短时交通量时间序列智能复合预测方法概述  

Survey of Traffic Volume Time Series Intelligent Compound Forecasting Methods

作　　者：张益[1] 陈淑燕[1] 王炜[2] 

机构地区：[1]南京师范大学江苏省光电重点实验室,江苏南京210097 [2]东南大学交通学院,江苏南京210096

出　　处：《公路交通科技》2006年第8期139-142,共4页Journal of Highway and Transportation Research and Development

基　　金：国家自然科学基金资助项目(50378016);江苏省教委自然科学基金资助项目(05KJB520056)

摘　　要：短时交通量预测是智能运输系统的核心研究内容之一,已成为交通工程领域重点研究课题。对国内外短时交通量时间序列的预测方法尤其是智能复合预测方法进行概述和总结,重点介绍灰色预测模型、模糊预测、遗传算法、神经网络、灰色神经网络、神经网络集成、统计学习理论、混沌预测、小波分解与重构的方法、以及由上述模型互相组合构成的各种智能组合预测模型等,并指出智能复合预测方法是解决短时交通量时间序列预测问题的有效途径和发展趋势。Traffic flow forecasting is one kernel study in Intelligent Transportation System and a research focus in traffic engineering field. The methods of traffic volume forecasting especially intelligent compound forecasting approach are summarized. Special stress was laid on several typical traffic volume forecasting methods including grey model, fuzzy prediction, genetic algorithm, neural network, grey neural network, neural network ensemble, statistical learning theory, chaos forecasting, wavelet decomposition and reconstruction, and some compound models built with the above mentioned models. Furthermore, it is pointed out that intelligent compound forecasting methods are the development trend and effective research solutions to short-term traffic volume forecasting.

关 键 词：交通工程 交通量时间序列 智能组合预测 综述 

分 类 号：U491.14[交通运输工程—交通运输规划与管理]