中文摘要：

高速列车初始晚点时长分布模型是行车指挥决策及相关理论研究的基本依据,具有重要的理论和实践意义.本文基于广铁集团高速列车运行实绩,研究不同致因情况下初始晚点时长的分布模型.以2014—2015年高速列车初始晚点时长为建模数据,通过Kolmogorov-Smirnov检验比选5类典型模型对不同致因导致高速列车初始晚点时长分布曲线的拟合效果,运用R软件对模型参数进行极大似然估计.研究发现：对数正态分布模型拟合效果最优;韦伯分布模型、伽马分布模型拟合效果较好.以2016年初始晚点时长数据进行模型校验,以2014—2016年数据进行双样本同分布检验,结果表明对数正态分布模型总体上能够满足校验数据的拟合度要求,也能通过双样本同分布检验.

英文摘要：

High-speed trains primary delay distribution models,with certain theoretical and practical signifi-cance,are the foundation for railway traffic control decision-making and relevant studies.The cause based pri-mary delay distribution models of high-speed trains were established based on the high-speed trains operation records of Guangzhou Railway（Group）Corporation.The Kolmogorov-Smirnov test was used for the compari-son of the curves＇fitness between the 5 kinds of typical candidate models based on the modeling data during 2014 to 2015,modeling the relationship between the influencing factors and the primary delay distributions,the parameters were obtained through maximum likelihood estimation using the R-proj ect.The results confirm that Log-normal distribution model has the best performances and both the Weibull distribution model and Gamma distribution model can fit the curves of influencing cause based primary delay distribution well.The model vali-dation was carried out based on the data of 2016.The independent homogeneity testing was implemented too, using the data during 2014 to 2016,which confirms that the Log-normal distribution model can fit the testing data well in general,and it can pass the independent homogeneity testing too.