Transcript
Page 1: Introduction to Numerical Analysis I MATH/CMPSC 455 PA=LU

Introduction to Numerical Analysis I

MATH/CMPSC 455

PA=LU

Page 2: Introduction to Numerical Analysis I MATH/CMPSC 455 PA=LU

PARTIAL PIVOTING

Why? (1) zeros pivot; (2) swapping rows reduces possibility of

rounding errors and poor scaling contaminating the solution.

How?Swap rows at each elimination step to

place the largest element on the diagonal.

Page 3: Introduction to Numerical Analysis I MATH/CMPSC 455 PA=LU

PERMUTATION MATRICES

Definition: A permutation matrix is a matrix consisting of all zeros, except for a single 1 in every row and column.

Fundamental Theorem of Permutation Matrices:Let P be the permutation matrix formed by a particular set of row exchanges applied to the identity matrix. Then, for any matrix A, PA is the matrix obtained by applying exactly the same set of row exchanges to A.

Page 4: Introduction to Numerical Analysis I MATH/CMPSC 455 PA=LU

PA=LU FACTORIZATION

Definition: PA=LU factorization is simply the LU factorization of a row-exchanged version of A.

The algorithm consists of two parts:• Factorization phase: apply to A only and is

designed to produce the LU decomposition of PA.

• Solution phase: update the right hand side Pb, and then solve LU = Pb by forward and backward substitution.

Page 5: Introduction to Numerical Analysis I MATH/CMPSC 455 PA=LU

Example:

Example:

Example:


Top Related