Implementation of the reconstruction process for constructive-synthesizing models of fractal time series

• Опубліковано в: PROBLEMS IN PROGRAMMING
• Дата: 2026
• Випуск: 4
• Сторінки: 3-11
• ISSN: 1727-4907
• Автори та афіліації: V.I. Shynkarenko — Ukrainian State University of Science and Technology; A.A. Zhadan — Ukrainian State University of Science and Technology
• Ключові слова: фрактальні часові ряди, генетична інженерія, конструктивно-продукційне моделювання, генетичний алгоритм, програмна інженерія, фрактальний, інженерія програмного забезпечення, фрактальний аналіз, програмні технології, квантовий генетичний алгоритм
• Автори: Shynkarenko, V.I., Zhadan, A.A.
• Формат: Стаття
• Мова: Українська
• Опубліковано: PROBLEMS IN PROGRAMMING 2026
• Теми: software; constrictive-synthesizing modeling; information technologies; fractals; fractal time series; genetic algorithm; L-system; cloud computing; UDC 004.4
• Онлайн доступ: https://pp.isofts.kiev.ua/index.php/ojs1/article/view/871
• Назва журналу: Problems in programming

The presented method focuses on reconstructing constructive models for predefined fractal time series. This work generalizes the approach to implementing the software system that ensures effective automation while considering the specifics of working with model series of various natures– both deterministic and stochastic. Two main approaches to system implementation based on a genetic algorithm were analyzed: the monolithic and the multi-agent. Considering the complexity of calculating chromosome viability indicators when working with stochastic series, it was decided to separate certain elements of the genetic algorithm– specifically, crossover and mutation– from the selection stage by introducing subpopulations. This made it possible to perform distrib uted computation of chromosome fitness indicators. The introduced solution enabled independent scaling of various elements of the reconstruction process, which increased the overall efficiency of the system. To ensure consistent interaction between elements, different types of communication were considered, among which the asynchronous approach proved to be the most effective. Mechanisms were implemented to optimize interaction between computational entities, as well as a node pattern for implementing crossover and mutation operations. This approach made it possible to eliminate problems associated with processing stochastic time series and en sure controlled and efficient horizontal scaling of the process of reconstructing constructive models.Problems in programming 2025; 4: 3-11