| Abstract: | 在消費者導向的時代下，食品企業為迎合顧客的喜好以獲取市場佔有率，食品工業產品推陳出新、品目眾多，快速地開發新產品、新製程的設計並兼及提昇品質、降低成本成為各食品工業生產者致力的方向。田口方法的特性在於以低成本、高效率及數量化的指標做好品質設計的工作，也就成為一項實行成效良好的最佳品質工程技術。但在另一方面，食品工業產品的設計益形複雜，線外品質的設計往往需要考量多個品質特性的最佳化，田口方法僅能針對單一品質特性進行最佳化分析已不敷實際使用，憑藉工程人員所判斷的整體適用的因子水準組合又太過於主觀與存在不確定性。雖然近年來學術界已提出若干解決多重品質特性問題的方法，但國內的食品工業仍屬較為傳統的民生基礎工業，負責製程品質改善的工程人員未必皆具備有相關的數學或統計的分析背景，因此，最佳化的分析技術若涉及太多複雜數學或艱深的統計理論，將難以被食品業界所採用。 本研究融合灰色系統理論中的灰色關聯分析整合出多重品質特性績效衡量指標，在保有傳統田口方法的原始精神下，以一套簡便又不失學理的演算模式，提供予食品業界面對多重品質特性要求時尋找製程最佳參數組合的實用品質工程技術。就案例公司的食糧群產品中肉雞飼料製程，找出在為同時達成硬度（望大特性）、含粉率（望小特性）、成品水份含量（望目特性）三個品質特性的整體製程品質考量情況下，最佳的製程參數組合，並與根據單一品質特性考量所得之三組傳統田口方法最佳因子組合做比較，發現利用傳統田口方法所獲得的製程參數，只能滿足該品質特性本身的目標，對於其他品質特性則無法兼顧。透過食品工廠實際製程證實此多重品質特性演算法在食品生產實務運用的可行性及有效性。
During this consumer-orientated era and in response to customers’ preference, the food industry aims to gain market share by creating varieties of new products and designs of new processes, enhancing product quality, and lowering cost, all of which have guided grocers to commit their efforts to this industry. Characteristics of Taguchi Method lie in its low cost, high efficiency, and quantitative index to lead a food industry to a good quality design that consequently enhances its engineering technology. On the other hand, design for food industry has been further complicated by the design of its off-line quality, requiring several quality characteristics to be taken into account during optimization, but the Taguchi Method can only manage analysis of the optimization for a single quality characteristic. This may not satisfy practical application, and furthermore, the Taguchi Method relies on a subjective overall applicable factor level combination judged by engineers, which brings an additional uncertainty variable. Although academic circles have brought forth numerous solutions for problems concerning the multi-response, the domestic food industry still defines itself as a traditional basic industry, and those engineers who are in charge of improving quality of process may not necessarily possess related mathematical or statistical analysis backgrounds. Consequently, the analytical technology of optimization can hardly be accepted by the food industry if it deals largely with abstruse mathematical or statistical theories. By working with the Grey Relational Analysis in the Grey System Theory, this study analyzed and integrated a number of performance measuring indices for multi-response to present a simple calculation model that adheres to the spirit of the traditional Taguchi Method. This model would benefit the food industry who seeks the practical quality engineering technology of the best parameter combination when it faces the requirements of a multi-response. For process of the broiler feed among the food products available to broilers among the case companies, the overall process quality to achieve three quality characteristics including hardness (Smaller-the-Better), powder content rate (Larger-the Better) and water content of finished products (Nominal-the-Best) were discovered, in light of overall process quality consideration. The best process parameter combination was then compared to the three sets of the best factor combination of the traditional Taguchi Method obtained based on the single quality characteristic consideration, and it was found that only the process parameter obtained from the traditional Taguchi Method met the goal of the quality characteristic itself, but not other quality characteristics. Through the practical process of the food plants, it can substantiate the feasibility and validity of this multi-response calculation method on the application of food production. |
