Truck Arrival Volume Prediction Method Research Of Container Terminals

With the fiercer competition among the ports,the container terminals put more and more attentions to the increase of its services ability,and provide the convenient services to the users to increase the attractive ability of the containers.Now a lot of ports use the concentrated arrival mode for the trucks which carry the export containers to the terminal.And it has many disadvantage compared with the random arrival strategies of export containers,such as the difficulty of the transportation arrangement of the truck fleets and the truck congestion at the port gates and the great transportation pressure of the port area.But with the random arrival mode of the export containers,the truck arrival would be more random,and it would bring a lot of troubles to the making of the operation plans,and because of the random of the trucks arrival,it is hard to leave the suitable space in the yard for the container to stock.So studying the truck arrival patterns and rules and predicting the truck arrival volume is very important and could work as the base of the resource allocation.And in this paper,based on the analysis of the historical data,we study the rule of the truck arrival and dig relations of the influential factors of the truck arrival,and find the influential degree of each factor by ANOVA.Besides,the machine learning is used to predict the truck arrival volume of a single container ship in each time period,and then based on the DBN to get the characteristics of them,and regards the new factors as the input factors of the machine learning method.And in the paper,based on the real dada of the Jinzhou port and Tianjin port,it proves the model we propose could have a better predict effect by the DBN&SVM.And according to the predict result,it is the most suitable method compared with the predict methods we motioned.