Abstract: Aiming at the deficiency of current forecasting of short-term traffic flow,the method of combining wavelet transformation and chaos theory is proposed to model and forecast short-term traffic flow.Firstly,using wavelet decomposition theory,traffic flow series are decomposed respectively into two parts:the low frequency part and the high frequency part.And the further analysis of decomposition indicates that a chaos feature exists in the traffic flow.Secondly,using chaos theory,the chaotic forecasting models are established to forecast the low frequency part and the high frequency part respectively.Finally,the forecasting result of chaotic model is reconstructed by wavelet theory.By doing so,the forecasting of the original traffic flow series can be made.The result demonstrates that this method is of high precision and has extensive potential applications.

Key words: wavelet decomposition, traffic flow, chaos, forecasting

摘要： 针对现有短时交通流预测模型的不足,提出了一种用于交通流短时预测的小波与混沌集成方法。首先对交通流序列进行小波分解,分别得到低频部分和高频部分,并在此基础上作进一步分析,结果表明交通流存在混沌特性。然后应用混沌理论分别建立低频部分和高频部分的预测模型,对低频部分和高频部分进行预测。最后应用小波理论对混沌模型预测的结果予以重构,实现对原始交通流序列的预测,与现有方法比较,结果表明该方法具有较高的精度和应用前景。

关键词: 小波分解, 交通流, 混沌, 预测