UNCERTAINTY QUANTIFICATION THROUGH A MODEL-BASED FUZZY SET MEMBERSHIP FUNCTION

An approach to uncertainty quantification is proposed. It is based on a fuzzy set membership function that is defined through a model (or models) response. That is, information about uncertain parameters is gathered from measured responses of one or more models, the membership function is inferred and used in uncertainty quantification of the output of a model of interest. The idea is illustrated by a beam deflection problem.