Research On The Prediction Model Of Seaport Container Throughput

The seaport container throughput is influenced by many factors such as market, economy and non-economy fackors. So it is a complicated nonlinear system. Meanwhile there is some inner rule over the throughput variation in the system. China's foreign trade and container transportation is continuing the trend of rapid growth along with the high development of country's economy. The grey theory can be well used to forecast the increased time series because of its own characteristics. However, some times, the forecasting result of grey theory is not satisfied. So, in the paper, the Artificial Neural Network is used to amended the model of GM(1,1) in order to reach the high forecasting precision.In the paper, the forecast model of remaining amended GM(1,1) based on Artifical Neural Network is proposed by the analysis of the componment of seaport container throughput, and then the model is used to forecast the seaport container throughput so as to confirm the validity of the forecast model.Firstly, the paper gives a complete introduction of the recent forecasting methods and the studies nowadays of container throughput home and abroad. In addition, it analyzes the elements that influence the container throughput and points out the seaport container thoughput that is a nonlinear system.Secondly, in the paper, it states the characteristics of the model of GM(1,1) and Artificial Neural Network. Consquently, the algorithm of remaining amended GM(1,1) based on BP Artificial Neural Network is proposed because of the defects of the model of GM(1,1).Thirdly, the forecast model of remaining amended GM-NN(1,1) is proposed on the based of nonlinear thimer series forecasting throry and the algorithm of remaining amended GM(1,1) based on Artificial Neural Network. Then the seaport container throughput forecasting system is designed and developed. Meanwhile it discusses the paraments of the forecast model.Finally, in the paper, it confirms the paraments of the forecast model. Then the validity of the forecast model is tested by the Dalian port container throughput forecast experiment. The Dalian port container throughputs in 2006-2010 are also forecasted by using the forecasting system. And then it gives a brief analysis upon the forecasting results.