15.11.2014 Views

MATLAB Mathematics - SERC - Index of

MATLAB Mathematics - SERC - Index of

MATLAB Mathematics - SERC - Index of

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Cholesky, LU, and QR Factorizations<br />

LU Factorization<br />

LU factorization, or Gaussian elimination, expresses any square matrix A as<br />

the product <strong>of</strong> a permutation <strong>of</strong> a lower triangular matrix and an upper<br />

triangular matrix<br />

A<br />

=<br />

LU<br />

where L is a permutation <strong>of</strong> a lower triangular matrix with ones on its diagonal<br />

and U is an upper triangular matrix.<br />

The permutations are necessary for both theoretical and computational<br />

reasons. The matrix<br />

0 1<br />

1 0<br />

cannot be expressed as the product <strong>of</strong> triangular matrices without<br />

interchanging its two rows. Although the matrix<br />

ε 1<br />

1 0<br />

can be expressed as the product <strong>of</strong> triangular matrices, when ε is small the<br />

elements in the factors are large and magnify errors, so even though the<br />

permutations are not strictly necessary, they are desirable. Partial pivoting<br />

ensures that the elements <strong>of</strong> L are bounded by one in magnitude and that the<br />

elements <strong>of</strong> U are not much larger than those <strong>of</strong> A.<br />

For example<br />

[L,U] = lu(B)<br />

L =<br />

1.0000 0 0<br />

0.3750 0.5441 1.0000<br />

0.5000 1.0000 0<br />

U =<br />

8.0000 1.0000 6.0000<br />

0 8.5000 -1.0000<br />

0 0 5.2941<br />

1-29

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!