russian special section of the Special Part were studied. The rank distributions obtained in the article for the corresponding frequency dictionaries of codes coincide with those obtained by V. P. Maslov’s law, which essentially clarifies Zipf’s law. This allows us to conclude both the good text organization and the adequacy of the selected punishments for crimes.

Modeling of carbon nanostructures by means of classical molecular dynamics requires a lot of computations. One of the ways to improve the performance of basic algorithms is to transform them for running on SIMD-type computing systems such as systems with dedicated GPU. In this work we describe the development of algorithms for computation of many-body interaction based on Tersoff and embedded-atom potentials by means of OpenCL technology. OpenCL standard provides universality and portability of the algorithms and can be successfully used for development of the software for heterogeneous computing systems. The performance of algorithms is evaluated on CPU and GPU hardware platforms. It is shown that concurrent memory writes is effective for Tersoff bond order potential. The same approach for embedded-atom potential is shown to be slower than algorithm without concurrent memory access. Performance evaluation shows a significant GPU acceleration of energy-force evaluation algorithms for many-body potentials in comparison to the corresponding serial implementations.