基于小波变换和遗传算法优化神经网络负荷预测

提出了采用小波变换和遗传算法优化神经网络的混合模型对电力负荷进行短期预测。首先通过小波变换,将原始负荷序列分解到不同的尺度上,然后根据不同的子负荷序列的特性分别建立相匹配的神经网络模型,采用遗传算法优化各神经网络模型的初始权值,最后对各分量预测结果进行重构得到最终预测值。采用成都某地区2009年的实际负荷对所提方法进行验证,实验结果表明基于该方法的负荷预测系统具有较高的预测精度。

Load Forecasting Based on Wavelet Transform and Neural Network Optimized by Genetic Algorithm

A novel short-term load forecasting method using wavelet transform and neural network optimized by genetic algorithm is proposed.Firstly,by the wavelet transform,the load series is decomposed into the subseries with different frequency characteristics,then according to the features of the decomposed components,the corresponding neural network models are constructed to forecast the components,finally the forecasting result is obtained by the reconstruction of the forecasting result of components.The genetic algorithm optimization is used to optimize the initial weights of neural network model of each decomposed subsequence.Experimental results show that the proposed forecasting method has a satisfactory accuracy.