1 / 60

Numerical Analysis

Numerical Analysis. EE, NCKU Tien-Hao Chang (Darby Chang). In the previous slide. Error estimation in system of equations vector/matrix norms LU decomposition split a matrix into the product of a lower and a upper triangular matrices

hop
Télécharger la présentation

Numerical Analysis

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Numerical Analysis EE, NCKU Tien-Hao Chang (Darby Chang)

  2. In the previous slide • Error estimation in system of equations • vector/matrix norms • LU decomposition • split a matrix into the product of a lower and a upper triangular matrices • efficient in dealing with a lots of right-hand-side vectors • Direct factorization • as an systems of n2+n equations • Crout decomposition • Dollittle decomposition

  3. In this slide • Special matrices • Strictly diagonally dominant matrix • Symmetric positive definite matrix • Cholesky decomposition • Tridiagonal matrix • Iterative techniques • Jacobi, Gauss-Seidel and SOR methods • conjugate gradient method • Nonlinear systems of equations • Exercise 3

  4. 3.7 Special Matrices

  5. Special matrices • Linear systems • which arise in practice and/or in numerical methods • the coefficient matrices often have special properties or structure • Strictly diagonally dominant matrix • Symmetric positive definite matrix • Tridiagonal matrix

  6. Strictly diagonally dominant

  7. Symmetric positive definite

  8. Symmetric positive definiteTheorems for verification

  9. Symmetric positive definiteRelations to • Eigenvalues • Leading principal sub-matrix

  10. Cholesky decomposition • For symmetric positive definite matrices • greater efficiency can be obtained • consider the symmetric of the matrix • Rather than LU form, we factor the matrix into the form • A=LLT

  11. Tridiagonal • Only 8n-7 operations • factor step 3n-3 • solve step 5n-4

  12. 3.7 Special Matrices

  13. Before entering 3.8 • So far, we have learnt three methods algorithms in Chapter 3 • Gaussian elimination • LU decomposition • direct factorization • Are they algorithms? • What’s the differences to those algorithms in Chapter 2? • they report exact solutions rather than approximate solutions question further question answer

  14. 3.8 Iterative Techniques for Linear Systems

  15. Iterative techniques • Analytic techniques is slow • O(n3) • Especially for systems with large but sparse coefficient matrices • As an added bonus, iterative techniques are less insensitive to roundoff error

  16. Iterative techniquesBasic idea

  17. Iteration matrixImmediate questions • When does T guarantee a unique solution? • When does T guarantee convergence? • How quick does {x(k)} converge? • How to generate T?

  18. Assume that I-T is singular, there exists a nonzero vector x such that (T-1I)x=0 • 1 is a eigenvalue of T • but ρ(T)<1, contradiction

  19. (in section 2.3 with proof) Recall that http://www.dianadepasquale.com/ThinkingMonkey.jpg

  20. Recall that http://www.dianadepasquale.com/ThinkingMonkey.jpg

  21. Iteration matrixFor these questions • We know that when ρ(T)<1, {x(k)} from x(k+1)=Tx(k)+c will converge linearly to a unique solution with any initial vector x(0) • What is missing? • remember the problem is to solve Ax=b • How to generate T? • like f(x) and g(x), different algorithms construct different iteration matrix question hint answer

  22. Splitting Methods

  23. Splitting methods • Ax=b (M–N)x=b  Mx=Nx+b • x=M-1Nx+M-1b • T=M-1N and c=M-1b • A class of iteration methods • Jacobi method • Gauss-Seidel method • SOR method

  24. Gauss-Seidel method

  25. Gauss-Seidel methodIteration matrix

  26. The SOR method (successive overrelaxatoin)

  27. Iterative Techniques for Linear Systems

  28. 3.9 Conjugate Gradient Method 37

  29. Conjugate gradient method • Not all iterative methods are based on the splitting concept • The minimization of an associated quadratic functional

  30. Conjugate gradient methodQuadratic functional

  31. http://fuzzy.cs.uni-magdeburg.de/~borgelt/doc/somd/parabola.gifhttp://fuzzy.cs.uni-magdeburg.de/~borgelt/doc/somd/parabola.gif

  32. Minimizing quadratic functional

  33. Choose the search direction d(m) • as the tangent line in Newton’s method • the gradient of f at x(m) • Choose the step size • as the root of the tangent line

  34. http://www.mathworks.com/cmsimages/op_main_wl_3250.jpg Global optimization problem

  35. Conjugate Gradient Method

  36. 3.10 Nonlinear Systems of Equations

  37. Nonlinear systems of equations

  38. Generalization of root-finding

  39. GeneralizationNewton’s method

  40. Generalization of Newton’s methodJacobian matrix

More Related