MODELLING STATIC SYSTEMS WITH NONLINEAR CHARACTERISTICS USING AN ONTOLOGICAL APPROACH

Abstract

In this study, a hybrid method for identifying interval models of static systems with nonlinear characteristics is substantiated and developed. The method is based on a knowledge-oriented approach to describing the domain concerning the class of identification problems and the rules for combining optimization methods to solve them. Rules for applying structural and parametric identification methods for interval models of static systems with nonlinear characteristics have been developed. These rules are based on the characteristics of the class of identification problems, such as the dimensionality of optimization problems and the properties of the objective function. The formalization of the proposed rules enabled the automated selection of the most effective identification methods and algorithms within the procedural part of the ontology for modelling static systems with nonlinear characteristics. The advantages of the proposed approach are illustrated using the example of modelling the generated power of the daily cycle of a solar power plant module.

The work substantiates and implements ontological descriptions of the subject area of ​​modeling static systems with nonlinear characteristics based on interval data to ensure automated control of the modeling process. As a result of the conducted research, the following results were obtained:

- on the basis of the analysis of temporal characteristics and the convergence of methods of structural and parametric identification of interval models of static systems with nonlinear characteristics, rules for the application of these methods have been developed, which are based on the properties of the identification problem, in particular, such as the dimension of the optimization problem and the complexity of the objective function. The formalization of the proposed rules enabled the automated selection of the most effective identification methods and algorithms in the implementation of the procedural part of the ontology modeling of static systems with nonlinear characteristics;

- the concept of identifying interval models of static systems with non-linear characteristics using an ontology, which includes a knowledge model, the task of which is to structure knowledge about the characteristics of problems of identifying static systems and optimization methods, and to determine the criteria for choosing a method depending on the characteristics of the problem, which collectively made it possible creation of a unified hybrid method of identification, which is the most effective from a computational point of view;

- a hybrid method of identifying interval models of static systems with nonlinear characteristics is substantiated and built, which is based on a knowledge-oriented approach to the description of the subject area of ​​identification tasks and rules for combining optimization methods, in particular global search based on gradient methods and methods of swarm intelligence (swarms of particles, behavioral models of bee colonies), based on the ontology, which collectively ensured a reduction in the computational complexity of identifying interval nonlinear models.

- as a result of the experimental studies, it was established that the hybrid method is based on the formalization of the identification process by choosing an effective optimization method based on the characteristics of the problem, which provides more effective modeling and identification of interval nonlinear models of static systems with nonlinear characteristics compared to existing ones. The advantages of the proposed approach are illustrated on the example of the problem of modeling the generated electricity power of the daily cycle of the power plant module.

Thus, the results obtained in the work contribute to the further development of methods for modeling static systems with nonlinear characteristics based on interval data, which has the potential for wide application in various fields of science and technology.