The Study On Inventory Classification Model And Its Aplication Based On D-S Evidence Theory And Grey Relational Analysis

With the development of computer technology applications, inventory management information system for inventory management has become an important tool for the work. In this highly integrated, the information age, how to use inventory data resources of management information system for inventory management effectively is a practical research significance of the subject. In this research, the first problem to be solved is the inventory classification problem. Inventory classification is the basis of inventory management. Therefore, based on the analysis of the problems for the current inventory classification, combined with domestic and foreign-related research in the field, this thesis proposed the study on inventory classification and its application based on D-S evidence theory and grey relational analysis .The research results of this thesis are as follows:First, it was described on the nature, procedures and framework for the classification analysis based on D-S evidence reasoning and classification of gray relational analysis principles. Secondly, in accordance with the gray system theory and evidence theory, it built the mold of the inventory classification based on D-S theory of evidence and grey relational analysis, and further study of the observed sample-related / conflict of evidence under the relevant / conflicting metrics and synthesis rules. Finally, took the spare parts of infrastructure projects for a the empirical analysis of the object, it found that the proposed method can verify the feasibility on the basis of classification results obtained this method compared the similarities with the results of other method of classification, and can be applied to the classification of business inventories.In the specific application process of this method, the observed sample size, resolution ratio, data transformation methods, standard reference series and the evidence of independence affected the model output as a certain extent. along with the increasing the number of evidence samples, it can significantly improve the discrimination of the model output; it could got more ideal model output when resolution ratio selected between 0 and 1 (generally choose 0.5), and data transformation method selected other ways in addition to the range of data transformation methods; standard reference series is the key factors to determine the model output, and it must be carefully chosen; and evidence related / conflict affected the accuracy of the model output as a certain extent, it could choose what kind of evidence combination rules according to the scope of its robustness by selecting the appropriate threshold to ensure the accuracy of the model output under the premise of the model to minimize the amount of computation in practical applications deal with conflict or the amount of relevant evidence measure.