Multimodel Approach to Forecasting Nonlinear Nonstationary Processes in Optimal Control Problems

This paper proposes different information technologies of forecasting, which are based on the step-by-step disclosure of uncertainties of different nature, refinement of modeling results at each stage of the study. The peculiarity of this approach is that the decision-maker has the opportunity to get not only a "ready-made" solution, but also to participate in the formation of models, to propose its own scenarios and forecasting horizons for processing using actual data. The best candidate models, input data are used: time series of the processes under study and additional factors significant at a given forecasting horizon.