Fuzzy inference with propagation of asymmetric convex constraints

A fuzzy inference method is proposed which provides control schemes for propagating the asymmetric constraints in given facts to inference consequences in convex forms. The proposed inference method is based on α-GEM (α-level-set and generalized-mean based inference). The parameters for the propagation in α-GEM are adjusted by using the fuzzy tautological rules generated from the antecedent parts of given fuzzy rules. The properties of the proposed inference method are clarified and then its feasibility is shown through simulation studies