ESTIMATION OF NON-STATIONARY PARAMETERS OF NONLINEAR EQUATIONS UNDER UNCERTAINTY

Abstract

This paper presents research of system of nonlinear equations. The algorithms of building a priory estimations, optimal functional estimations and guaranteed estimations of non-stationary parameters of differential equations are offered. These results are spread for discrete-time models. The algorithms of building optimal estimations and guaranteed estimations of non-stationary parameters of difference non-linear equations are offered. The approaches to construct optimal estimations based on Bellman functions and Kalman-Bussi filter. For each algorithms of building optimal functional estimations and guaranteed estimations the error of estimation are offered. We have presented as an example the results of numerical experiments to build guaranteed and optimal estimates for mathematical model of spreading one type of information with external influence. The number of information is taken as key parameter promoting accomplishment of aim. Information is spread in the community along internal (interpersonal communication of the members of social community) and external threads (mass media). Also for simplicity models with a constant number of individuals who are intentionally able to perceive and further spread an informational massage are explored. The model takes the form of non-linear ordinary differential equation with stationary parameters. The peculiarity of such models is that they allow a reasonable level of precision to model the subject area and obtain he results that can be uses in practice. The numerical experiments demonstrated the practical meaning of offered results. The offered approaches except theoretical interest has an important practical meaning. The results can be useful for algorithm development for estimation of dynamic of process in the information-communicative space.