規則式專家系統求解多廠區半導體工業生產計劃問題

NCU Institutional Repository > 管理學院 > 企業管理研究所 > 博碩士論文 > Item 987654321/72893

請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/72893

題名:	規則式專家系統求解多廠區半導體工業生產計劃問題
作者:	何宗穎;Ho, Tsung-Ying
貢獻者:	企業管理學系
關鍵詞:	生產計劃;產銷協同;供應鏈管理;Production planning;Sales and operation planning;Supply chain management
日期:	2017-01-11
上傳時間:	2017-05-05 17:15:13 (UTC+8)
出版者:	國立中央大學
摘要:	半導體製造是為高度資本與知識密集之產業，而半導體最終產品之生命週期大幅縮短，廠商無不希望保持產能利用率同時維持既定的生產週期，如期交貨以滿足市場瞬息萬變之需求。企業資源規劃系統 (Enterprise resource planning system) 應用在半導體工業來整合協調各個部門的資訊外，基於ERP系統之限制，我們更需要一個決策模型，來協助管理者達成生產與銷售協同之目標，以提昇顧客滿意度以及獲利。現有之規劃系統僅考量了原物料需求滿足生產所需之情況，並未考量到生產產能限制、多個廠區、工作中心之指派問題，也無法進行生產規劃模擬，或者考量其他生產限制進行規劃調整，因此相關的調整生產規劃之邏輯仍需要一套模型或者系統。學者提出先進規劃系統 (Advanced planning system) 之概念，將可針對生產規劃問題之結構模組化，並可以套用相對應的資訊軟體，與ERP系統的資訊整合來解決上述之問題。本研究提出了一套以APS為架構之生產規劃模型，考量半導體製造前、後製程產能限制、生產產能、良率、生產週期時間、交期限制來解決上述的問題。在管理者進行總體規劃時，能夠納入考量整個供應鏈之能力，以維持較低之庫存，降低公司資金積壓造成之營運風險。本研究採用數值驗證，驗證模型與規則式專家系統演算法之可行性，並與最佳化軟體Lingo之規劃結果進行比較，證實本研究提出之模型與解法能夠維持較低之庫存，具有一定的求解品質。本研究最後針對模型與解法考量不周之處進行說明，期待能使未來之研究更臻完善，使半導體產業之管理者能夠提升管理績效。關鍵字 : 生產計劃、產銷協同、供應鏈管理。 ;Semiconductor manufacture is an industry consisting of high capital and knowledge. Maintaining high utilization rate is persuaded by companies. Meanwhile, companies are eager to meet due date constraint. Enterprise resource planning (ERP) system is utilized in the semiconductor manufacture industry to integrate information. Owing to the system limitation, decision models are needed to achieve the balance of sales and operation. However, the existed ERP system has encountered certain drawbacks. We divided the problems into two parts. First, the system ignored not only the production limitation but multi plants、multi work center dispatch problem and so on. Second, it could neither stimulate the production plan nor adjust it. Stadtler (2005) stated the concept of advanced planning system (APS) had exhibited an architecture based on the principles of hierarchical planning. The main focus is on supporting the material flow amidst a supply chain and is related to business functions: procurement, production, transport and distribution as well as sales. This research provides a production planning model which integrates front-end and back-end production plan based on the capacity. The whole capacity of supply chain can be considered by managers when making aggregate level planning. Moreover, companies can maintain lower inventory level, and also avoid the operation risk of backlog holding cost. Besides, this research offers numerical experiment to verify the model and Rule-based expert algorithm. We compare the result that we got from algorithm and the one out of optimization software Lingo. At last, we found out that two methods have their advantages respectively. Using the optimization software can reach the lower inventory level, whilst our algorithm is able to gain feasible solution. In the end of the research, we illustrate the points which we did not take into consideration. Nevertheless, in future researches, thinking over and solving all the problems to make the model and algorithm more practical are necessary jobs to do. Key words : Production planning, Sales and operation planning, Supply chain management.
顯示於類別:	[企業管理研究所] 博碩士論文

文件中的檔案:

檔案	描述	大小	格式	瀏覽次數
index.html		0Kb	HTML	353	檢視/開啟

在NCUIR中所有的資料項目都受到原著作權保護.

社群 sharing

資料載入中.....