The Samples Updated Real-time Neural Network and Its Application in Short Term Load Forecasting
运用样本更新的实时神经网络进行短期电力负荷预测

A multi-layer neural network based on Single Parameter Dynamic Search Algorithm (SPDS) is proposed for short term load forecasting in this paper. The model proposed can overcome BP algorithm's dissatisfied convergence for large-scale problem,training samples is dynamically built according to the information of the weather forecast. The calculations show that the method is efficient for short term load forecasting.