DRP Decision Support System And WNN's Application In It

With the continuous developing of corporation, the amount of accumulated data was very large.Facing this situation, decision-maker need to know the important knowledge hiding behind the data to help decision supporting. Traditional DSS has some shortcomings such as lack of organizing and low using rate of data. Data Warehouse and Data Mining provide a new way for DSS.Data Warehouse can be the basic data source of decision supporting, and Data Mining is one of the tools of obtaining knowledge.They are both effective technique in decision supporting. It will greatly improve the quality and level of making decision of corporation to build a decision support system using Data Warehouse and Data Mining technique.This paper first introduces the theory of Data Warehouse and Data Mining as well as other relative technique, then discusses the decision support system based on Data Warehouse and Data Mining technique.On the basis of analysing the shortage of domestic DRP system and analysing Ningxia Xiajin dairy produce corporation's DRP system,this paper designs a DRP decision support system ,establishes a sale-oriented multidimensional data, evaluates the dealers of this company using classifying wavelet neural network model and predicts the sale of it using predicting wavelet neural network model. It provides strong support and reference on the selection of dealer and production and sale for this company. This paper analyses the BP algorithm. To overcome disadvantages of the traditional BP algorithm such as slow convergence speed and easily running into local optimum,two improved algorithms are implemented: a mixed algorithm which combines rapidly descent method with conjugate gradient and a BP algorithm based on genetic algorithm.Experiment results show that the improved algorithms are better than BP algorithm.The two improved algorithms show good effectiveness when used to train...