Research On Judgment Model And Demand Forecasting Of Slow Moving Spare Parts

Inventory control of slow moving spare parts is one of the most important works in business management. Based on the good control of the slow moving spare parts, we can cut the cost of the inventory sharply and ensure the normal production. But the research on the slow moving spare parts is not enough, especially on the judgment model and forecasting. There are many works we have to do.According to the feature of the slow moving spare parts, we make a delimitation of it and propose a judgment model based on the Fuzzy Comprehensive Evaluation which can take into account both the fuzzification and the weight factors. Using this model we can abstract the slow moving spare parts from the normal spare parts more effective comparing with the old method. It builds a solid foundation for the management of the slow moving spare parts.Having a stiff price and high importance, slow moving spare part is the core of the forecast for device requirement. If the forecasting of the demand for the slow moving spare parts is accurately, not only can the enterprises ensure a smooth production, but also they can cut the cost of the inventory sharply. But the demand of the slow moving spare parts is often uncertain, so forecasting the demand accurately using a conventional method is very difficult. For making the forecasting more reasonable this paper use linear multiple regressions combined with pre-classification to forecast the requirement of spare parts. The pre-classification above is that before the forecast, the spare parts which have the similar principle, structure, working environment and requirement should be classified into one group according to their homology and succession. On this basis it can make a prediction using linear multiple regression. Then the Economic Order Quantity of spare parts could be obtained based on the result of prediction. The basis of this model is the factors which affect the demand of the slow moving spare parts. It's suitable to forecast the demand of the slow moving spare parts when the enterprise has a detailed maintenance record. But many of the enterprises do not have the detailed data for their slow moving spare parts. Most of them have the time series data of the spare parts demand. So we develop a BP neural network based on the time series data to forecast the demand of the slow moving spare parts.In this method, we divide the forecast into two parts. On one hand we forecast the occurrence of the demand for slow moving spare parts; On the other hand we forecast the quantity of the demand. At last, we combine the two parts to make a better and more available prediction for the future demand. This method not only reduces the difficulty of the forecast but also help us to find the potential rules of the demand for slow moving spare parts.To sum up, this paper make a research on the slow moving spare parts form two points, one occurrence, the other demand. The model proposed is high performance and can improve the inventory management of slow moving spare parts.