透過您的圖書館登入
IP:54.236.245.71
  • 學位論文

應用基因演算法與差分演算法與灰關聯在製程最佳化-以微流體晶片製程為例

Study on Optimal Process of Microfluidic Chip by Genetic Algorithm and Differential Evolution Algorithm and Grey Theory

指導教授 : 邱垂昱
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


隨著科技發展日新月異一日千里,人們研究範疇也從肉眼所及之事物演進到顯微鏡下的世界乃至基因定序與蛋白質分析等等,微電機技術的應用是一大關鍵,隨著研究發展,檢測儀器的敏感度雖以達到水準,但檢測時間過長、操作不易、儀器大小攜帶不便、儀器昂貴等,以至於無法普及,分析儀器微小化(Miniaturization)的概念進而產生,將取樣、樣品前處理、樣品分離、偵察等功能整合在晶片之上透過外加電壓或離心力等,驅使樣品在各元件中的微管道中移動,完成檢測。 微流體晶片主要優勢在於微小化後而提升性能、品質、可靠度,可以快速得到檢測結果,減少樣品的消耗,並同時可以降低製造成本。目前微流體晶片應用的方面有新藥開發、疾病檢測、基因定序、蛋白質分析、感染病原檢測、血液篩檢、法醫辨識鑑定、環境檢測及食品檢驗等。若無法保證微流體晶片的產出可以處在穩定的高品質高良率的狀況下,則造成更多的成本去彌補。所以本研究建立一套模式,針對微流體晶片製程參數的最佳化的應用,使用田口方法減少實驗次數與降低成本。透過灰關聯法、基因演算法以及差分演算法針對製程最佳化,得到最佳微流體晶片製程參數。灰關聯分析實驗結果得到最佳化參數組合與各因子影響性,得到最重要製程參數為壓印溫度.而演算法實驗結果所得到的最佳化參數,也明顯較大幅改善原本製程結果。

並列摘要


In recent years, the emerging technologies such as DNA sequencing, protein analysis and micro-electronics draw a lot of research attention. The application of the micro-electronics is among the most important ones. The sensitivity of machine has reached a certain level. Due to lengthy inspecting time, difficulty of carrying, expensive cost, some inspecting devices are not universal. The concept of miniaturization becomes very critical to the development of micro-electronic technology which integrates the sampling, pretreatment, detachment, testing process into the micro-fluidic chip that gives an impetus to the sample to move between the micro channels of the parts to detect by centrifugal force and impressed voltage. The main advantage of micro-fluidic chip is to increase the performance, quality, reliabilityand reduce the sample consumption and cost. The applications of micro-fluidic chip include new medicine development, disease examination, DNA sequencing, protein analysis, infection cause of disease, blood sieves examines, environment examination, food examination and so on. A lot of cost may happen if the micro-fluidic chip manufacturing process does not result in high quality and yield rate.In this project, we proposed a model which tries to optimize the parameters of the manufacturing process of micro-fluidic chips.The Taguchi method is used to reduce the number of experiments. Genetic algorithm, differential evolution algorithm and grey theory are utilized to optimize the parameters of manufacturing process. The result of gray relational analysis obtains the optimize the parameters of the manufacturing process of micro-fluidic chips and the effect of each parameters.the most important parameters of the manufacturing process of micro-fluidic chips is the embossing temperature. The result of the algorithms are apparently improved the result of the original process.

參考文獻


12. 葉偉弘,應用基因演算法及灰色理論於微針陣列製程最佳化之研究,碩士論文,國立台北科技大學工業工程與管理研究所,台北,2012。
13. 李國華,結合灰關聯分析法與田口法於壓鑄製程參數最佳化,碩士論文,國立台北科技大學材料科學與工程研究所,台北,2009。
17. 陳怡珊,應用田口灰關聯分析於生物冷凍櫃多目標最佳化設計,碩士論文,國立高雄大學亞太工商管理學系碩士班,高雄,2009。
3. 林信成,「軟性計算理論與知識處理技術之研究」,教育資料與圖書館學,第三十八卷,第二期,2000,第148-173頁。
18. Ali Wagdy Mohamed, Hegazy Zaher Sabry, Tareq Abd-Elaziz, "Real parameter optimization by an effective differential evolution algorithm" Egyptian Informatics Journal, Vol 14, 2013, pp. 37–53

延伸閱讀