Evolutionary methods in modelling behaviour of complex system

In this paper, a recursive-regression approach is formulated in the formation of a functioning model of a complex system. The complex system consist of the object of study represented by a set of related factors. Some of these factors can be specified from the outside, whereas other factors contain information generated by system. We demonstrate that time-dependence of these factors can be reproduced by the nonlinear regression model. The proposed sub-course combines the method of group accounting of arguments with standard regression analysis adapted to fast dynamic processes. As an example, recursive-regression modelling of the functioning of an industrial enterprise was performed. This allows us to predict a possible behaviour of the system and to identify the so-called significant factors that have a significant impact on the behaviour of the system. To demonstrate validity of the method, we apply it to analyse the data characterizing a manufacturing company. Also paper includes meteorological data analysis. An important advantage of the proposed approach is that models for effective factors are formed automatically at each time step. The given results make possible to determine the optimal parameters of introduced algorithm and number of time observations for analysis.