Computational Complexity of Structural Identification Method for Interval Discrete Models
Mykola Dyvak, Natalia Porplytsya, Yurii Maslyiak · 2019
One of the stages of building of discrete models based on experimental data, is structural identification. Structural identification means the procedures of finding of general form of models. In cases if experimental data are inaccurate, represented in an interval form, and mathematical models of the processes are represented in the form of difference equations, structural identification tasks have a very high computational complexity. One of the methods of structural identification of interval discrete models, is a method based on bee colony behavioral models. However, the complexity of this method has not been investigated. The problem of high computational complexity of this method constrains its wide application for a large spectrum of modeling tasks in technical, environmental, medical, economic and other fields. The paper is devoted to investigation of this problem, the theoretical substantiation of the causes of high computational complexity and ways to reduce it. The paper is of a theoretical nature, but will be widely used in the case of adding the computational procedures based on the results of the research to the method of structural identification of interval discrete models.