IC封裝廠中期生產規劃系統之構建

Abstract

隨著電子產業的發展，半導體相關需求也日益增加，連帶使下游封裝廠需求也水漲船高。目前國內IC封裝廠多屬代工生產型態，在生產管理上常面臨生產預測不易、訂單動態到達等問題。因此，如何配置具有良好彈性的產線規模，已成為IC封裝廠中期規劃不可或缺的重點。 本研究所提之IC封裝中期生產規劃系統，共包括：產線配置、接單兩個模式。在本規劃系統中，打破傳統的規劃時距，以短週期之生產規劃理念解決生產預測困難的問題。在中期產線配置模式上，考量規劃瓶頸具有相當長的設置時間之特性，動用群組技術及限制理論等概念，使各种產品於所屬的產線生產，減少不必要之機器設置時間。接著，依初步產線配置之結果，設定各產線之規劃瓶頸及利用率瓶頸之預定利用率，用以求算各產線之規模及每日可投入量，作為接單時每日可承接訂單量之依據。 由於IC封裝廠訂單動態到臨，因此，本文發展短期產線微調模式，依訂單到臨的不準確性，調整已配置之各產線利用率瓶頸規模，來增加系統產出及產線之彈性，用以觀察各產線可承受之變異程度，作為下一規劃週期規劃瓶頭利用率設定之參考，確保中其規劃之可行性。 範例結果顯示，在整體需求不變的情況下，各種到料不準確比例經由本文產線微調模式之調整，皆能提供30%以上的產線彈性。並且，各產線每日最適投入量的規劃，能確實有效地降低產品的生產週期時間。

With the rapid development fo electronic products, the demand fro IC chips is increasing. IC packaging thus is getting more and more important. Due to the OEM production type, today, most IC pacaging factories face the problems that orders are arrived dynamically and mot easy to be forecasted. Therefore. the key point of mid-term production planning of the IC packaging factories is to plan the flexible production lines such that the capacities among lines can be adjusted. In the thesis, the mid-term production planning is executed through two models, production lines allocating and orders accepting. To allocate system capacities to prouction lines, the concept of group technology and theory of constraint is used so as to reduce setup times. Then, following the result of production lines allocating, the bottleneck machine utilization rate is set to plan the daily releasing quantities in each line. This justifies that orders accepting is fit for the planned capacities in each line. To find out the capacity for handling the variability of order arrivals, a short-term production lines capacity adjusting model is designed. When the actual order arrival is different from the planned one, the numbers of bottleneck machines are adjusted between production lines considering the capacity needed for actual order arrivals. Running the model for different cases, we can observe the degree of variability that the system can handle so as to make sure the feasibility of the mid-term proluction planning. This degree of variablity is then used as reference for setting the bottleneck utilization rate in next planning horizon. To verify the performance of proposed mid-term production planning system, an example is tested. The simulation results indicate that this model can provide at least 30% flexibility to handle the dynamically arrived orders given that the total capacity requirement is fixed. Also, the lead time can be decreased effectively by the releasing strategy.