Possibilistic evidential reasoning systems on systolic arrays

Abstract

Suggests a systolic array implementation of fuzzy expert systems based on possibilistic evidential reasoning methodology. Evidential reasoning systems are computationally-bounded in general, and fuzzy systems match a larger number of rules than in a simple symbolic reasoning system. The proposed possibilistic evidential system, which is a combination of both types of system, needs much more computation time than either of these two independently. To speed up the processing in such systems, the suggested implementation is useful.<>