摘要
重构相空间理论对我国铁路客流量数据时间序列进行相空间重构,通过计算分形维数和提取最大Lyapunov指数,分析得出了铁路客流量时间序列数据的演化具有混沌特征,结合我国铁路客流量在全年中的波动特点,提出适合我国国情的综合混沌相空间重构与相似性原理的铁路客流量预测算法.该算法能自动参考过去相似年度同期的客流量变化趋势,对预测值进行智能化调整.通过以某火车站客流量为预测对象进行验证,结果证明了该预测算法的准确性和实用性.
This paper develops a new algorithm with promising performance to forecast short term railway passenger demand. Based on the theory of phase space restructuring, the phase space of railway passenger demand time series data is restructured, their fractional correlation dimension and maximum Lyapunov exponent are calculated, and the conclusion that railway passenger demand time series data has chaotic property is deduced, and then according to the annual fluctuation characteristic of China railway passenger demand, the algorithm based on integrating chaotic phase space restructuring and principle of similarity is proposed. In the proposed algorithm, the inside law of passenger demand evolvement is abstracted from the history data of railway passenger demand, the trend of passenger demand evolvement in the same term of similar year is taken into account, and the forecasting value can be automatically optimized. Test results for using this algorithm to forecast the short term railway passenger demand of an actual railway station are reported, and show that the proposed algorithm can attain promising performance and practicality.
出处
《武汉理工大学学报（交通科学与工程版）》
2007年第4期684-687,共4页
Journal of Wuhan University of Technology（Transportation Science & Engineering）
基金
江西省教育厅科技研究基金项目资助(批准号:2007188)
关键词
铁路客流量
混沌预测
相空间重构
LYAPUNOV指数
相似性
railway passenger demand
chaotic forecasting
phase space restructuring
lyapunov exponent
principle of similarity
