24.03.2013 Views

Chapter 3 Solution of Linear Systems - Math/CS

Chapter 3 Solution of Linear Systems - Math/CS

Chapter 3 Solution of Linear Systems - Math/CS

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

62 CHAPTER 3. SOLUTION OF LINEAR SYSTEMS<br />

Here the superscripts on the entries a (k)<br />

ij<br />

are used to indicate entries <strong>of</strong> the matrix that are modified<br />

during the kth elimination step. Recall that, in general, the multipliers are computed as<br />

mij =<br />

element to be eliminated<br />

current pivot element<br />

If the process does not break down (that is, all pivot elements, a11, a (1)<br />

22<br />

the matrix A can be factored as A = LU where<br />

⎡<br />

⎢<br />

L = ⎢<br />

⎣<br />

1<br />

m21<br />

m31<br />

.<br />

0<br />

1<br />

m32<br />

.<br />

0<br />

0<br />

1<br />

· · ·<br />

· · ·<br />

. ..<br />

⎤<br />

0<br />

0 ⎥<br />

0 ⎥<br />

. ⎦<br />

and<br />

⎡<br />

⎢<br />

U = ⎢<br />

⎣<br />

mn1 mn2 mn3 · · · 1<br />

The following example illustrates the process.<br />

Example 3.3.1. Consider the matrix<br />

⎡<br />

A = ⎣<br />

1 2 3<br />

2 −3 2<br />

3 1 −1<br />

Using Gaussian elimination without row interchanges, we obtain<br />

⎡<br />

1<br />

⎣ 2<br />

2<br />

−3<br />

⎤<br />

3<br />

2 ⎦ −→<br />

⎡<br />

1<br />

⎣ 0<br />

2<br />

−7<br />

⎤<br />

3<br />

−4 ⎦ −→<br />

3 1 −1<br />

0 −5 −10<br />

and thus<br />

⎡<br />

L = ⎣<br />

1 0 0<br />

m21 1 0<br />

m31 m32 1<br />

m21 = 2<br />

m31 = 3<br />

3 5<br />

7<br />

⎤<br />

⎦ .<br />

.<br />

, a(2) 33 , . . ., are nonzero) then<br />

upper triangular matrix<br />

after elimination is<br />

completed<br />

⎡<br />

⎣<br />

m32 = 5<br />

7<br />

⎤ ⎡<br />

1 0 0<br />

⎤<br />

⎡<br />

⎦ = ⎣ 2 1 0<br />

1<br />

⎦ and U = ⎣<br />

It is straightforward to verify2 that A = LU; that is,<br />

⎡<br />

1<br />

⎣ 2<br />

2<br />

−3<br />

⎤ ⎡<br />

3 1<br />

2 ⎦ = ⎣ 2<br />

0<br />

1<br />

⎤ ⎡<br />

0<br />

0 ⎦ ⎣<br />

3 1 −1<br />

1<br />

3 5<br />

7<br />

If we can compute the factorization A = LU, then<br />

1 2 3<br />

0 −7 −4<br />

0 0 − 50<br />

7<br />

Ax = b ⇒ LUx = b ⇒ Ly = b, where Ux = y.<br />

Therefore, the following steps can be used to solve Ax = b:<br />

• Compute the factorization A = LU.<br />

• Solve Ly = b using forward substitution.<br />

• Solve Ux = y using backward substitution.<br />

1 2 3<br />

0 −7 −4<br />

0 0 − 50<br />

7<br />

1 2 3<br />

0 −7 −4<br />

0 0 − 50<br />

7<br />

2 Recall from Section 1.2 that if A = LU, then the i, j entry <strong>of</strong> A can be obtained by multiplying the ith row <strong>of</strong> L<br />

times the jth column <strong>of</strong> U.<br />

⎤<br />

⎦ .<br />

⎤<br />

⎦<br />

⎤<br />

⎦ .<br />

⎤<br />

⎥ .<br />

⎥<br />

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

Saved successfully!

Ooh no, something went wrong!