Freight Prediction Based on BP Neural Network Improved by Chaos Artificial Fish-Swarm Algorithm

Back propagation (BP) neural network has widely application because of its ability of self-studying, self-adapting and generalization. But there are some intrinsic defaults, such as low convergence speed, local extremes and so on. Artificial fish-swarm algorithm (AFSA) is an up-to-date proposed optimal strategy, which possesses good capability to avoid the local extremum and obtain the global extremum. In order to improve the search efficiency of AFSA, Chaos system is introduced. A quantitative forecast method based on the BP network improved by chaos artificial fish-swarm algorithm is proposed in the paper. The model is trained with the freight data of a city and then used to forecast the freight. Compared the simulated results with BP network and BP network improved by other algorithm, it concludes that CAFSA-BPN has smaller error in forecasting. And it indicates that CAFSA has the capability of fast learning the weight of network and globally search, and the training speed of the improved BP network is greatly raised.