Research into the effectiveness of optimization methods for building models of complex systems
Computational Problems of Electrical Engineering · 18.04.2026 · 16 (1) · 32-41
DOI: 10.23939/jcpee2026.01.032 ↗Kurz gefasst
Fachartikel in der Zeitschrift Computational Problems of Electrical Engineering, online erschienen am 16.06.2026. Verfasst von Bohdan Melnyk. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 34 Literaturangaben.
Abstract (Originalsprache)
The mathematical description of the behavior of many complex technical objects or systems involves building models in the form of a system of equations with a high degree of nonlinearity. The variables in these equations are the observed indicators of the object behavior. In the case of a “black box” model, such indicators are the signals observed at the input and output of a complex object. The form of the equations in the model, which is the essence of structural identification, is often determined by the analysis of the nature of the general object behavior. The model will adequately reflect this behavior by selecting the appropriate coefficients in the equations, which are the parameters of this model. This stage of model building is called parametric identification. In many cases, the problem of parametric identification is solved by applying optimization methods. It is important for the researcher to choose the most effective method for solving a certain class of such problems. There can be several criteria for the effectiveness of the method, namely the accuracy of reaching the global extremum point, the number of iterations required to find this point, the number of calculations of the objective function and each iteration, the ability of the method to leave the zone of attraction of the local extremum, the speed of directed movement towards the global extremum point. Determining the advantages of various optimization methods with regard to these criteria requires detailed research. This article considers the application of several optimization methods for a parametric model of the “black box” type of an electromechanical object, in particular, several modifications of the directional cone method, as well as methods implemented in the scipy.optimize library of the Python language. To study the efficiency of these methods, appropriate software was developed. A comparative analysis of their efficiency was conducted according to several criteria, which allowed for concluding that it is appropriate to use a specific method to solve the optimization problem for building parametric models of the “black box” type of electromechanical objects.
Zitieren
APA
Melnyk, B. (2026). Research into the effectiveness of optimization methods for building models of complex systems. Computational Problems of Electrical Engineering, 16(1), 32-41. https://doi.org/10.23939/jcpee2026.01.032
Vancouver
Melnyk B. Research into the effectiveness of optimization methods for building models of complex systems. Computational Problems of Electrical Engineering. 2026;16(1):32-41. doi:10.23939/jcpee2026.01.032
BibTeX
@article{Melnyk2026,
author = {Melnyk, Bohdan},
title = {Research into the effectiveness of optimization methods for building models of complex systems},
journal = {Computational Problems of Electrical Engineering},
year = {2026},
volume = {16},
number = {1},
pages = {32--41},
doi = {10.23939/jcpee2026.01.032}
}
EUVIMED