Uncertainty representation in nonlinear and hybrid systems

Information of any kind always involve some kind of uncertainty, inaccuracy. The model used to represent the information during the information processing can affect the achievable preciseness and can determine the usability of different computing methods, as well. The data model must also be able to represent the uncertainty, inaccuracy of both the input data and the results. The two most popular data models for the representation of uncertain data are the "classical", probability based and the recently introduced fuzzy data model. With the increasing complexity of the computing problems and with the appearance and spreading of new modeling and computing techniques, the evaluation and representation of the uncertainty in different systems become an increasingly important question and methods for the mixed use of different data models can also be needed. This paper deals with the limits of the "classical", probability theory based uncertainty representation and examines possible solutions for information processing based on mixed data models together with different conversion methods between fuzzy and probability theory based representations.