| Summary: | 碩士 === 國立成功大學 === 工業與資訊管理學系碩士在職專班 === 104 === Although most enterprises have practiced Six Sigma Project for a while, the products that have about 0.3% defectives are still shipping to their customers. When defectives are
detected at customer side, the product returning processes will correspondingly activate. However, the operating costs rise because most enterprises have to prepare certain
amount of stock to return products in time to satisfy their customers. Accordingly, it becomes an important issue to find the breakeven point to both prevent the stock from being hoarded to futility, and to return products in time. The grey models are widely taken to make the next term predictor in short-term time series data; though, there still exists the possibility to make their predictions more precise. Therefore, this research reveals a new grey model to improve the accuracy of predictors by setting the coefficient sets of in traditional grey model. The technique this paper employs to find the sets is the mega-trend-diffusion, and the grey model thus named MTDGM(1,1). In the experiments, the examined dataset that contains eight product types was taken from the leading company in the TFT-LCD industry. The results compared with several improved grey models show that MTDGM(1,1) has more precise predictors.
|
|---|
