• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 1
  • 1
  • Tagged with
  • 2
  • 2
  • 2
  • 2
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

縮基法初始值問題之數值研究 / Numerical studies of reduced basis methos for initial value problems

陳揚敏 Unknown Date (has links)
縮基法(RBM) 是對參數化的曲線求逼近解的一個方法,基本上乃使用投影法將解曲線投射到解空間的一子空間中,如此一來,可將原問題轉換成一較小的系統,並經由數值計算出小系統的解,來求得大系統的一逼近解。在本篇論文中主要的乃探討RBM在常微分方程組初始值問題上的應用,並發展一套含有誤差控制的演算法。 本篇論文中所採用的ODE Solver 乃由Gordon 和Shampine 基於Adams PECE方法所發展的。在求解的過程中,對於計算解誤差的控制我們除了利用ODE Solver 的誤差估計,另外我們又發展對縮基解(reduced basis solution) 的後(Aposteriori) 誤差估計,以確保數值計算解的準確性。我們所考慮使用的子空間有三種Taylor, Lagrange , Hermite 。同時為了要增加數值的特定性及簡化小系統的求解工作,我們先行將子空間的基底直交化。因此,除了誤差的控制外,我們也討論了roundoff error 對向量直交化及形成小系統時所造成的影響,並設立誤差標準以判別何時誤差過大到嚴重影響縮基解的準確度。 本篇論文的目的是希望利用RBM發展出一套解常微分方程組初始值問題的求算法,以期計算解能在較短的時間內準確的被計算出來。 / The reduced basis method(RBM) is a scheme for approximating parametric solution curves. The basic technique of RBM is projection. By applying the method, we can find an approximate solution of the original system which satisfies a system of smaller size. In this paper, we mainly concern the applications of RBM for ODE initial value problems and develop an algorithm which contains a set of error controls. The ODE solver used in this paper is developed by Gordon and Shampine based on Adams PECE formulas. To assure the accuracy of the reduced basis approximation, we set up an appropriate automatic error control in calling GS solver and develop an a posteriori error estimate to keep the reduction error under control. The subspaces considered are Taylor, Lagrange and Hermite subspaces.In the meantime, in order to improve the numerical stability and simplify the computation of the reduced basis solution, we orthogonalize the generators of reduced subspaces. We also discuss the roundoff errors in the orthogonalization process and build up a criterion for identifying the case the accuracy of the reduced basis solution up a criterion for identifying the case the accuracy of the reduced basis solution is destroyed by the errors. The aim of this paper is to develop an algorithm to solve the ODE initial value problems efficiently.
2

線性規劃求解方法之研究—個案分析 / Search Direction Analysis in Linear Programming -- Case Study

莊文華 Unknown Date (has links)
線性規劃的求解方法依幾何觀點可區分為單體法(Simplex Method)和起源於Karmarkar的內點法(Interior Point Method),無論在理論上或應用上這兩者的許多變體仍然不斷地在改進中。Dantzig的單體法和它變體的演算法是從邊上角點走到另一角點直至到達最佳點,內點法則是透過這個可行地區的內部的可行點走到可行點的方法。眾所週知,當遭遇最壞案例時單體法求解速度的函數表示成指數形式;而內點法卻成多項式形式。就一般案例而言,實驗證據提議,當問題大小為小型時,以單體法求解速度較佳。而內點法僅僅在很大規模的線性規劃時,其解法優於以單體法為基礎的方法。 儘管內點法求解速度較單體法為快,但大多數運用內點法求得最佳解的原理,都使用一個可行解區域內部的點作為起點(或稱為中心),然而就原始線性規劃問題而言,求得一可行解其複雜度和求得最佳解的複雜度是相同的。倘若能以單體法製造起始點的方法,搭配內點法的優點,產生一不同於過去單體法和內點法之新演算法,能降低求解速度,則是吸引人的課題。是故作者希望能以線性規劃問題的搜尋方向為主題加以整理,探討如何結合單體法及內點法的特性,深入研究搜尋方向在邊界上可能會發生的問題,並以一演算法為例,設計程式,作數據實驗,實際觀察問題發生的原因及提供解決的方法。

Page generated in 0.0234 seconds