Study Of Shipping Cargo Flux Predicting For Northeast China

A cargo flux predicting model based on RBF Neural Network is presented in this thesis, and the predicting results for Dalian port and Northeast China cargo throughput are gained. A forecasting result extent is obtained by using Monte Carlo.Lots of scholars domestically and abroad predict cargo throughput by using time series models, grey models and regression models. Because the RBF Neural Network has ability of self-organizing, self-learning, self-adapting and association, so it is properly to use it in analysis of cargo throughput predicting, but the application examples of RBF in cargo throughput area are really rare. The author's tasks in this thesis are building RBF Network cargo throughput predicting model, analyzing the accuracy of this model with cargo throughput history data and finally predicting cargo throughput for Dalian port and Northeast China from 2006 to 2010, plus 2020 single year. In addition, another task in this thesis is to analyze the credit of predicting result by using Monte Carlo, finally the probabilities of predicting result in every result extent are gained.The predicting experiment indicates that it is proper to use RBF Network in cargo throughput predicting. Conclusions are that the probability of 2010 Dalian port cargo throughput between 2710.2 and 2782.4 million tons is 0.6433, the probability of 2020 Dalian port cargo throughput between 3588.9 and 3661.9 million tons is 0.8133, the probability of 2010 Northeast China cargo throughput between 7111.5 and 7666.0 million tons is 0.7567 and the probability of 2020 Northeast China cargo throughput between 21097 and 22595 million tons is 0.85.The author hopes that the result can contribute to the port projects planning.