The Research Of Logistics Enterprise Alliance Partner Selection Based On PCA-BP Neural Network

With the opening of China’s logistics market, more and more foreign logistics giantinvestment into the domestic logistics market, which brings a huge impact to the developmentof China’s logistics industry, especially that China’s logistics market are mostly small andmedium-sized logistics enterprises, in the face of fierce market competition at home andabroad, it seems difficult to survive and develop. So corporate alliance becomes an effectiveway to improve the competitiveness of small and medium-sized logistics enterprisesvulnerable, and to strengthen the organization of our country’s logistics market, and the choiceof alliance partners is the most important and critical step in the formation of alliances.This article is based on the above background, describes the current development of thelogistics industry, the current development of domestic and international logistics alliance,and proposed that logistics alliance is the inevitable choice for the development of small andmedium-sized logistics enterprises in china. On the basis of the in-depth analysis to the theoryof logistics enterprise alliance, the article proposed that the most suit alliance model for chinalogistics market are leader-member type and federal mode. Then, combined with analysis ofthe previous evaluation system, as well as logistics alliance partner selection factors, weestablish an evaluation index system for the choice of logistics alliance partner. And aftercomparing the pros and cons of a variety of evaluation methods, this article introduce theprincipal component analysis combining BP neural network, using principal componentsanalysis to simplify the evaluation data,while using neural network itself nonlinear mapping,adaptive learning and associative memory properties to train the samples,in order to get aneffective intrinsic link between expert knowledge and indicators, and reduce the subjectiveevaluation of human factors, and finally establish a practical and effective and sustainableimprovements in the use of the evaluation model. The implementation of the model using theSPSS software and Matlab tool, to analyze the data and train BP neural networks. But theneural networks algorithm is very complex, the programming of modeling is workload, thearticle comes directly use of the Matlab neural network software Toolbox, which can simply and intuitively complete the BP neural network model’s establishment and training. Finally,the article take Zhejiang YL logistics alliance as an example to test the model, select thelogistics enterprise alliance partners and provide a practical method for operationalevaluation.