摘要
为增进港口泊位船舶到达分布模型准确性,分析总结传统分布模型的局限性,引入对数正态分布模型,使用极大似然估计方法进行参数估计完成模型构建.选取天津港作为实例,分析天津港2012—2014年船舶到达数据,为避免船型差异对结果造成影响,选定散货船、集装箱船、油船三种目标船型,分别研究散货、集装箱和原油泊位的船舶间时距概率分布.使用对数正态分布、埃尔朗分布、γ分布对样本拟合,通过卡方拟合优度检验得出对数正态分布的拟合程度最优,验证了对数正态分布的有效性和优越性.
In order to promote the accuracy of ship arrival distribution model in port berths,the limitations of the traditional distribution models were analyzed and the lognormal distribution model was introduced to estimate parameters and construct the model with the maximum likelihood estimation methord.Taking Tianjin Port as the example,the vessel arrival data from 2012 to 2014 was analyzed.In order to avoid ship type difference on the results,three target ship types:bulk carrier,container ship and oil tanker were chose to analyze the probability distribution of time-interval between ships in port berths.The lognormal distribution,Erlang distribution and gamma distribution were adopted to fit the above sample respectively.The fitting degree of lognormal distribution is the most optimal by goodness-of-Fit tests.
作者
罗肖霞
刘奕
刘敬贤
LUO Xiaoxia;LIU Yi;LIU Jingxian(School of Navigation,Wuhan University of Technology,Wuhan 430063,China;Hubei Key Laboratory of Inland Shipping Technology,Wuhan 430063,China)
出处
《武汉理工大学学报（交通科学与工程版）》
2018年第2期313-317,共5页
Journal of Wuhan University of Technology（Transportation Science & Engineering）
基金
国家自然科学基金项目(51479156)
国家自然科学基金青年科学基金项目(51709219)
湖北省自然科学基金项目(2016CFB270)资助
关键词
船舶间时距分布
对数正态分布
传统分布模型
极大似然估计
拟合优度检验
Time-Interval between ships
lognormal distribution
traditional distribution model
maximum likelihood estimation
Goodness-of-Fit test
