- 學位論文
應用資料挖礦於瓶頸站排程的研究–以U公司為例
The study of data mining in bottleneck scheduling-a case study of U company
摘要
中國最近這些年快速崛起與低成本優勢，已對印刷電路板產業帶來巨大的影響，低成本競爭對於大多數台灣業者來說都不是強項，2008年開始的金融海嘯，接著2009年全世界電子消費市場的緊縮，都是對台灣之印電路板業廠商生產管理能力的考驗。 現在的企業環境不像以往單純，過去的企業有穩定的訂單來源，穩定的交期以及擁有充足的物料，然而，在此刻取而代之的是更複雜的生產環境以及競爭更激烈的市場。因此，如何在競爭日趨激烈的時代，讓企業更有效率的運作，是企業主管應該優先考量的。 本文運用資料挖礦中的倒傳遞類神經網路模型，分析印刷電路板製程中的瓶頸站與各工作站的在製品數量兩者的關聯性。除了利用分析所得之關聯性預測各工作站短期的在製品數量，藉以此執行及時監控整體生產是否有異常，另外可加入每月出貨需求的考量，訂定瓶頸站各產品的配置與產出供排程參考。
並列摘要
In recent years China rapid-rise and low-cost advantages, has made the greatest influence on the printed circuit board industry, low-cost competition for the most of Taiwanese companies are weakness, the financial tsunami began in 2008, followed by 2009, the consumer electronics market tightening, are tested for the production and management of printed circuit board industry in Taiwan . Recently, the business environment becomes more complex with the past business environment. There were stable source of orders, delivery period, and sufficient materials. However, the situation becomes more complex and the consumer market is more competitive. Therefore, it is important that a company how to be more efficient and stronger in this competitive environment. The study applied back propagation neural network model of data mining to analyze the relationship between bottleneck station and the WIP（work in process）quantity of else stations in printed circuit board industry. Besides predicting the WIP quantity by this model in short term for monitoring whether any problems in production process in time, we also consider shipping requirement to allocate the WIP and production in bottleneck station for scheduling reference.