本論文係利用人工免疫系統法來辨識管制圖趨勢分類模式。研究中採用Matlab程式撰寫人工免疫系統分類演算法,考慮四種輸入變數來辨識Normal、Mixtures、Shift in level、Stratification和Trends等五種不同類型之管制圖趨勢。研究結果顯示演算法演化出之抗體皆可有效辨識不同類型的管制圖趨勢,其中分類效果顯著的是Normal、Mixtures、Stratification和Trends四類,而Shift in level類型則分類成效有限。本論文之研究成果可以迅速提供製程相關資訊給予決策者,適時地採取因應的改善措施,進而提昇產品品質。
In this thesis, we apply the artificial immune system(AIS) to establish a pattern classification of control chart tendency model. Classification algorithm is a port of the pattern recognition. In the research, the immune algorithm is formed by the Matlab program. Four different input variables are used to differentiate the five different pattern tendencies, which are Normal, Mixtures, Shift in level, Stratification and Trends models. The results of the study show that the immune classification algorithm could differentiate the four control charts pattern out of five effectively. The types of Normal, Mixtures, Stratification and Trends are recognized significantly while Shift in level type is failure to be identified as well as the other four types. We wish the research results could provide worthy information to the manufacturing procedure, and the decision maker could promote the quality of products correctly in the early stage.