Abstract: The selection of logistics supplier is an importantlink in supply chain management, while the assessment of capacity of logisticssupplier help client select logistics supplier in a better way. In view of theassessment and selection, this article proposes supplier assessment andselection based on support vector machine in special connection with smallsample event, which realizes the unification of assessment factors and results,improves the precision of selection and provides an effective method for theassessment of the capacity logistics supplier.
Keywords: Support VectorMachine, logistics supplier, assessment index
1. Introduction
With the sustaining economic development in china, themarket competition is increasingly globalized and intensive. To fit in with theneeds of the fast changing market, enterprises increasingly pay more attentionon the selection of suppliers. Logistics supplier is a cource that the productwas delivered from manufacturers to sellers. This course would not be able tobe completed by the manufactures, but by other logistics enterprise which iscalled logistics supplier. As an import force of logistics industry, its statusbecomes increasingly more important. The assessment and selection of supplieris not only a basis for the operation of supply chain, but also an import partof supply chain management. It has become an important task of an enterprise toassess the comprehensive performance of suppliers in an accurate and scientificway. There are many scientific assessment methods, such as experts' scoringmethod, linear programming, multi objective program, fuzzy integration method,analytic hierarchy process (AHP) and analytic network process (ANP). Theexpert’s scoring is full of human factors and lots of randomness. It is veryhard to create accurate mathematical model in linear programming and multiobjective program; though literature[1] selects logistics supplier by fuzzy integration method, it is hardto determine the weight of each index; the analytichierarchy process used in literature [2] can weaken human factors, however, itrequires that the factors in the hierarchy system of index be independent oneach other; though the network analysis method used in literature [3-4] canremedy some defects of the analytic hierarchy process, all the data used increating positivereciprocal matrixis determined by decision-makers based on theirsubjective awareness, thus there is still too strong subjectivity in the weightobtained; while there are many problems with the neural network method [5-6]such as the structure is hard todetermine, excessive study, insufficient study and localminimum. In view of this, this article proposes supplierassessment and selection based on support vector machine in special connectionwith small sample event.