Local search for learning algorithm in adaptive fuzzy inference system

Abstract

In this paper, a local search for learning mechanism was proposed to improve the adaptive fuzzy inference system. The adaptive fuzzy inference system is a complementary technology based on the concept of fuzzy theory, if-then rules and fuzzy reasoning. The learning or training capability of this system is provided by the neural network through a learning mechanism. The learning algorithm we have proposed in this paper is based on the local search. The simulation is carried out based on the famous Mackey-Glass time series. Our results show that the local research for learning algorithm in adaptive fuzzy inference system is useful and effective because it requires less memory and it is able to overcome the disadvantages of the gradient descent. This demonstrates that the local search is very suitable for learning mechanism in the adaptive fuzzy inference system.