本研究旨在探討集群分析在縮短記憶體模組換線時間的應用，並以A公司為研究對象。A公司長期深耕記憶體模組研發與製造，成功在全球打出自有品牌，成為全球前十大記憶體模組製造商。由於記憶體模組的種類繁多，市場需求的轉變快速，訂單模式由大量生產轉為少量多樣，造成生產換線次數增加，製造成本提高。本研究利用集群分析的方式將物料相似度高的產品分群，透過集中生產同一群的產品，並觀察減少換線時間。首先用集群分析中的華德法找出A公司產品料號最佳集群數，再以K-means法將料號分群，經由PDCA逐步執行生產驗證，換線時間明顯減少，生產效率也有提升。本研究也找出了一些關鍵因素及條件，這些因素和條件對於生產線的運作非常重要，可以試圖應用在其它產品線上，提升整體生產線的效率。總體而言，本研究採用集群分析配合PDCA方法，在產品需求變化的情況下，成功地改善了生產線換線時間和生產成本的問題。本研究提供了實證研究，可以作為其他企業在生產線排程上進行決策和改善的實務參考。 ;This study aims to investigate the application of cluster analysis in reducing changeover time for memory module production, focusing on Company A as the research subject. Company A has long been dedicated to memory module research and manufacturing and has successfully established its own brand, becoming one of the top ten memory module manufacturers globally. With the wide variety of memory module types and the rapid changes in market demand, the production order pattern has shifted from mass production to small-batch and diverse production, resulting in an increase in changeover frequency and higher manufacturing costs. This study utilizes cluster analysis to group products with high material similarity and observes the reduction in changeover time through concentrated production within each group. Initially, the Ward’s method in cluster analysis is employed to determine the optimal number of clusters for Company A′s product part numbers, followed by grouping the part numbers using the K-means method. Through the step-by-step implementation of production validation using the PDCA approach, the study demonstrates a significant reduction in changeover time and an improvement in production efficiency. The study also identifies several key factors and conditions that are crucial for the operation of the production line. These factors and conditions can be explored for application in other product lines to enhance overall production line efficiency. In conclusion, this study successfully addresses the issues of changeover time and production costs on the production line by employing cluster analysis in conjunction with the PDCA method. The empirical research provided in this study can serve as practical references for other enterprises in making decisions and improvements in production line scheduling.
