Vol 32, No 2 (2026)

An algorithm for finding optimal polymer synthesis process parameter values in the presence of multiple optimality criteria has been developed. The algorithm utilizes the method of moments and a genetic algorithm. Using the method of moments, the mathematical description of the polymerization process is transformed to a final form. The computational procedure for multi-criteria optimization is based on Pareto set approximation using a genetic algorithm. The proposed approach enables efficient processing of classes of process optimization problems whose mathematical models require preliminary analytical processing for the application of optimization methods. The algorithm’s operation is demonstrated using a multi-criteria problem for the industrially significant process of isoprene polymerization over a neodymium-containing catalyst system.

A hybrid approach for accelerating detailed routing of very large-scale integration (VLSI) circuits is proposed. The method combines a neural network model based on the U-Net architecture enhanced with Self-Attention and the classical Rip-Up and Reroute (R&R) algorithm. Experimental results demonstrate a significant acceleration of the routing process without loss of quality. The proposed solution illustrates the practical efficiency of machine learning methods in the field of physical design automation. The proposed approach represents the detailed routing task in a tensor form that preserves complete spatial information required for constructing routing paths. А modified deep learning segmentation model is developed to predict routing patterns for multiple nets simultaneously within a shared topological region. The predictions of the neural network serve as an initial approximation for the heuristic R&R algorithm, which substantially reduces the number of iterations needed to reach convergence. The neural network is trained on data derived from the results of global routing and physical design parameters extracted from LEF/DEF and Guide files. А new data decomposition method is introduced that allows the neural model to be adapted to any process design kit (PDK) by partitioning the routing layers into independent stacks. Tests on real integrated circuits show that the proposed method achieves up to a fivefold speedup compared to the open-source router OpenLane, particularly for large-scale designs. The study highlights the potential of deep learning in reducing the computational cost of detailed routing, one of the most time-consuming stages in VLSI physical synthesis. The approach demonstrates scalability, adaptability to different design rules, and opportunities for further performance gains through model optimization and integration into existing EDA workflows.

This study addresses the challenge of enhancing the structural secrecy of information transmission systems employing code-division multiplexing. To increase the structural secrecy of information transmission systems with coded channel separation, the necessity of increasing the number of used ensembles of orthogonal code sequences is proved. One of the possible ways to expand the number of orthogonal code sequences is the use of eigenvectors of Hermitian matrixes. Due to the complex component, multiphase orthogonal code sequences are obtained, the expansion of the number of which can be achieved by changing both the modules and arguments of the Hermitian matrix coefficients. However, not all multiphase orthogonal code sequences obtained from the eigenvectors of the Hermitian matrix will have the established requirements for correlation characteristics. The objectives of the paper are to determine the constraints for 4-th order Hermitian matrices at which ensembles of orthogonal multiphase code sequences described by their eigenvectors satisfy the first and second "condition of L. E. Varakin" by their correlation functions. The possible values of side peaks of correlation functions for different values of arguments of Hermitian matrix coefficients are calculated. Establishment of the ranges of changes in the side peaks of autocorrelation DR_j and cross-correlation DR_jk functions of ensembles of orthogonal multiphase code sequences have been established, which respectively lie within the limits from R_jmin = 0.25 to R_jmax = 0.75 and from R_jkmin = 0.5 to R_jkmax = 0.75.

An important task of materials science is to study the strength of complex objects used in structures and applied areas. These objects are created within the framework of additive manufacturing, where control at all stages is required. А computer-aided design system for additive manufacturing is proposed, and its functional diagram is presented. The features of multi-criteria optimization of the technological process are considered, and the results of applying the adaptive control algorithm with model adjustment are presented.

The article proposes modifications to the method for predicting abrupt changes in activity in software products with network effects, aimed at improving the accuracy and timeliness of detecting critical events. The focus is on adaptive parameters of the method, including a dynamic activation threshold for predictive signals dependent on current data volatility and automatic determination of the moving average window width based on local variability of activity metrics. А comparative analysis of the effectiveness of each modification was conducted using performance metrics defined in the article and visualization of results. It was found that the adaptive threshold reduces the proportion of false positives, while the algorithm for automatically selecting the moving average window width enables earlier detection of predictive signals. The results demonstrate that combining the proposed modifications ensures a balance between sensitivity and reliability of predictions, which is particularly important for social network monitoring systems and forecasting the risk of coordinated destructive actions by users.