【作者】 叶剑；

【导师】 宋向群； 郭子坚；

【作者基本信息】 大连理工大学 ， 港口、海岸及近海工程， 2005， 硕士

【摘要】 随着世界经济与贸易的发展,港口集装箱吞吐量不仅是衡量一个港口在国际经济贸易中地位的重要标志,也是一个国家与地区经济繁荣程度的晴雨表。研究各种港口集装箱吞吐量预测方法的选择理论,对于正确地预测未来港口集装箱吞吐量,指导港口的规划建设,确定港口投资规模和促进地区经济的发展具有极其重要的战略意义。但是长期以来对各种港口集装箱吞吐量预测方法展开的研究,大都集中在对具体的预测方法的改进上,却忽视了对方法本身的选择理论的研究,特别是忽视了不同地区港口运量增长规律的差异性,因而造成了港口集装箱吞吐量预测的不准确性。本论文提出了一种建立在对港口进行类型化分析的基础上对港口的集装箱吞吐量进行预测的方法。本研究着重分析了时间系列模型、灰色系列模型、回归模型和RBF神经网络模型四大类预测模型的特点,将灰色马尔科夫链模型、灰色傅利叶模型、时序残差修正模型、灰色优化模型和灰色非线性模型引入到港口集装箱吞吐量预测中。根据各港口集装箱运量增长的特点,应用基于聚类分析的类型化方法将我国沿海主要集装箱港口分成了三大类型,即普通增长型、加速增长型和波动增长型三大类,并对各类型港口增长特点进行了分析。最后应用时间系列模型、灰色系列模型、回归模型和RBF神经网络模型四大类预测模型,对普通增长型、加速增长型和波动增长型三大类港口分别进行了预测,通过对预测结果的比较和分析,归纳总结出了各类港口相对适用的集装箱吞吐量的预测方法。更多还原

【Abstract】 With the development of the world economy and trade, the port container handling becomes not only an important sign of judging the station of a port in the international economy and trade, but also a rain glass of the flourishing degree in economy of a country or an area Researching the methods to forecast the container hanging and doing the prediction properly are very important to guide the programming and construction of the ports, to confirm the investment scale of the ports and moreover to accelerate the development of the areas. However, for a long time, the research on the methods to forecast the container handling of port mainly focus on the amelioration of the idiographic forecasting methods, neglecting the otherness of the increasing rule of the port’s throughput in different areas, which leads to the port’s forecasting error.This paper brings forward a new research method of port container throughput based on the port’s cluster analysis. First of all, in the paper, the author introduces the basic theories and models, and especially does the systemic research on time series models, grey models, regression models and RBF neural network model, and moreover brings the grey Markov-chain residual modification model, the grey Fourier residual modification model, the grey model on time series error corrected, GOM model and grey nonlinear model to the port container throughput forecasting for the first time. And then, based on the characteristic of the port’s container increasing, the clustering analysis method is applied to classify the major seaports of our country into three types: the normally increasing port, the accelerative increasing port and the fluctuating increasing port. The characteristic of the port’s container increasing is also analyzed. Finally, the author applies the time series models, grey models, regression model and RBF neural network model to forecast the container throughput of the three type ports, and gets the conclusion of applicable forecasting methods for special container-port through the comparison and analysis.更多还原