Font Size: a A A

Study On Intelligent Prediction For A Container Terminal Gate

Posted on:2016-10-26Degree:MasterType:Thesis
Country:ChinaCandidate:S S ZhangFull Text:PDF
GTID:2272330479498991Subject:Traffic and Transportation Engineering
Abstract/Summary:PDF Full Text Request
Container terminal gate is the inevitable corridor for the container trucks to get in and out a terminal, the key node and bottleneck for the operation of port collecting and distributing system. Its irrational setting will not only makes customers to wait for a long time in a queue, but also causes some severe jamming problem in port area’ roadway because of the long queue length and great traffic volume pressure. In the meaning time,these terminal enterprises feel great pressure on their transportation operation, i.e., a more reasonable arrangement problem of transport organization in order to avoid seriously imbalance of operation and huge economic losses because of demand uncertainty.Therefore, the problem of predicting the traffic volume for a container terminal gate is extremely significance in both theoretical study and practical application. At the same time,although traffic demand of container terminal gate has nonlinear characteristics, there are some regularity and inner randomness of transport organization especially the gate traffic demand with sailing schedule. How to make use of modern information intelligence technology and high nonlinear characteristics to predict the traffic demand of gate is the starting point of the study. The main work in this paper could be summarized as the following aspects:(1) The collection and distribution traffic system in the port area was analyzed. Based on the discussion of system fundamental factors, system environment and traffic characteristics, the intelligent prediction system of terminal gate was designed including information subsystem, database, knowledge base, off-line predicting subsystem and on-line predicting subsystem. The functions and theories of these subsystems and the intelligent prediction system were explained. It provided a basic platform and application possibility to study the traffic volume intelligent prediction of container terminal gate.(2) The predicting method of traffic demand on a container terminal gate was studied by using on curve fitting method and support vector machine(SVM). A probability distribution model was set up based on historical information and probability distribution fitting method. The total number of container truck in each boat by using SVM. The number of container truck can be got in each time period and it was verified by concrete examples.(3) The predicting method of traffic demand on a container terminal gate was studied using real-time information. A probability correction predicting model with real timeinformation was proposed based on the probability distribution by using probability distribution fitting method. The concrete testing example shows its high precision by using the method.(4) The predicting method of traffic demand on a container terminal gate was studied by using on seasonal artificial neural networks method. Based on the seasonal and nonlinear characteristics of traffic demand on a container terminal gate originating from sailing schedule, the thought was proposed to predict the traffic demand of terminal gate based on sailing schedule including each scheduling line. The traffic volume for container truck arriving at terminal gate was transferred into the normal time series data by using seasonal time series method. A neural network model, which maps the nonlinear between the transferred time series data and the predicted traffic volume, was established. The results of the concrete example on a container terminal gate in Tianjin Port show that the method is prior than that probability distribution fitting method and probability correction predicting method based on real-time information. The feasibility of the method was proved.
Keywords/Search Tags:intelligent transportation system(ITS), container terminal, gate traffic demand, sailing schedule, seasonal time series, artificial neural networks(ANN)
PDF Full Text Request
Related items