Development of a software architecture concept for solving singularly perturbed problems in semiconductor electronics

The software architecture for automated solving of singularly perturbed problems in semiconductor electronics is presented. The system is based on the boundary correction method and domain-driven design (DDD). The developed software architecture integrates an artificial neural network for analyzing types of mathematical subproblems and selecting the most suitable method for solving them. The use of modern information technologies such as parallelization, cloud computing, and symbolic computations enables a new level of automated modeling. The research results demonstrated the effectiveness of the proposed approach in solving complex mathematical problems with singular perturbations.