基于粒计算的交通流参数预测

发布时间：2018-04-22 21:10

本文选题：城市交通 + 信息颗粒　；参考：《交通运输系统工程与信息》2015年06期

【摘要】：交通流参数预测是交通流诱导和交通信息发布的重要依据.以信息颗粒为基础数据分析单元,针对以往模糊时间序列模型存在的缺陷,提出一种新方法构建模糊时间序列模型,该方法在挖掘数据内在信息关联的基础上,考虑时间变量影响下分析动态可变的区域间隔长度.此方法主要特点是基于Gath-Geva模糊聚类的时间序列分割,利用模糊分割构造信息颗粒,以信息颗粒为数据单元,通过粒计算分析交通流参数动态变化趋势.实验结果表明,基于粒计算的交通流参数预测可以预测合理的交通流参数置信区间,比传统的参数数值预测可靠度更高,对于交通状态的动态分析具有指导意义.
[Abstract]:Traffic flow parameter prediction is an important basis for traffic flow guidance and traffic information release. This paper presents a new method to construct the fuzzy time series model, which is based on the information particle as the basic data analysis unit, aiming at the defects of the previous fuzzy time series model. The method is based on mining the internal information association of the data. Considering the influence of time variable, the dynamic variable region interval length is analyzed. The main feature of this method is the time series segmentation based on Gath-Geva fuzzy clustering. The information particle is constructed by fuzzy segmentation. The information particle is taken as the data unit and the dynamic change trend of traffic flow parameters is analyzed by particle calculation. The experimental results show that the prediction of traffic flow parameters based on granular computing can predict reasonable confidence interval of traffic flow parameters, which is more reliable than the traditional numerical prediction of traffic flow parameters, and has guiding significance for the dynamic analysis of traffic state.
【作者单位】：四川省交通运输厅公路规划勘察设计研究院;西南交通大学交通运输与物流学院;
【基金】：国家自然科学基金资助项目(51308475) 四川省科技支撑计划资助项目(2011FZ0050)
【分类号】：U491

【相似文献】