A hybrid chaos algorithm and simulation study for fuzzy neural network

A hybrid optimization algorithm about chaos search for fuzzy neural networks (FNN) is proposed in this paper. Off-line optimization uses chaos algorithm, chaotic variables are applied to searching for parameters of membership function and weights for FNN, and an approximate optimal FNN is found. On-line optimization used gradient descent algorithm, the original values of gradient descent searching are approximately global optimal values of chaos search, and parameters of controller are optimized by means of combination of chaos global searching and gradient descent partial searching. Simulation results show that the optimistic method is efficient to search for parameters of controller.