Seyed Mohammad Attaran
Shahrum Shah Bin Abdullah

In this paper we provide a short review of radial basis function (RBF) and its properties. In addition RBF NN was used to estimate the non-Isothermal CSTR. For achieving this goal because of on lining data it was necessary to use the RLS each time that new data come to system. By using this method (RLS), RBF NN updates its weights for mapping the input and output of the system. The result shows that this neural network can be worked as a good estimator. It is clear that the RBF NN can be used as modeling of the non-linear isothermal CSTR. According to the activation function that consists of Gaussian function it can solve the non-linear part with good approximation.

Neural network,RLS,RBF,Estimate