Abstract. The issues related to granulating information and granular calculations (granular computing) in presenting FS-granules as a Kronecker product are considered. The proposed model is shown to be universal, in particular, in calculating the inverse value of fuzzy variable X = {x / μ^x} and solving other control problems under uncertainty. It is shown that the proposed form of FS-granules, tensor (Kronecker) product of components x⊗μ^x, allows preserving the integrity of granules and promotes the solution of the problem of granular computing and transformation of granules.

Keywords: tensor, fuzzy set, Kronecker product, norm, information granule, singular value decomposition.