成本導向之多品質混合實驗最佳化演算法

Abstract

工業界常利用實驗設計(Design of Experiments, DOE)或田口方法(Taguchi Method)來規劃實驗及找尋最佳參數組合，但此兩個方法只適用於最佳化單一品質特性。由於全球化使得產業競爭越來越趨激烈，著重單一品質特性之產品已經不足以達到消費者的要求。但若考慮多個品質特性同時最佳時，則可能使成本越來越高而售價往往因市場競爭無法提高過多而使公司難以生存，故在找尋最佳配方時，如何利用最小成本以提升多個品質特性最佳化是業界一個重要的問題。此外，某些產業如材料、食品、化工等，其產品在設定因子水準時著重在各因子間之比例關係，而非因子之特定水準值，這類產品並不適用傳統的實驗設計來規劃實驗及求得最佳因子水準組合，而是屬於混合實驗(mixture experiment)之範疇。本研究利用資料包絡法(Data Envelopment Analysis)結合模糊邏輯法(Fuzzy logic)及自組性演算法(Group Method of Data Handling, GMDH)，針對具多品質特性之混合實驗，發展一套低成本高品質之最佳配方演算方法。本研究最後應用新竹某廠商所提供之煞車皮碗實例說明本研究所提出方法之確實有效。

Design of Experiments(D.O.E.) and Taguchi Method are often utilized to find the optimal factor-level combination in industries. These two methods can optimize single quality characteristic. With the increasing demands of the consumers, product design is becoming more and more complicated and cost-oriented. Optimization of a single quality characteristic can no longer satisfy the needs of customers. However, the conventional designs of experiments methods are not appropriate for applying to some industries, such as chemical or material industries. Their responses of the experiments are affected by the proportional relationship among the factors rather than the quantities of the factors, called mixture experiments. Therefore, this study first utilizes Data Envelopment Analysis (DEA) and Fuzzy Logic to integrate the multiple responses into a composite index, and then employs Group Method of Data Handling (GMDH) to develop a procedure to pursue ideal proportion of components with minimum cost under the restriction of good composite index and proportion limits of components. Therefore, by using the proposed procedure, superior proportion of components to manufacture low cost and high quality products can be obtained. A real case of rubber bowl production from a Taiwan automobile company is utilized to demonstrate the effectiveness of the proposed procedure.