Optimization of Fuzzy Neural Network Using Multiobjective NSGA-II

In the area of computational intelligence as like Artificial Neural Networks (ANNs) or Fuzzy logic have been used for the construction of an effective and reliable system in order to solve a real world problem where appropriate outcome along with certainty as well as precision are highly required. In this article, we present an integrated approach based on a fast elitist non-dominated sorting genetic algorithm and ANN for constructing optimal fuzzy systems. At First, the neural network with clustering method, used as a fuzzy rule generator to generate training fuzzy logic rules for the NSGA-II (Non dominated sorting genetic algorithm II). Multi-objective NSGA-II is used to optimize the fuzzy model involving more than three objective constraint to be augmented concurrently that are directly related to the fitness factor of the controller. In contrast with other conventional fuzzy model, this multi-objective fuzzy-NSGA-II controller achieves benefits over the control performance with an oppositeness and probability.