良率好壞直接影響生產成本，因此生產者在產品發展及量產之過程中，期望能找出迅速提升良率之方法。換言之，在發展及量產之階段，當生產良率不佳時就應該迅速將造成良率不佳之根本原因找出並加以改善調整，如此才能提升產品良率。而資料蒐集的單位從單一機台至整個製程廠區，資料相對上是越來越複雜且數量亦越來越大。在分析的難易度上當資料維度越大，其資料收集和分析是越困難的。因此，本研究團隊提出了關鍵參數搜尋法 (Key-variable Search Algorithm, KSA)，用以解決上述所提及之問題。
當良率損失發生時，傳統的作法是藉由蒐集造成良率損失的相關製程資料進行大數據分析，及利用長久以來累積的經驗建立起的失效記錄資料庫，讓後人得以參照相關的歷史資訊。但這些都建立在目前已知的情況上，對於製程中未知的情況仍須利用更多的資訊進行分析並與製程專家討論分析結果的合理性。然而，即使在已知的問題上進行大數據分析，仍可能無法有效率地找出根本原因；在另一方面，若是未知的問題則可能連問題發生所屬的範圍都未能確定，最終甚至連發生問題的根本原因也無從了解。而關鍵參數搜尋法可解決上述傳統的缺失，提供使用者一個快速且有效地找出影響良率根本原因的工具。此外對於使用者來說，一個工具的好壞取決於該工具使用上的便利性、即時性以及產生的結果品質。
因此，本論文參考「全自動虛擬量測系統 (Automatic Virtual Metrology, AVM) 在執行製程品質推估工作時，可藉由信心指標 (Reliance Index, RI) 提供使用者推估結果的信心度」之觀念，提出了一套適用於關鍵站點搜尋法的『信心指標』。此外，本論文亦提出盲站搜尋演算法，藉以輔佐 KSA機制解決因資料型態的限制造成機制中之核心演算法運算上產生之盲點。關鍵參數搜尋法為良率管理系統中之一種方法，本論文提出的兩種機制使此方法在執行時能更完善且效益更高。

Production yield directly affects production cost. Thus manufacturers seek to quickly enhance product yield during the development and mass-production phases. In other word, when production yield is poor, the root causes should be found rapidly and the problem should be resolved in both the development and mass-production phases so that production yield could be improved. Data collection ranges from a single equipment to all factory-wide, which means the data quantity is extremely large and the data analysis will become more difficult. To resolve the problem mentioned above, the scheme of Key-variable Search Algorithms (KSAs) is proposed.
When a yield loss is encountered, the traditional yield enhancement approach is to collect all production-related data that cause yield loss to perform big data analysis and to create historical database of yield loss by expert experience for posterity reference. But this approach may work only for an already-known situation. For unknown circumstances, we still need more information for analysis and need to discuss result reasonability with process experts. However, it is not efficient to search for the root causes of a known yield loss problem with big data analysis. On the other hand, if it is an unknown problem, then we may not confirm where does the problem come from and cannot even search for the root causes. The KSA scheme can resolve the shortcomings of the traditional approach and provide user a tool which is rapid and efficient to search for the root causes. For users, the convenience, immediacy, and result quality of using a tool are the factors to determine it is a good tool or not.
To better improve the performance of the KSA scheme, this paper refers to the concept of Reliance Index (RI) of the Automatic Virtual Metrology (AVM) system to develop the Reliance Index of the KSA scheme, denoted as RIK. In addition, Blind-stage-Search (BS) algorithm which resolves the blind point of the KSA scheme caused by the constraint of data types is also proposed in this paper. With the additions of RIK and BS algorithms, the KSA scheme will become more complete and more effective.
The illustrate examples taken from a thin film transistor array (TFT-Array) foundry are presented. Experimental results demonstrate that the proposed RIK and BS algorithms are applicable to the KSA Scheme.

[1] L. Fowler and T. Davis, “Engineering data analysis using Discovery,” IEEE/SEMl Advanced Semiconductor Manufacturing Conference, pp 416 - 422, November 1996.
[2] F.-T. Cheng, Y.-S. Hsieh, J.-W. Zheng, S.-M. Chen, R.-X. Xiao and C.-Y. Lin, “A Scheme of High-Dimensional Key-Variable Search Algorithms for Yield Improvement,” IEEE Robotics and Automation Letters, vol. 2, no. 1, pp. 179-186, June 2016.
[3] F.-T. Cheng, H.-C. Huang, and C.-A. Kao, “Developing an automatic virtual metrology system,” IEEE Transactions on Automation Science and Engineering, vol. 9, no. 1, pp. 181-188., January 2012.
[4] F.-T. Cheng, Y.-T. Chen, Y.-C. Su, and D.-L. Zeng, “Evaluating reliance level of a virtual metrology system,” IEEE Transactions on Semiconductor Manufacturing, vol. 21, no. 1, pp. 92-103, February 2008.
[5] 鄭春生，《品質管理-現代化觀念與實務應用》，全華圖書股份有限公司，2012。
[6] D. Port, Rick. Kazman, H. Nakao and M. Katahira, "Practicing What is Preached: 80/20 Rules for Strategic IV & V Assessment", Computer Software and Applications Conference, 2005. COMPSAC 2005. 29th Annual International, EQUITY, pp.45-54, 2007.
[7] J. M. Juran, & A. Blanton Godfrey "Juran’s Quality Handbook", McGraw-Hill, New York, 1999.
[8] L. Huang, L. Huang, J. Nie, “Using Pareto Principle to Improve Efficiency”, 2010 7th IEEE Consumer Communications and Networking Conference (CCNC), 2010.
[9] M. Singson, & P. Hangsing, "Implication of 80/20 Rule in Electronic Journal Usage of UGC-Infonet Consortia", The Journal of Academic Librarianship, 41(2), pp. 207-219, 2015.
[10] S. Krasteva, P. Sharma, & L. Wagman, "The 80/20 rule: Corporate support for innovation by employees", International Journal of Industrial Organization, 38, pp. 32-43, 2015.
[11] H. Cheng, M. P. L. Ooi, Y. C. Kuang, E. Sim, B. Cheah, S. Demidenko, “Automatic Yield Management System for Semiconductor Production Test”, 2011 Sixth IEEE International Symposium on Electronic Design, Test and Application (DELTA), pp. 254-258, 2011.
[12] G. Ducotey, A. Couvat, V. Audran, D. Pepper, L. Couturier, D. David, “In-line methodology for defectivity analysis from dark field wafer inspection to defect root cause analysis using FIB cut”, 2008 IEEE/SEMI Advanced Semiconductor Manufacturing Conference, pp. 138-141, 2008.
[13] F. Lee, S. Smith, “Yield analysis and data management using Yield ManagerTM”, 1998 IEEE/SEMI Advanced Semiconductor Manufacturing Conference and Workshop, pp. 19-30, 1998.
[14] H. Koyama, M. Inokuchi, “Yield Management for Development and Manufacture of Integrated Circuits”, 1998 IEEE/SEMI Advanced Semiconductor Manufacturing Conference, pp. 208-211, 1998.
[15] C.-K. Ing and T. L. Lai, "A stepwise regression method and consistent model selection for high-dimensional sparse linear models," Statistica Sinica, vol. 21, pp. 1473-1513, 2011.
[16] 黃文隆、黃龍，《統計學(上)(下)，修訂一版》，保成出版事業有限公司，2000。
[17] F.-T. Cheng, C.-A. Kao, C.-F. Chen, W.-H. Tsai, “ Tutorial on Applying the VM Technology for TFT-LCD Manufacturing,” IEEE Transactions on Semiconductor Manufacturing, vol. 28, no. 1, pp. 55-69, 2015.
[18] 田民波，林怡欣校訂，《TFT LCD面板設計與構裝技術》，五南圖書出版社，2010，第七章。
[19] 蕭宏，《半導體製程技術與導論第三版》，全華出版社，2014。