论文部分内容阅读
对于服从对数正态分布的同一组样本,不同估计方法可能会产生不同的参数估计结果。针对这种情况,运用高速公路收费数据分别对客车和货车的行程时间分布进行了对数正态分布拟合。以误差平方和、累积分布函数的误差平方和、确定系数、偏度系数的相对误差和峰度系数的相对误差为衡量指标,比较最大似然估计和最小二乘估计的拟合结果。结果表明:最大似然参数估计在拟合累积概率分布时效果更优。建立了路段单位距离行程时间可靠性模型,提出了参数及阈值标定适用的估计方法。最后,对陕西省的高速公路网进行了实例分析。分析表明:出行距离对货车和客车的行程时间可靠度均有影响,出发时段对客车的出行延误影响较货车显著。
For the same set of samples that obey the lognormal distribution, different estimation methods may produce different parameter estimation results. In view of this situation, the use of highway toll data separately for passenger cars and trucks travel time distribution of the lognormal distribution fitting. The square error sum of squares, the sum of squared error of the cumulative distribution function, the relative error of the coefficient of determination, the coefficient of skewness and the relative error of the kurtosis coefficient are used as the indexes to compare the fitting results of the maximum likelihood and the least-square estimation. The result shows that the maximum likelihood parameter estimation is more effective in fitting the cumulative probability distribution. The reliability model of distance traveled per unit distance was established, and the estimation methods suitable for parameter and threshold calibration were put forward. Finally, an example is given to the expressway network in Shaanxi Province. The analysis shows that the travel distance has an impact on the reliability of the travel time of trucks and buses, and the departure time has a more significant impact on the travel delays than the trucks.