Terentiev O. Models, methods and information technologies of forecasting nonlinear nonstationary processes in conditions of uncertainty

The thesis focuses on solving urgent scientific and applied problems related to development and usage of models and methods for forecasting nonlinear nonstationary processes of various origin to be further used in modern information decision support systems. In the frames of the thesis new information technology was developed for implementation and application of high adequacy and quality forecasting mathematical models that allow for forecasting nonlinear nonstationary processes of various types and origin. The information technology developed is based upon the principles of multi-model approach, integration of structured and non-structured information, systemic applications of the methods for intellectual data analysis, modeling, forecasting and decision making. The principal scientific results include the following ones: information technology for solving the problem of modeling and forecasting nonlinear nonstationary processes; improved methods of forecasting the processes mentioned and of various origin on the basis of usage the probabilistic, statistical and combined forecasting techniques; the method for model parameters estimation as well as for their ensembles; the method for uncertainty identification and taking into consideration the uncertainty by making use of probabilistic and statistical approaches with the final purpose of correct solving the problem of forecasting selected processes; the approaches that provide for adequate formal description of causal relationships between various groups of factors and determining possible future alternatives for development of nonlinear nonstationary processes in various areas of human activity; the information technology was developed and practically implemented for solving the problem of forecasting nonlinear nonstationary processes, and its further use in decision support systems. Multiple application examples based on statistical data collected are given illustrating practical usage of the methods, models and systemic methodology developed.