張淳智林伶嬑2025-08-282025-08-282012-06-08U0061-1207201213584600https://www.airitilibrary.com/Publication/alDetailedMesh1?DocID=U0061-1207201213584600https://nutcir-lib.nutc.edu.tw/handle/123456789/1171為因應近年經濟動盪不安，無法快速成長的情況下，高品質、低成本與短期交貨已是企業生存的基本條件，為了壓低成本，多數會在庫存可容許範圍內量產。有別於一般產品，在探討間歇性需求特性的機車備用零件問題時，其需求量及間隔相當不穩定，為維持一定服務水準的前提下，企業需負擔很高的庫存及損壞風險，因此準確的備用零件需求預測模式顯得格外重要。 許多學者針對此類議題，除了使用不同方法建構模式以降低誤差外，也嘗試假設常態分配配適備用零件壽齡曲線，但備用零件種類及影響因素很多，部分重要參數不易取得，因此僅探討常態壽齡模式是不足的。 本研究在國內首度嘗試以不同分配配適機車備用零件壽齡曲線，以對數常態及其鏡射翻轉分配建構需求預測模式。根據實證結果，部分零件因配適不同曲線分配，可得到較優異的結果，改善率最高達42.99%，表示壽齡曲線有探討的必要性，當壽齡曲線貼近實際狀況時，本研究模式的正廠零件使用率的參數校估值也更符合實務運作狀況，在提供業者制定策略及備料參考時也具更高的可信度。In response to recent economic turmoil and can’t fast-growing, high-quality, low-cost and short-term delivery are business survival basic conditions, in order to lower the costs, mostly in the inventory to the extent permitted mass production. Unlike ordinary products, the characteristics of motorcycle spare parts is intermittent demand, the demand for and intervals are rather unstable, companies need to afford to maintain a certain standard of service under the premise of high inventory and the risk of damage, accurate spare parts demand forecasting model is particularly important. Many scholars for such issue, in addition to using different methods construction mode to reduce the error, also try to assumption normal distribution fit the spare parts product life curve. There are many types of spare parts and the impact factors, some of the important parameters is not easy to obtain, therefore only explore the normal product life curve model is inadequate. In this study, the first attempt to fit motorcycle spare parts product life curve with different distributions, using lognormal and mirror lognormal distribution to construct the demand forecasting model. According to the empirical results, our models decrease the MSE up to 42.99% in some spare parts. It indicates the product life curve is not always the Normal distribution. When the product life curve is more close to the actual situation, the estimated value of utilization of genuine parts will be more in line with the practice. It can provide enterprise to develop strategies and prepare inventory with higher confidence.zh需求預測備用零件壽齡分配馬可夫鏈蒙地卡羅Demand ForecastingSpare PartsProduct Life DistributionMarkov Chain Monte Carlomaster thesis