Font Size: a A A

Strategy Of Inventory Control Based On Improved Algorithm Of BP Neural Networks

Posted on:2004-06-22Degree:MasterType:Thesis
Country:ChinaCandidate:W J HuangFull Text:PDF
GTID:2156360122475522Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
The competition of market brings new management ideas, and boost study and application of more advantage and more practical management methods. Under the market environment that based on time and consumer requirement, the inventory management is the key to improve the customers' satisfaction and lower the inventory cost. It is a bold attempt to apply the BP neural networks technology to inventory management. To eliminate the enterprises' high inventory and improve the customers' service, this paper discussed the new idea and method that manage the inventory by using the BP neural networks technology.To modeling the inventory, three aspects of net training were prepared: sample data, improved algorithm and net structure. With the background of CQDP hospital inventory, the main factors that affected 17GY Liuzhi aces' using quantity in CQDP hospital inventory were confirmed. The data mining technology was adopted. The original data was preprocessed, including data clearing, data integration and data transforming. Then the sample data that used in net training was obtained; the principle and disadvantage of BP neural networks was introduced. In order to improve the performance of BP neural networks, the network weight matrix W and the parameters T ,θ of neuron's tan-sigmoid transfer function are adjusted. Information is stored in the weight matrix and the transfer functions dispersedly. The improved algorithm has stronger nonlinear mapping capability than traditional algorithm. The final formula of the improved algorithm is presented after rigorous mathematical deducing. An example of function adaptation is presented. The original structure of BP neural networks was constructed. Through repeat net training, the optimal number of hidden layer node and learning rate were found based on training set and test set's mean square error. Then the optimal structure of net was ascertained. After training of the net, the final inventory model was obtained. The demand value was forecasted, and compared with two traditional methods' forecast value (regression analysis and EOQ model). The sensitivity of model was processed, and the conclusion was obtained. Finally, it is programmed that the computation program for the improved BP algorithm with C++ computer language, which can be used in many industries and is good to decision-making of inventory management.
Keywords/Search Tags:BP neural networks, data mining, improved algorithm, inventory control, demand forecast
PDF Full Text Request
Related items