Short-Term Load Forecasting Based on RBFNN and QPSO

Abstract

Coping with the questions of radial basis function neural network (RBFNN) in short-term load forecasting, a new training method of the RBF neural network based on quantum behaved particle swarm optimization (QPSO) algorithm was introduced. In the algorithm, all network parameters were coded into individual particles which can search optimal-adaptive values at random in the overall space. So, the parameters can be quickly and accurately identified. The application in power load forecasting show that the method can accelerate convergence speed of the network and increase accuracy of predicting compared with traditional RBFNN.