Research On The Method Of Supplier Evaluation For Project Purchasing

Purchasing management is very important to project management and company operation. Supplier evaluation and selection is the most significant activity of purchasing management. Program flawless launch and business success in the competition depend on selecting a good supplier.This thesis analyzes the methods on the supplier evaluation based on BP Neural network. First step is to create the integrated system of supplier evaluation indexes by comparing some supplier evaluation checklists from automobile companies. Second one is to calculate the weight on each index by the method of AHP. Third step is to create the supplier evaluation model based on the BP neural network and take the Levenberg-Marquardt optimization algorithm as training method, which could cause the rapid convergence. Finally, the trained model is verified by test samples.Supplier evaluation is a multi-objective decision problem. The research shows the BP evaluation model does benefit to define the weight of network with diversity and variability. The BP neural network based model could provide an operational evaluation method for supplier selection.