Download as doc, pdf, or txt
Download as doc, pdf, or txt
You are on page 1of 1

Gaussian Elimination Homework Assignment (with problem #’s fixed, 9-26-2012)

Read in Watkins: pages 3-7 (operation count) and 70-106 (Gaussian elimination)

1. Consider , , and .

(a) Apply Gaussian elimination without pivoting to A to determine L and U such that A = LU.
(b) Use part (a) to solve Ax = b1.
(c) Use part (a) to solve Ax = b2.

2. For the same matrix as in problem 1:


(a) Apply Gaussian elimination with partial pivoting to A to determine P, L and U such that PA = LU..
(b) Use part (a) to solve Ax = b1.
(c) Use part (a) to solve Ax = b2.

3. For the same matrix as in problem 1 apply complete pivoting to decompose A. Determine L, U, P and Q such that
PAQ = LU.

4. Solve Ax = b for and by

(a) Using Gaussian elimination with no pivoting assuming all calculations are done with 3 significant digit arithmetic
(so in all your calculations keep three digits starting at the leftmost non-zero digit).
(b) Using Gaussian elimination with pivoting assuming all calculations are done with 3 significant digit arithmetic
(c) Which answer is better? Why?

5. Count, approximately, the floating point operations (flops) in : Form y = Ax where A is m by n and x is n by 1 using:
for i = 1, 2, … m

yi =

6. Count, approximately, the floating point operations (flops) in : Form C = AB where A, B and C are n by n using:
for i = 1, 2, … n
for j = 1, 2, … n

7. Count, approximately, the floating point operations (flops) in : Form y = Ax where A is n by n upper triangular and x
is n by 1 using:
for i = 1, 2, … n

yi =

8. Count, approximately, the floating point operations (flops) in : Form C = AB where A, B and C are n by n upper
triangular matrices using:
for i = 1, 2, … n
for j = i, i+1, … n

9. Watkins exercise 1.7.39, page 87.

10. Watkins exercise 1.7.44, page 88

11. Watkins exercise 18.20, page 102.

You might also like