Solving fuzzy relational equations through logical filtering

The paper introduces an idea of logical filtering viewed as a new tool for solving fuzzy relational equations. Considering the panoply of the existing methods, the proposed approach can be classified as semi-analytic in the sense it departs from the individual analytical solutions to the individual equations in the system and combines them through an optimization process of logical filtering (masking). Several types of filtering are studied and provided with detailed learning schemes.