Research And Implemention Of Auto Parts Demand Prediction System Based On BA-BP Algorithm

With the rapid development of auto industry,enterprises face more complex environment and stronger competitors.Therefore,car companies not only need to raise their manufacturing technology,but also need to improve the quality of after-sales service.Want to be more stable and better after-sales service,auto parts inventory is relatively increased,which brings higher cost and more risks.So the effective and accurate auto parts demand forecasting can efficiently reduce the inventory cost,which can also improve the service quality and gain greater benefits.First of all,this thesis is committed to design and realize a reasonable demand forecasting system of auto parts.Taking the manufacturer as the core,based on the study of automobile parts supply chain and main business processes,to find the key factors that affecting the demand of automobile parts.According to the actual research demand and the comparative analysis of forecasting programs of automobile parts at home and abroad,we select BA(BAT Algorithm)to optimize the BP neural network to establish the prediction model and the validity of the model is verified by the training and test results of the model.Secondly,the basis of the predict is data,while the traditional one is only based on local data to predict and ignore the mutual influence of each node in the whole parts supply chain,which often results in inaccurate demand prediction.The system utilizes JAX-WS framework to develop WebService,combined with RSA encryption algorithm to achieve the data safely exchange between system and enterprise,and also to improve the accuracy of model prediction by using the integrated data to carry out model training.Finally,according to the demand of the system realization of the System Management Module,Basic Data Query Module,Data Exchange Module,Parts Prediction And Query Module,Model Control Module,and complete the function test and performance test of the system,to verify the accuracy and stability of the system.In this thesis,the combination of data exchange technology,prediction technology and actual demand improves the prediction accuracy and feasibility,which provides users with a simple,clear visual interface and multi-dimensional graphics display,and users can get a more intuitive understanding of data that eases their later demand decision-making help.