1 / 31

Giansalvo EXIN Cirrincione

unité #5. Analyse numérique matricielle. Giansalvo EXIN Cirrincione. U.  1. . 0. V. . . . =. 0. .  n. . . 0. . 0. Full SVD. a 11. . a 1n. . . . . valeurs singulières. . . a m1. . a mn. Décomposition en valeurs singulières ( SVD ). U.  1.

Download Presentation

Giansalvo EXIN Cirrincione

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. unité #5 Analyse numérique matricielle Giansalvo EXIN Cirrincione

  2. U 1  0 V    = 0  n   0  0 Full SVD a11  a1n     valeurs singulières   am1  amn Décomposition en valeurs singulières (SVD)

  3. U 1  0 V    = 0  n   0  0 a11 a11   a1n a1n U 1  0 V        ^ =     0  n     am1 am1   amn amn Reduced SVD Décomposition en valeurs singulières (SVD)

  4. Full SVD Reduced SVD Décomposition en valeurs singulières (SVD)

  5. Approximation au sens des moindres carrées Example: polynomial data fitting

  6. f(x) 2 1 0 yi -1 -2 -3 -4 xi 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 Approximation au sens des moindres carrées

  7. discrete square wave interpolation least squares m = n = 11 m = 11 , n = 8 Approximation au sens des moindres carrées

  8. Approximation au sens des moindres carrées Posons le problème matriciellement

  9. Matrice de Vandermonde (1735-1796) Approximation au sens des moindres carrées système linéaire de n équations et n inconnues erreur d’approximation

  10. Approximation au sens des moindres carrées forme quadratique Équations normales

  11. r = b - A x b y = A x range(A) = Pb The system is nonsingular iff A has full rank.

  12. The system is nonsingular iff A has full rank.

  13. Solution par les équations normales factorisation de Cholesky AHA est une matrice n x n hermitienne strictement définie positive 1. Form the matrix AHA and the vector AH b 2. Compute the Cholesky factorization AHA = RHR 3. Solve the lower-triangular system RH w = AH b for w 4. Solve the upper-triangular system R x = w for x

  14. reduced QR factorization 1. Compute the reduced QR factorization 2. Compute the vector 3. Solve the upper-triangular system for x Solution par la factorisation QR (Householder)

  15. 1. Compute the reduced SVD 2. Compute the vector 4. Set 3. Solve the diagonal system for w Solution par la SVD

  16. Comparison of algorithms • speed : normal equations • standard : QR factorization • A close to singular : SVD • Drawbacks • normal equations : not always stable in the presence of rounding errors • QR factoriz.: less-than-ideal stability properties if A is close to singular • SVD : expensive for mn

  17. Conditionnement et précision

  18. r = b - A x b  closeness of the fit = Pb y = A x range(A) Conditionnement du problème des moindres carrées Données : A , b Solutions : x , y

  19. Conditionnement du problème des moindres carrées Données : A , b Solutions : x , y 2-norm relative condition numbers exact for certain  b upper bounds

  20.  highly ill-conditioned basis  very close fit Stabilité des méthodes des moindres carrées exemple Least squares fitting of the function exp(sin(4)) on the interval [0,1] by a polynomial of degree 14  x15 = 1

  21. reduced Stabilité des méthodes des moindres carrées exemple factorisation QR (Householder) The rounding errors have been amplified by a factor of order 10 9. This inaccuracy is explained by ill-conditioning, not instability.

  22. Stabilité des méthodes des moindres carrées exemple factorisation QR (Householder) implicit calculation of the product QH b

  23. Stabilité des méthodes des moindres carrées exemple factorisation QR (Householder) implicit calculation of the product QH b

  24. Stabilité des méthodes des moindres carrées exemple factorisation QR (Householder) backward stable

  25. Stabilité des méthodes des moindres carrées exemple SVD It beats Householder triangularization with column pivoting ( MATLAB's \ ) by a factor of about 3 backward stable

  26. Stabilité des méthodes des moindres carrées exemple équations normales factorisation de Cholesky not even a single digit of accuracy unstable

  27. Stabilité des méthodes des moindres carrées BS least squares algorithm The condition number of the LS problem may lie anywhere in the range  to 2 .

  28. Stabilité des méthodes des moindres carrées BS least squares algorithm Cholesky factorization (BS) The normal equations are typically unstable for ill-conditioned problems involving close fits.

  29. Stabilité des méthodes des moindres carrées The solution of the full-rank least squares problem via the normal equations is unstable. Stability can be achieved, however, by restriction to a class of problems in which (A) is uniformly bounded above or (tan)/is uniformly bounded below. The normal equations are typically unstable for ill-conditioned problems involving close fits.

  30. FINE

More Related