Article
Keywords:
fuzzy set; membership function; uncertainty quantification
Summary:
Since the common approach to defining membership functions of fuzzy numbers is rather subjective, another, more objective method is proposed. It is applicable in situations where two models, say $M_1$ and $M_2$, share the same uncertain input parameter $p$. Model $M_1$ is used to assess the fuzziness of $p$, whereas the goal is to assess the fuzziness of the $p$-dependent output of model $M_2$. Simple examples are presented to illustrate the proposed approach.