ВИКОРИСТАННЯ СУЧАСНИХ ІНФОРМАЦІЙНИХ ТЕХНОЛОГІЙ В СИСТЕМАХ АВТОМАТИЗОВАНОГО РОЗВ’ЯЗАННЯ СИНГУЛЯРНО ЗБУРЕНИХ ЗАДАЧ

The paper presents the architecture of a software system for automated solution of singularly perturbed problems of semiconductor electronics, developed based on the principles of domain-driven design (DDD). The DDD technology was applied to the mathematical model of the technical system "plasma diode" to demonstrate its potential in the context of complex engineering problems. The range of software design problems was expanded using DDD methods, which made it possible to model complex systems and processes more accurately. A special feature of the presented architecture of the software system is the integration of an artificial neural network designed to recognize the types of recurrent sequences of subproblems. These subproblems are formed as a result of decomposition of the basic problem using the method of boundary corrections of perturbation theory. The neural network analyzes the obtained subproblems and classifies them by known types, which allows automating the selection of the most effective methods for their solution and optimizing computing resources. The use of DDD in developing a software system for solving singularly perturbed problems allowed us to deeply understand the subject area and reflect its concepts in the program code. This contributes to creating flexible and scalable systems that can easily adapt to changing requirements and operating conditions. The "plasma diode" mathematical model served as a case study confirming the approach's effectiveness. The developed system opens up new prospects for automating and optimizing modeling and analysis processes in semi- conductor electronics. The use of DDD and the integration of intelligent components, such as artificial neural networks, contribute to the expansion of software capabilities and increased efficiency. Further system development involves expanding the database of subtasks and methods for solving them, improving neural network training algorithms, and adapting the architecture to other technical systems and problems. This will contribute to a wider application of the developed approach in various fields of science and technology.