六標準差方法論之發展與應用：以半導體廠為例

Development and Application of Six Sigma Methodology in Semiconductor Manufacturing Companies

周家任(Chia-Jen Chou)

指導教授 ： 蘇朝墩

國立清華大學/工學院/工業工程與工程管理學系/博士(2008年)

摘要

六標準差已經廣泛的應用在全球不同的企業，並儼然成為在品質的範疇中最重要的議題。六標準差是一具有完整架構的方法論，其透過持續的改善協助企業達到長期目標。然而，在推行六標準差活動時仍會遭遇到許多的困難，例如，如何有效產生可行的六標準差專案、在有限的資源情況下如何選取最重要的專案。另一方面，半導體產業在國內政府倡導下蓬勃發展，這類產業具有分工細密、組織架構複雜，在導入六標準差時會有缺乏效率且無法確實達到營運目標的困擾。因此一套有系統、有架構的六標準差導入方案對於半導體業的幫助必然有其助益。
本研究發展出一套六標準差方法論，包括macro模式與micro 模式。Macro model主要探討的是六標準差策略方面，期許透過此模式將方案進行分類，以找出對企業本身具有高度利益與可行的專案。Macro model主要是透過風險及利益兩方面建構整個評估模式。另一方面，micro model是針對六標準差專案的執行層面，利用六標準差有效的執行步驟來達到預期的專案目標。本研究也導入用以解決複雜問題的計算智慧技術，更進一步的協助企業解決專案的問題。
本研究以半導體廠為例透過所提出的macro model，為個案公司快速的篩選有用的專案，並成功的將專案分成黑帶、綠帶及無效益的專案種類。此外，運用micro model的問題解決過程以及計算智慧(computational intelligence)的技術，本研究有效地降低個案公司的不良問題點，並大大的提昇了良率。

關鍵字

六標準差 ； 半導體製造 ； 計算智慧

並列摘要

Six Sigma has been widely adopted in a variety of industries in the world and it has become one of the most important subjects of debate in quality management. Six Sigma is a well-structured methodology that can help a company achieve expected goal through continuous project improvement. Some challenges, however, have emerged with the execution of the Six Sigma. For examples, how are feasible projects generated? How are critical Six Sigma projects selected given the finite resources of the organization? On the other hand, semiconductor manufacturing industries in Taiwan have had significant growth over the last decade in the world and also make up for the bulk of economic benefits. Semiconductor foundries are multi-layer organizations with complex systems and complicated division of labor. Due to the organization structure is huge; when a semiconductor foundry executes the Six Sigma methodology is ineffective and difficult to achieve the organization objectives. A completely and system procedure for implementing the Six Sigma methodology into semiconductor manufacturing industries is necessary and beneficial.
This study aims to develop the knowledge based Six Sigma methodology which is involved macro model and micro model. The macro model is focus on the strategy of Six Sigma which is including project generation, project evaluation, project selection and so on. The objective of macro model is to effectively identify the vital project. In addition, the micro model is aimed at the practical of Six Sigma. The micro model is to implement the DMAIC cycle to achieve the objective of each project. Furthermore, we conduct the computational intelligence approach to resolve the complex project problem.
This proposed procedure of macro model was successfully help the case company divided project into Black belt project, green belt project and profitless project. In addition, the micro model which is conducting the computational intelligence approach is effectively decrease the waste of manufacturing process and reaches the optimization situation.

並列關鍵字

Six Sigma ； semiconductor manufacturing ； computational intelligence

參考文獻

[1] Abouhamze, M., and Shakeri, M., “Multi-objective stacking sequence optimization of laminated cylindrical panels using a genetic algorithm and neural networks,” Composite Structures, vol. 81, pp. 253-263, 2007.

[4] Arita, T. and McCann, P., “The spatial and hierarchical organization of Japanese and US multinational semiconductor firms,” Journal of International Management, vol. 8, pp. 121-139, 2002.

[5] Ashu, J., Sanaga, S., Kumar, B. R., “Determination of an optimal unit pulse response function using real-coded genetic algorithm,” Journal of Hydrology, vol. 303, Issue 1-4, pp. 199-214, 2005.

[7] Berry, M. J. A., and Linoff, G. S., Data Mining Techniques for Marketing, Sales, and Customer Relationship Management, Wiley Publishing, Inc., 2004.

[8] Box, G., and Wilson, K. B., “The exploitation of response surfaces: some general considerations and examples,” Biometrics, vol. 10, pp. 16-60, 1954.