晶圓廠的生產設備在五到七年會因製程演進而折舊淘汰，此外半導體廠商的產能大小會影響其供貨速度，如果產能不足，將嚴重影響相關電子產品的生產與供應速度，電子產品的生命週期短，產品價格隨時間快速下滑，在龐大的投資成本與景氣波動明顯下，正確的機台組合與快速的供貨能力，可以提升資本的使用效率，增強企業的競爭力。 本研究提出一套完整的半導體最佳產出決策，在考慮產品需求與機台組合的變動，來讓在製時間(cycle time)減到最小，讓產出保持在一個高水準，利用eM-plant模擬軟體構建實際半導體廠之生產系統，進行模擬實驗，產生出生產績效指標(在製時間)的模擬資料，再藉由實驗設計（DOE）的方法，將影響生產績效指標的各個機台區列為實驗控制因子，設計成每因子各兩個水準的一個單一重複的2k-5設計來分析因子影響效應，並透過統計方法尋找最適因子組合，來達到生產績效指標(在製時間)最小化之目的，最後以經濟學領域中的規模經濟與不經濟理論，找出最適規模的產能。 The production equipment of semiconductor industry need to depreciation and elimination about 5 to 7 years because of the evolution of technology. Moreover the capacity of manufacturer will effect its delivery time. If throughput are shortage it will critically affect the production and supply speed of related electronic products. The life cycle of electronic products is short, wafer prices rapid reduce by time. With the astronomical capacity cots and high demand volatility, exact tool combination and rapid delivery time ability can make best use of capital and strength the competition of enterprise. The thesis addresses a total solution to make best decision for semiconductor manufactory. Considering customer demand and variation tool sets to minimize the cycle time and keep throughout at a high level. By using eM-plant simulation tool to construct a real semiconductor manufacturing system, and obtain the data of cycle time. The tool sets that significantly influence production cycle time were identified through factor screening experiments. A 2k-5 un-replicate factional factorial design was performed. Base on the results from the factor screening experiments, using statistical analysis to find the best factor combination to reach minimize the cycle time. Finally, the optimal scale through the economies of scale and diseconomies of scale in economic field can find.
