半導體產業其資本支出都較為龐大，且電子產業其產品生命周期較短的緣故，半導體製造商在面對龐大的投資金額以及身處高度競爭的環境之中，不僅要滿足顧客的要求，還必須能夠在短期之內進入市場才能在產品市場中具有競爭力，為此需要透過正確的機台組合達到快速供貨的能力，即提升資本的使用效率來增強企業的競爭力，因此萌生了研究如何讓晶圓廠達到經濟效益最優化的產出模式。 本研究之研究模型為將半導體晶圓廠預設十個加工工作區，假設當半導體市場需求資訊給定的情況下，根據起始解得出的機台數量，應用電腦模擬進行廠區模擬產生出相關的生產績效資料，再利用迴歸分析找出晶圓廠區中需要進行優化改善的工作區域，並運用模擬退火演算法有效率地做機台數量調整，以利在最短的時間內使生產及資本的效益最佳化。在本研究中之起始解及驗證解中可以發現，生產績效部分，產出率增加了0.18%、在製品數量減少0.13%、製造時間也減少了0.30%，證實本研究方法找出之關鍵影響工作區對於廠區之整體作用是有顯著影響力的。 ;The semiconductor industry requires significant capital expenditures, and due to the short product lifecycle in the electronics industry, semiconductor manufacturers face immense investment costs and a highly competitive environment. They not only need to meet customer demands but also must enter the market quickly to remain competitive in the product market. To achieve this, they need the capability to rapidly supply products through the right combination of equipment, thus improving the efficiency of capital utilization and enhancing the competitiveness of the company. As a result, there is a need to study how to optimize the production mode of wafer fabs to achieve the best economic benefits. The research model in this study assumes ten processing zones in a semiconductor wafer fab. Under the given semiconductor market demand information, the initial solution determines the number of equipment required. Computer simulations are then applied to generate relevant production performance data for the fab. Regression analysis is used to identify the areas in the wafer fab that require optimization and improvement, and an efficient simulated annealing algorithm is employed to adjust the number of equipment, aiming to optimize production and capital efficiency in the shortest possible time. Through the initial and validation solutions in this study, it was found that in terms of production performance, the output rate increased by 0.18%, the work-in-process quantity decreased by 0.13%, and the manufacturing time decreased by 0.30%. This confirms that the key affected areas identified by the research methodology have a significant impact on the overall operation of the wafer fab.
