Multivariant Forecasting Mode of Guangdong Province Port throughput with Genetic Algorithms and Back Propagation Neural Network

Abstract

For better accurate forecasting of port throughput, a back propagation neural network model with genetic algorithms is proposed. By means of analysis of influence factors for port throughput, the structure of the BP neural network model is determined. Then the connection weight matrix of the BP network is designed for chromosomes of genetic algorithms, which is proved to optimize BP network. The port throughput of Guangdong province in China is used for verification, and the result of the experiment shows that GA-BP neural network model has better accuracy, but consumes more time than a traditional BP network model does.