A design method based on radial basic function(RBF) neural network combining immune flocking with immune evolution planning has been put forward. In this method

the number of indistinct nodal points and the centre position of data for RBF can be adaptively determined according to input data collection by adopting artificial immune flocking mechanism and the weight value from output layer of the network determined by taking the width of nuclear function as vaccine and selected operation for immune. Adopting the design method based on RBF neural netowork

a RBF network model for the outlet temperature and the outlet flow rate of a medium speed coal pulverizer has been established. Results of simulation show that the established model boasts higher accuracy.