小波神经网络预测电价的新改进

Electricity Price Forecasting Using Wavelet Neural Networks Optimized by GA

预测市场边际电价对于电力市场的参与者有十分重要的意义。该文首先分析了BP神经网络在电价预测方面的优劣势,然后基于小波分析,即用母小波取代Sigmoid函数建立了小波神经网络的电价预测模型,并用遗传算法优化神经网络的拓扑结构和各权重系数,从而避免BP神经网络的预测电价陷入局部极小值。实际计算表明,改进后的预测模型有效地提高了预测精度。

Forecasting the future electricity price is very important for every participators in the power market.This paper analyses the advantages and disadvantages of BP neural networks,then provides a wavelet-improved neural network model base on wavelet analysis and the topology structure and weight coefficient optimized by GA.Application to the real system shows that this model can improve the forecasting precision and avoid the limitation of the BP neural networks.