Application of BP neural network in logistics forecasting

The cost of logistics is a dominant part in total cost of an enterprise. In this paper, the forecast model of BP neural network is established according to the historical data of purchasing quantity, market supply, volume of exports and market price for some material supply in an enterprise. The network has three input vectors and three neural layers. The historical data of material supply is used as the training samples of BP neural network and then the market price of material is forecasted using the BP neural network. The simulation results indicate that the model can forecast accurately the price of material under different market situation conditions. Furthermore, the logistics forecasting based on BP neural network is an effective prediction method for the purchase decision for enterprises.