You are on page 1of 1

MATH 2346C - Computational Linear Algebra Tutorial

Tutorial
   
1 2 5
1. Consider the linear system Ax = b, where A = and b = . Let δb =
3 4 11
 
1
10−3 and consider the perturbed system A (x + δx) = b + δb.
1

Verify that
kδxk2 1
≥ kδbk2 .
kxk2 κ2 (A)
p
Note: kAk2 = maximum eigenvalue of AT A.
   
4 2 2
2. Consider the linear system Ax = b, where A = and b = . Let ∆A =
3 2 1
 
10−4 0
and consider the perturbed system (A + ∆A) (x + δx) = b.
0 0

Verify that
kδxk2 κ2 (A) k∆Ak2 /kAk2
≤ ·
kxk2 1 − κ2 (A) k∆Ak2 /kAk2

3. Consider the linear system


    
5 1 −1 x1 −6
 −1 −3 1   x2  =  2  . (1)
2 0 4 x3 2

Starting with x(0) = 0, perform two iterations of the


(a) Jacobi method;
(b) Gauss-Seidel method;
(c) SOR method with ω = 1.2.
For each the above-mentioned methods, compute kx(2) − xe k2 , where xe is the exact solution
of (1).
4. Solve the linear system
    
1 −1 −1 −1 x1 −2
 −1 26 11 −14  x2 
 =  −33
  
  ,
 −1 11 14 −2   x3   −27 
−1 −14 −2 15 x4 22

using
(a) the inner product formulation of Cholesky’s method.
(b) the outer product formulation of Cholesky’s method.

Page 1 of 1

You might also like