A Comparison of PSO and Backpropagation Combined with LS and RLS in Identification Using Fuzzy Neural Networks

In this article using a population-based method, particle swarm optimization in training the standard deviation and centers of radial basis function fuzzy neural networks is put into practice and the results are compared with training the same networks´ standard deviation and centers using backpropagation. We have applied Least Square and Recursive Least Square in training the weights of this fuzzy neural networks . There are four sets of data used to examine and prove that according to the convergence speed and the identification error particle swarm optimization works better and as its complexity is much less, it can be suggested as a good solution for training the parameters.