Application of fuzzy inference algorithm in artificial neural networks forecasting

In order to describe the non-linear characteristic relation of power load varying, a short-term load forecasting model based on fuzzy neural networks is presented. The model consists of a multi-inputs and one-output neural networks model which is not entirely linked. The fuzzy inference and defuzzification of the model are both realized by neural networks. The membership function of fuzzification layer is selected and in fuzzy inference layer fuzzy getting smaller inference algorithm is put forward to finish fuzzy inference. For the sake of extending the scope of weight acting, the non-linear feedback item of weight value is employed to form chaos learning algorithm, which can make the system find the global minimal point or its approximation quickly and enhance the stability of the model. The simulation results indicate that the maximal error and the mean absolute percentage error of the forecasting daily load for next day are 2.90% and 1.26%, respectively. The model proposed shows preferable forecasting capability.