摘要 「田口方法」(Taguchi Methods)與「反應曲面法」(Response Surface Methodology;RSM)為目前常用之實驗設計方法,前者主要是利用直交表( Orthogonal Array)來規劃實驗進行產品品質之改善,並以信號雜音比(Signal-to-Noise Ratio)作為衡量指標,以找出最佳之製程參數設計降低產品品質變異;後者則是一結合數學應用、實驗設計與迴歸分析等工具的方法論,有助於問題建構與分析並加以求解模擬。 本研究鑒於研究與實務發展之需要,提出一「PEC應用模式」(Preparation-Experiment-Completion Application Model),提供以田口方法或反應曲面法之改善專案,在進行實驗規劃與執行時一可參考之依據,另外此模式包含「田口方法與反應曲面法之整合應用模式」(Taguchi Methods and Response Surface Methodology Integration Application Model )和「田口-反應曲面法」(Taguchi-Response Surface Methods)兩個方法,用以搭配「PEC模式」之進行。整個「PEC應用模式」以平版印刷廠(Offset Printing)之品質改善案作為驗證。
Abstract “Taguchi Methods (TM)” and “Response Surface Methodology (RSM)” are the popular designs of experiments. Taguchi Methods that search the optimal parameters of the process or system use “Orthogonal Array” to plan experiments for quality improvement and “S/N Ratio (Signal-to Noise Ratio)” as an evaluating index. RSM is a helpful approach to establish and analyze problems in mathematical models by merging mathematical methods, design of experiments, regression analysis, etc. This research proposes an integration model based on research gap and practical need for improvement projects by TM or RSM as a reference when planning experiments and executing ones. The model is named “Preparation-Experiment-Completion (PEC) Application Model”. It is involves in “Taguchi Methods and Response Surface Methodology Integration Application Model” and “Taguchi-Response Surface Methods” for operating collaboratively with “PEC Model”. The “PEC Model” is verified by a quality improvement project in the offset printing factory.