Rule based fuzzy logic inferencing
Yiming Yu, E.R. Fowler · 2002
This paper presents a new method of rule based fuzzy logic inferencing solely based on a many-valued logic, Lukasiewicz many-valued logic. Fuzzy logic can be considered as an extended set many-valued logic. Many-valued logic is extended from two-valued logic but differs from it in that each has different sets of primitives. The authors believe that fuzzy logic should be solely based on many-valued logic. In this paper, an implication connective of Lukasiewicz many-valued logic is proposed to be a basis of the conditional proposition in a new fuzzy logic inferencing method. Corresponding to this implication connective, a max-/spl delta/ composition is derived from Lukasiewicz many-valued logic and it is shown to fit human intuition. The advantage of the new fuzzy logic inferencing, over other fuzzy logic inferencing, is that it provides a solid theoretical basis for further study of its inferencing behavior when used in a certain application, such as in a fuzzy controller. An example, using machine vision to detect surface flaws of machined parts, is given.>