完整後設資料紀錄
DC 欄位語言
dc.contributor.author馮堯欽en_US
dc.contributor.authorFeng, Yao-Chinen_US
dc.contributor.author蘇朝墩en_US
dc.contributor.author楊大和en_US
dc.contributor.authorSu, Chao-Tonen_US
dc.contributor.authorYang, Ta-Hoen_US
dc.date.accessioned2014-12-12T02:19:19Z-
dc.date.available2014-12-12T02:19:19Z-
dc.date.issued1997en_US
dc.identifier.urihttp://140.113.39.130/cdrfb3/record/nctu/#NT863031023en_US
dc.identifier.urihttp://hdl.handle.net/11536/63324-
dc.description.abstract由於導體製造機器設備昂貴,一般半導體製造廠的生管人員皆以提高機器利用率,追求系統最大產出為要目標,其次,由於系統平均在製品存貨影響晶圓製造廠能與生產週期時間,因此,控制系統平均在製品存貨水準亦是生管人員積極追求的目標之一。晶圓代工廠的製造特性是屬於訂單式生產(Make-to-Order)模式,在訂單式生產環境下,產品設料組合因訂單的不同而改變。產品組合的不同,直接影響到各機器設備的利用率與系統平均在製品存貨水準,甚至導致系統瓶頭的漂移,以至於生管人員無法控制系統的瓶頭,進而降低了半導體製造廠的生產績效。在本論文中,考慮晶圓製造廠生產狀況的隨機性,採用模擬方法以獲得動態的績效指標,再藉由倒傳遞類神經網路的學習能力,學習晶圓製造廠中的績效指標間的相互關係,以建構一快速反應產品投料組合與機器利用率和系統平均在製品存貨水準之關係的類神經網路模式,同時,亦建構一資料庫搜尋模式。透過此兩种模式的運用,使得生管人員能夠透過期望的設備利用率與系統平均在製品存貨水準的設定,尋求最佳的產品投料組合,以利即時的(Real-time)線上生產活動控制。本論文實例說明顯示,透過類神網路模式與資料庫搜尋模式,一則可控制機臺利用率與系統平均在製品存貨,二則可快速地決定適合且可行的產品投料組合。zh_TW
dc.description.abstractThe expensive nature of equipment associated with semiconductor manufacturing necessitates that a production planner should enhance equipment utilization and maximize the throughput. Also, average work-in-process (WIP) affects throughput and cycle time. The semiconductor foundry is a make-to-order production type. Under such production surroundings, product mix depends on the customer's order. Different product mixes directly affect equipment utilizqtion and average WIP level, causing a shift in the bottleneck. In this thesis, we consider the stochastic characteristics of productio in semiconductor fabrication factories and develop a product mix prediction model by applying the back-propagation neural network. The neural network model can respond rapidly to the relationship between product mix, equipment utilization and average WIP level. In addition, a database searching model is presented herein. Assighing the expected equipment utilization and average WIP level allows the developed database to obtain the feasible product mix. Moreover, a case study emonstrates the proposed models' effectiveness.en_US
dc.language.isozh_TWen_US
dc.subject產品組合zh_TW
dc.subject機器利用率zh_TW
dc.subject平均在製品存貨zh_TW
dc.subject類神經網路zh_TW
dc.subjectProduct Mixen_US
dc.subjectEquipment Utilizationen_US
dc.subjectWIPen_US
dc.subjectNeural Networken_US
dc.title晶圓製造廠產品組合、機器利用率與系統平均在製品存貨水準模式之構建-考慮隨機性模式zh_TW
dc.titleConstruction of Product Mix, Equipment Utilization and WIP Models for Semiconductor Fabrication Factoriesen_US
dc.typeThesisen_US
dc.contributor.department工業工程與管理學系zh_TW
顯示於類別:畢業論文