中文摘要：

针对因风速具有很强的波动性和间歇性而导致其难以预测的问题．提出了一种新的基于小波分解和微分进化支持向量机的预测方法，通过小波变换对风速数据进行多分辨率分解，并以微分进化优化的支持向量机对各分解层的风速分别建立预测模型，然后将各模型的预测结果叠加后作为最终的预测值。用某风电场实测风速数据进行仿真预测，结果表明，所提方法与交叉验证支持向量机和BP神经网络等常用的预测方法相比．具有更高的预测精度。

英文摘要：

Since it is difficult to forecast the wind speed because of its fluctuation and intermittence,an approach based on wavelet decomposition and DE-SVM（Differential Evolution-Support Vector Machine） is proposed,which carries out the multi-resolution decomposition of wind speed data by the wavelet transform, builds the forecasting model based on DE-SVM for each scale,and combines the forecasts of different models to get the final forecasting result. The proposed approach is applied to the real wind speed data of a wind farm and its forecasting result is compared with those of the cross validation SVM and BP neural network, which demonstrates its higher forecasting precision.