Simulation and Analysis of FuzzyParameterized Models with the Extended Transformation Meth.pdfVIP

Simulation and Analysis of Fuzzy-Parameterized Models with the Extended Transformation Method Michael Hanss Institute A of Mechanics, University of Stuttgart Stuttgart, Germany m.hanss@mecha.uni-stuttgart.de Abstract tion of fuzzy-parameterized models, the complete informa- tion about the uncertainties in the model can be included The transformation method has been pr oposed as a pr acti- and one can demonstrate how these uncertainties are prop- cal toolfor the simulation and the analysis of systems with agated through the calculation procedure. The transforma- uncertain,fuzzy-valued modelpar ameters usingfuzzy arith- tion method avoids the possibly serious drawbacks of con- metic. Up to now, this method has been available in two ventional fuzzy arithmetic [2, 4], and, as an additional ad- forms: in a general form, which can be used for the simu- vantage, it can also be used to determine the degrees of lation and the analysis of arbitrarily non-monotonic pr ob- influence of each fuzzy parameter, i.e. the proportion to lems, and in a reduced form, which reduces the computa- which the uncertainty of each model parameter contributes tional costs of the method to a large extent, but requires, to the overall uncertainty of the model output. instead, an additional condition to be fulf illed. In this pa- Up to now, the transformation method has been available per, the extended transformation method will be introduced in two forms: in a general form, which can be used for as an advanced version