透過您的圖書館登入
IP:3.129.22.135
  • 期刊
  • OpenAccess

結合基因演算法與田口實驗法於飛行控制設計

Application of Genetic Algorithms and Taguchi Methods to Flight Control Design

摘要


在衆多的數值最佳化方法中,基因演算法則爲眾人所熟知且已廣泛地應用於不同領域之最佳化研究中。然研究發現,若搜尋的空間愈廣或所選擇之適存函數非線性程度愈高,則基因演算法所得之解愈有可能會隨著其蟲口數、交換率、突變率與母代保留數設定的不同而異。在這篇論文中,吾人引用田口品質工程實驗設計法的理念,針對不同的飛行控制問題找出基因演算法中各參數之最適組合,使所搜尋之解更具強健、更趨近於最佳值之效果。本文先以範例說明理念之可行性後,再將基因演算法與田口實驗法的結合體與不同之控制理論(LQR與H(下標 ∞))配合,應用於F16飛行控制之設計中。除達成控制目的之最佳設計外,同時成功驗證基因演算法與田口實驗法兩者結合應用之優越性。

並列摘要


Genetic algorithms (GAs) are well known in most kinds of numerical methods and are widely applied m different areas of optimal studies. It is found that if the search space is wide or the fitness function that we select has high degree of nonlinear, the GA solutions could strongly depend on the setting parameters which include population size crossover rate mutation rate and the remaining size of parent in GAs. This paper applies Taguchi methods to GAs to find the best combination of the setting parameters for different flight control problems. The purpose of such a combination is to make the control more robust and close to optimal solution. To demonstrate this new idea, we consider its applications to different flight control problems for F16 fighter by using auto-stabilization LQR and H(subscript ∞) design. The simulation results not only achieve the expected optimal design for flight control problems but also justify the reliability and feasibility of the proposed method.

被引用紀錄


黃怡君(2006)。模擬退火實數編碼利基遺傳演算法應用於河川水理模式阻力參數自動率定之研究〔碩士論文,國立臺灣大學〕。華藝線上圖書館。https://doi.org/10.6342/NTU.2006.02329

延伸閱讀