基于非参数核密度估计的集装箱码头交通需求预测模型
|
摘要
针对现有集装箱集疏港时间概率分布建模中需要假设参数分布的问题,基于非参数核密度估计理论建立集装箱码头交通需求预测模型.非参数核密度估计的核函数选取高斯核,最优带宽由交叉验证法求得,通过检验、K-S检验和后验检验对比分析了核密度估计与两种传统参数模型的估计效果,并应用该模型预测DCT码头的交通需求.结果表明:非参数核密度估计模型具有更高的拟合精度、稳定性和适用性,得到的概率密度曲线能更加准确反映出口箱集港时间和进口箱疏港时间的整体分布形态,基于非参数核密度估计的集装箱码头交通需求预测模型具有比传统的参数模型更高的预测精度,可为集装箱码头基础设施规划、集疏港通道的道路交通管理、码头资源优化配置和调度等问题的研究提供更准确的交通量和作业任务量预测.
To address the inadequacies associated with present parametric density estimations for containers' delivery and pick up time distributions, a traffic demand forecasting model based on non-parametric kernel density estimation was developed.Gaussian kernel was chosen as the kernel function and the optimal window width was obtained by the cross-validation method.The χ~2 test, K-S test and posteriori test were used to compare the goodness-of-fit of the proposed probabilistic model and two conventional parametric distribution models, and the proposed model was applied to forecast the traffic demand of DCT. The results demonstrate that the proposed non-parametric estimation has better accuracy, stability and applicability. The probability density curve obtained can more accurately reflect the overall distribution pattern of containers' delivery and pick up time. The traffic demand forecasting model based on non-parametric kernel density estimation has higher prediction accuracy than that of the conventional parametric distribution models, which can provide more accurate traffic volume and task volume prediction for the infrastructure planning of container terminal, road traffic management, allocation and scheduling of terminal resources.
To address the inadequacies associated with present parametric density estimations for containers' delivery and pick up time distributions, a traffic demand forecasting model based on non-parametric kernel density estimation was developed.Gaussian kernel was chosen as the kernel function and the optimal window width was obtained by the cross-validation method.The χ~2 test, K-S test and posteriori test were used to compare the goodness-of-fit of the proposed probabilistic model and two conventional parametric distribution models, and the proposed model was applied to forecast the traffic demand of DCT. The results demonstrate that the proposed non-parametric estimation has better accuracy, stability and applicability. The probability density curve obtained can more accurately reflect the overall distribution pattern of containers' delivery and pick up time. The traffic demand forecasting model based on non-parametric kernel density estimation has higher prediction accuracy than that of the conventional parametric distribution models, which can provide more accurate traffic volume and task volume prediction for the infrastructure planning of container terminal, road traffic management, allocation and scheduling of terminal resources.
引文
[1] KOUROUNIOTI I, POLYDOROPOULOU A. Identification of container dwell time determinants using aggregate data[J]. International Journal of Transport Economics, 2016, 44(4): 567-588.
[2] KOUROUNIOTI I, POLYDOROPOULOU A, TSIKLIDIS C . Development of models predicting dwell time of import containers in port container terminals – an artificial neural networks application[J]. Transportation Research Procedia, 2016, 14:243-252.
[3] 杨龙海,章锡俏,李社教,等.深圳盐田港保税区货运交通需求预测[J].城市交通,2007,5(2):51-54. YANG H L, ZHANG X Q, LI S J, et al. The freight traffic demand forecasting of Shenzhen Yantian Harbor Free Trade Zone [J]. Urban Transport of China, 2007,5(2):51-54.(in Chinese)
[4] BOER C A, VERBRAECK A, DE A, et al. Distributed E-services for road container transport simulation [C]//Proceedings 15th European Simulation Symposium. 2003:541-550.
[5] 滕藤, 曾庆成, 陈文浩. 集装箱码头集卡到达时间预测模型[J]. 武汉理工大学学报(交通科学与工程版), 2015, 39(6):1221-1224.TENG T, ZENG Q C, CHEN W H. Prediction model for truck arrival time in container terminals[J]. Journal of Wuhan University of Technology (Transportation Science & Engineering) , 2015, 39(6):1221-1224. (in Chinese)
[6] 张莎莎, 庞明宝, 杜宏亮, 等. 基于船期表的一种集装箱码头闸口交通需求智能预判方法[J]. 交通科技, 2014 (5):148-151.ZHANG S S, PANG M B, DU H L, et al. An intelligent traffic demand forecasting method for container terminal based on line schedule [J]. Transportation Science & Technology, 2014 (5):148-151. (in Chinese)
[7] 陈家源, 郑仁环. 港口通过能力系统理论研究——港口交通流的分布与生产状态的分析方法[J]. 大连海事大学学报, 1993,19(3):297-309.CHEN J Y, ZHENG R H. System theories research of the port’s throughout capacity[J]. Journal of Dalian Maritime University, 1993,19(3):297-309. (in Chinese)
[8] 杨忠振, 陈刚, 陈康. 基于码头作业形态的港区道路交通需求预测[J]. 大连海事大学学报, 2009, 35(4):35-38.YANG Z Z, CHEN G, CHEN K. Traffic demand forecasting model for large port based on terminal operation situation[J]. Journal of Dalian Maritime University, 2009, 35(4):35-38. (in Chinese)
[9] YANG Z, CHEN G, MOODIE D R. Modeling road traffic demand of container consolidation in a Chinese port terminal[J]. Journal of Transportation Engineering, 2010, 136(10):881-886.
[10] 郭子坚, 方泽兴, 王文渊,等. 集装箱港区集疏港货运交通随机时变特性仿真研究[J]. 港工技术, 2017, 54(5):11-15.GUO Z J, FANG Z X, WANG W Y, et al. Simulation research on freightage changing randomly with time in container harbor[J]. Port Engineering Technology, 2017, 54(5):11-15. (in Chinese)
[11] 薛留根. 应用非参数统计[M]. 北京:科学出版社, 2013.
[12] 孙志华. 非参数与半参数统计[M]. 北京:清华大学出版社, 2016.
[13] PARZEN E. On estimation of a probability density function and mode[J]. The Annals of Mathematical Statistics,1962, 33(3):1065-1076.
[14] 徐玉琴, 张扬, 戴志辉. 基于非参数核密度估计和Copula函数的配电网供电可靠性预测[J]. 华北电力大学学报, 2017, 44(6):14-19.XU Y Q, ZHANG Y, DAI Z H. Power supply reliability prediction of distribution network based on nonparametric kernel density estimation and copula function[J]. Journal of North China Electric Power University, 2017, 44(6):14-19. (in Chinese)
[15] RUDEMO M. Empirical choice of histograms and kernel density estimators[J]. Scandinavian Journal of Statistics, 1982, 9(2):65-78.
[16] BOWMAN A W. An alternative method of cross-validation for the smoothing of density estimates[J]. Biometrika, 1984, 71(2):353-360.
[17] QIN Z, LI W, XIONG X. Estimating wind speed probability distribution using kernel density method[J]. Electric Power Systems Research, 2011, 81(12):2139-2146.
[18] 颜伟, 任洲洋, 赵霞,等. 光伏电源输出功率的非参数核密度估计模型[J]. 电力系统自动化, 2013, 37(10):35-40.YAN W, REN Z Y, ZHAO X, et al. Probabilistic photovoltaic power modeling based on nonparametric kernel density estimation[J]. Automation of Electric Power Systems, 2013, 37(10):35-40. (in Chinese)
[2] KOUROUNIOTI I, POLYDOROPOULOU A, TSIKLIDIS C . Development of models predicting dwell time of import containers in port container terminals – an artificial neural networks application[J]. Transportation Research Procedia, 2016, 14:243-252.
[3] 杨龙海,章锡俏,李社教,等.深圳盐田港保税区货运交通需求预测[J].城市交通,2007,5(2):51-54. YANG H L, ZHANG X Q, LI S J, et al. The freight traffic demand forecasting of Shenzhen Yantian Harbor Free Trade Zone [J]. Urban Transport of China, 2007,5(2):51-54.(in Chinese)
[4] BOER C A, VERBRAECK A, DE A, et al. Distributed E-services for road container transport simulation [C]//Proceedings 15th European Simulation Symposium. 2003:541-550.
[5] 滕藤, 曾庆成, 陈文浩. 集装箱码头集卡到达时间预测模型[J]. 武汉理工大学学报(交通科学与工程版), 2015, 39(6):1221-1224.TENG T, ZENG Q C, CHEN W H. Prediction model for truck arrival time in container terminals[J]. Journal of Wuhan University of Technology (Transportation Science & Engineering) , 2015, 39(6):1221-1224. (in Chinese)
[6] 张莎莎, 庞明宝, 杜宏亮, 等. 基于船期表的一种集装箱码头闸口交通需求智能预判方法[J]. 交通科技, 2014 (5):148-151.ZHANG S S, PANG M B, DU H L, et al. An intelligent traffic demand forecasting method for container terminal based on line schedule [J]. Transportation Science & Technology, 2014 (5):148-151. (in Chinese)
[7] 陈家源, 郑仁环. 港口通过能力系统理论研究——港口交通流的分布与生产状态的分析方法[J]. 大连海事大学学报, 1993,19(3):297-309.CHEN J Y, ZHENG R H. System theories research of the port’s throughout capacity[J]. Journal of Dalian Maritime University, 1993,19(3):297-309. (in Chinese)
[8] 杨忠振, 陈刚, 陈康. 基于码头作业形态的港区道路交通需求预测[J]. 大连海事大学学报, 2009, 35(4):35-38.YANG Z Z, CHEN G, CHEN K. Traffic demand forecasting model for large port based on terminal operation situation[J]. Journal of Dalian Maritime University, 2009, 35(4):35-38. (in Chinese)
[9] YANG Z, CHEN G, MOODIE D R. Modeling road traffic demand of container consolidation in a Chinese port terminal[J]. Journal of Transportation Engineering, 2010, 136(10):881-886.
[10] 郭子坚, 方泽兴, 王文渊,等. 集装箱港区集疏港货运交通随机时变特性仿真研究[J]. 港工技术, 2017, 54(5):11-15.GUO Z J, FANG Z X, WANG W Y, et al. Simulation research on freightage changing randomly with time in container harbor[J]. Port Engineering Technology, 2017, 54(5):11-15. (in Chinese)
[11] 薛留根. 应用非参数统计[M]. 北京:科学出版社, 2013.
[12] 孙志华. 非参数与半参数统计[M]. 北京:清华大学出版社, 2016.
[13] PARZEN E. On estimation of a probability density function and mode[J]. The Annals of Mathematical Statistics,1962, 33(3):1065-1076.
[14] 徐玉琴, 张扬, 戴志辉. 基于非参数核密度估计和Copula函数的配电网供电可靠性预测[J]. 华北电力大学学报, 2017, 44(6):14-19.XU Y Q, ZHANG Y, DAI Z H. Power supply reliability prediction of distribution network based on nonparametric kernel density estimation and copula function[J]. Journal of North China Electric Power University, 2017, 44(6):14-19. (in Chinese)
[15] RUDEMO M. Empirical choice of histograms and kernel density estimators[J]. Scandinavian Journal of Statistics, 1982, 9(2):65-78.
[16] BOWMAN A W. An alternative method of cross-validation for the smoothing of density estimates[J]. Biometrika, 1984, 71(2):353-360.
[17] QIN Z, LI W, XIONG X. Estimating wind speed probability distribution using kernel density method[J]. Electric Power Systems Research, 2011, 81(12):2139-2146.
[18] 颜伟, 任洲洋, 赵霞,等. 光伏电源输出功率的非参数核密度估计模型[J]. 电力系统自动化, 2013, 37(10):35-40.YAN W, REN Z Y, ZHAO X, et al. Probabilistic photovoltaic power modeling based on nonparametric kernel density estimation[J]. Automation of Electric Power Systems, 2013, 37(10):35-40. (in Chinese)