RNDr. Libor Dostálek, Ph.D.

Publications

A Cloud-Based Software Architecture for Mathematical Modeling based on Interval Data Analysis

Authors
Dyvak, M.; Manzhula, V.; Melnyk, A.; Dostálek, L.
Year
2024
Published
2024 14th International Conference on Advanced Computer Information Technologies (ACIT). IEEE eXplore, 2024. p. 89-93. ISSN 2770-5226. ISBN 979-8-3503-5003-6.
Type
Proceedings paper
Annotation
In this paper, we present a software architecture for mathematical modeling based on the analysis of interval data using cloud technologies. The key features of the proposed architecture include the integration of an interval modeling subsystem for static systems into a cloud-based service-oriented architecture, optimization of computational schemes using the Google Cloud Run platform, the use of the MapReduce distributed computing model, free software-interpreted tools, and the application of RESTful APIs at all stages of mathematical modeling.

Parallel Computations in the Problem of Identification of Interval Discrete Models based on Swarm Intelligence of a Bee Colony

Authors
Dyvak, M.; Kindzerskyi, O.; Dostálek, L.
Year
2023
Published
2023 13th International Conference on Advanced Computer Information Technologies (ACIT). IEEE eXplore, 2023. p. 23-28. ISSN 2770-5226. ISBN 979-8-3503-1167-9.
Type
Proceedings paper
Annotation
The article discusses the problem of parametric identification of interval discrete models of complex objects. A new scheme of parallel computations for solving the problem of parametric identification based on the behavioral model of an artificial bee colony is proposed and justified. The effectiveness of the proposed parallel computing scheme is considered using the example of parametric identification of a mathematical model of the anaerobic microbiological fermentation process in biogas plants. A nonlinear second-order discrete model describes the dynamics of residual organic waste mass in the biogas reactor during fermentation. According to the results of the experiments, a 2-fold acceleration of the algorithm’s computational time was achieved.