Research On NingBo Harbor's Ship Traffic Volume Forecasting

Ningbo port is the main transport station in our country, along with the development of shipping economy, the proving of shipping traffic volume rapidly contributes to development of economy and society, but the risk of shipping traffic is enhanced at the same time because of the proving. More requirements are being forward to programming design and traffic management in the port water area. The searching of shipping traffic forecast just provides the basic evidence in water area programming, design and shipping traffic management. So the searching of shipping traffic forecast in Ningbo port is very obligatory.On the base of deeply analysis and compare many kinds of methods of shipping traffic volume, the thesis studies the modeling of shipping traffic forecast using neural network. The first, introducing the common methods of forecast in the shipping traffic, comparing the characteristics of them, expatiate on the basic of theoretic and principium. Discussing the way how to build the model of BP and RBF neural network, the choice the number of conceal layers, confirming the value of center wide and so on, having built the BP and RBF neural network forecast models, and comparing of the method of multinomial forecast. The results of the test can prove: RBF neural network forecast has the smallest errors in the forecast of shipping traffic in Ningbo port, and the forecast value close to the value of fact. At last, the RBF NN model applies to the forecast of shipping traffic volume in Ningbo port.