This preview shows pages 1–7. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: EE103 (Fall 200910) 5. The Cholesky factorization positive (semi)definite matrices examples the Cholesky factorization solving Ax = b with A positive definite inverse of a positive definite matrix permutation matrices sparse Cholesky factorization 51 Positive (semi)definite matrices A is positive definite if A is symmetric and x T Ax > for all x negationslash = 0 A is positive semidefinite if A is symmetric and x T Ax for all x Note: if A is symmetric of order n , then x T Ax = n summationdisplay i =1 n summationdisplay j =1 a ij x i x j = n summationdisplay i =1 a ii x 2 i + 2 summationdisplay i>j a ij x i x j The Cholesky factorization 52 Examples A 1 = bracketleftbigg 9 6 6 5 bracketrightbigg , A 2 = bracketleftbigg 9 6 6 4 bracketrightbigg , A 3 = bracketleftbigg 9 6 6 3 bracketrightbigg A 1 is positive definite: x T A 1 x = 9 x 2 1 + 12 x 1 x 2 + 5 x 2 2 = (3 x 1 + 2 x 2 ) 2 + x 2 2 A 2 is positive semidefinite but not positive definite: x T A 2 x = 9 x 2 1 + 12 x 1 x 2 + 4 x 2 2 = (3 x 1 + 2 x 2 ) 2 A 3 is not positive semidefinite: x T A 3 x = 9 x 2 1 + 12 x 1 x 2 + 3 x 2 2 = (3 x 1 + 2 x 2 ) 2 x 2 2 The Cholesky factorization 53 Examples A = B T B for some matrix B x T Ax = x T B T Bx = bardbl Bx bardbl 2 A is positive semidefinite A is positive definite if B has a zero nullspace diagonal A x T Ax = a 11 x 2 1 + a 22 x 2 2 + + a nn x 2 n A is positive semidefinite if its diagonal elements are nonnegative A is positive definite if its diagonal elements are positive The Cholesky factorization 54 Another example A = 1 1 1 2 . . . . . . . . . . . . . . . 2 1 1 1 A is positive semidefinite: x T Ax = ( x 1 x 2 ) 2 + ( x 2 x 3 ) 2 + + ( x n 1 x n ) 2 A is not positive definite: x T Ax = 0 for x = (1 , 1 ,..., 1) The Cholesky factorization 55 Resistor circuit y 1 y 2 x 1 x 2 R 1 R 2 R 3 Circuit model: y = Ax with A = bracketleftbigg R 1 + R 3 R 3 R 3 R 2 + R 3 bracketrightbigg ( R 1 ,R 2 ,R 3 > 0) Interpretation of x T Ax = y T x x T Ax is the power delivered by the sources, dissipated by the resistors The Cholesky factorization 56 A...
View Full
Document
 Spring '08
 VANDENBERGHE,LIEVEN

Click to edit the document details