Study On Supplier Evaluation And Selection And Optimization Problems In Lean Supply Chain

In the environment of lean supply chain management, based on the newest research results of the field of artificial intelligence, this thesis conducted systematic and extensive studies on a supply chain management theories and applications. It has theoretical aspects and practical usefulness in scienticfic decision of lean supply chain management to modem enterprices in socialist market economics in modern China.The thesis established a new evalution factor system of supply chain partner based on lean supply chain management, put forward evaluation and selection method of supply chain partner based on Radial basis function artificial neural network (RBFN). Using definition sensitive index of supply chain partner evalution factor, this dissertation originally made sensitivity analysis of supply chain partner evalution factors based on RBFN, which are calculated by the partial derivatives of output variables with repect to the input variables. Models of supply chain partner number selection and decision model of oder distribution are jointly presented, which is based on uncertainty demands. Models of interval planning of supply chain partner dispersion center layout are put forward, which is based on uncertainty demands and uncertainty production capacity. A few examples were given to highlight the theories and models. The theoretical results are consistent with the practical results, which shows that the theory and models in this thesis are correct and useful in supply chain management. The main results are as follows:(1) A new evalution factor system of supply chain partner based on lean supply chain management is established. These evalution factors are selected using software SSPS, which include 26 factors such as quanlity and cost, etc. The thesis quantifies these evalution factors, which makes it very convinent in practical lean supply chain management.(2) This thesis presents the mathematic principles and application model of the radial basis function neural network (RBFN) method. An example is provided to illustrate the proper selection of supply chain partner. The results obtained from the RBFN network were also compared with the results obtained using BPN networks. It was found that the proposcd method is correct and can be particularly used in the Supply chain partner selection whose parameters have randomness and relevancy.(3) For the first time, RBFN method is extended to compute the sensitive index of supply chain partner evalution factors, which are calculated by the partial derivatives of output variables with repect to the input variables in RBFN. Then, the respective sensitivity of 15 factors are analyzed quantitively. The results are consistent with the practical analysis. The proposed new method will play great role in the correct selection of supply chain partners.(4) The impact of minimum supplier order quantities is examined. The results indicate that single sourcing is a dominant strategy only when supplier capacities are large relative to the product demand and when the firm does not obtain diversification benefits. In all other cases, we find that multiple sourcing is an optimal sourcing strategy. This thesis also characterize a non-intuitive trade-off between supplier minimum order quantities, costs, and supplier reliabilities. Finally, an extensive numerical analysis of the key parameters of our model is conducted.