Recurrent neural network applied in dynamitic process identification based on RPROP and chaos optimization coupling algorithm

A new style of dynamic recurrent neural network based on RPROP and chaos optimization coupling algorithm is provided and is applied to dynamic process identification in this paper. RPROP algorithm is good at improving neural network´s convergence speed and chaos optimization algorithm can prevent efficiently from partial minimum. Moreover, a new kind of object function is applied to this network to improve the generalization capability of neural network. By the simulation experiment, the feasibility of this method was demonstrated.