Research On Statistical Classification Model On Demand Patterns Of Rarely-used Spare Parts

The thesis has provided a practical analysis on the demand patterns of the spare parts in the Guangdong Nuclear Power Station (GNPS). This nuclear power station is facing the inventory problems with a large number of rarely-used key items. The empirical analysis aims to identify the usefulness of models put forward in the academic literature and motivates new ideas on some of the subjects.The study begins with the description of the inventory characteristics of the GNPS, including the features of the spare part items and its inventory control process. Then detailed analysis on demand pattern is carried on by splitting it into two components: demand size and interval between occurrences of demand.One contribution of this paper is that an improved model for demand pattern classification has been developed based on the literatures forward. Intermittence, irregularity and magnitude are considered into a three-dimensional model to classify the demand pattern.To assess the empirical utility of my theoretical findings, an empirical analysis needs to be done specified for simulation purposes. The data set is derived from the inventory records of the GNPS. First the demand data is observed from both a macroscopic and microscopic point of view. Then, one item is selected to explain how the three-dimensional construction works and to assess the classification model put forward in0 the academic literature. Finally, some discussion on parameter selection is made according to the demand pattern.