中文摘要：

为提高风速时间序列的预测性能，针对具有混沌特性的风速时间序列提出一种混合预测方法。通过分析风速时间序列的动力学特性，求解风速时间序列蕴含的最佳不稳定周期轨道，将前一最佳不稳定周期轨道附近的值作为当前预测结果，从而得到基于混沌不稳定周期轨道的预测结果。另外，将混沌算子网络应用于风速时间序列预测分析中，通过调节网络参数改变预测网络的动力学特性，从而实现风速时间序列预测分析。将这两种具有不同机理的预测方法通过优化融合指标函数的方式实现预测结果融合，从而实现风速时间序列的混合预测。仿真结果表明，混合预测方法能够进一步提高风速时间序列的预测性能。

英文摘要：

A hybrid prediction method for wind speed time series with chaotic characteristic is proposed to improve the forecasting performance. The best unstable periodic orbit is resolved by analyzing the dynamic characteristic of the wind speed time series. The prediction value is obtained by the data points near the former best unstable periodic orbit. Another prediction method, chaotic operator network, is also applied to predict the wind speed time series. The dynamic characteristic of the prediction network can be changed by adjusting the control parameters. The fusion prediction results of two different prediction methods are obtained by optimizing the fusion index function. In this way, hybrid prediction of wind speed series can be realized. Simulation results show that the hybrid prediction method can further improve the prediction performance.