Novel Fuzzy Logic Controller Based on Weighting of Partially Inconsistent Rules Using Neural Network

Abstract

In this paper a new method of fuzzy logic control based on possibly inconsistent if-then rules representing uncertain knowledge or imprecise data is studied. In most cases of practical applications adopting fuzzy if-then rule bases, inconsistent rules have been considered as ill-defined rules and, thus, not allowed to be in the same rule base. When it is hard to obtain consistent rule bases, we propose a fuzzy logic controller based on weighted rules depending on output performances using the neural network and we will derive a weight updating algorithm. To guarantee convergence of the weights, a learning rate is developed by introducing a Lyapunov function. With the final weight change informations, we can make better decisions by taking into consideration conflicting rules. The proposed method is applied to simple problems and simulation results are included. And real application problems are also discussed.