MATH2033 (2023–2024)

Introduction to Scientific Computation

Linear Systems: Direct Methods I

PROBLEM SHEET SOLUTIONS

Solution to Problem 1
a Forward elimination:

1 −1 3
 
2
 3 −3 1 −1 
1 1 0 3
1 −1
 
3 2  3

R2 − 1
R1 → R 2
∼  0 0 −8 −7 
R3 − 11 R1 → R3
0 2 −3 1
1 −1
 
3 2


∼  0 2 −3 1  R2 ↔ R3
0 0 −8 −7
1 −1
 
3 2


∼  0 2 −3 1  R3 − 02 R2 → R3
0 0 −8 −7

Backward substitution:
−7 7
x3 = = = 0.875,
−8 8
 
1 1 21 29
x2 = (1 − (−3)x3 ) = 1+ = = 1.8125,
2 2 8 16
    
1 1 29 21 1 13 19
x1 = (2−((−1)x2 +3x3 )) = 2− − + = 2− = = 1.1875.
1 1 16 8 1 16 16

b Forward elimination:

2 −1.5 3 1
 
 −1 0 2 3 
4 −4.5 5 1

Page 1 of 5
 2 −1.5
 
3 1  −1

R2 − 2
R 1 → R2
∼  0 −0.75 3.5 3.5 
R3 − 42 R1 → R3
0 −1.5 −1 −1
2 −1.5
 
3 1  
−1.5
∼  0 −0.75 3.5 3.5  R3 − −0.75 R2 → R3
0 0 −8 −8

Backward substitution:
−8
x3 = = 1,
−8
1 1
x2 = (3.5 − 3.5x3 ) = (3.5 − 3.5) = 0,
−0.75 −0.75
1 1 −2
x1 = (1 − ((−1.5)x2 + 3x3 )) = (1 − 3) = = −1.
2 2 2
c Forward elimination:
 
2 0 0 0 3
1 1.5 0 0 4.5
 
 
0 −3 0.5 0 −6.6
 
 
2 −2 1 1 0.8
 
2 0 0 0 3
R2 − 12 R1 → R2
 
0 1.5 0 0 3  
 
∼  R3 − 02 R1 → R3 

0 −3 0 −6.6 
0.5

R4 − 22 R1 → R4

0 −2 1 1 −2.2
 
2 0 0 0 3
R3 − −3
  
0 1.5 0 0 3  R → R3

∼ 
 1.5 2
0 0 0.5 0 −0.6  R4 − −2 R → R4

1.5 2

0 0 1 1 1.8
 
2 0 0 0 3
0 1.5 0 0 3 
 

1
∼   R4 − 0.5 R3 → R4

 0 0 0.5 0 −0.6 
0 0 0 1 3

Backward substitution:
3
x4 = = 3,
1
1 −0.6
x3 = (−0.6 − 0x4 ) = = −1.2,
0.5 0.5
1 3
x2 = (3 − (0x3 + 0x4 )) = = 2,
1.5 1.5
1 3
x1 = (3 − (0x2 + 0x3 + 0x4 )) = = 1.5.
2 2

Page 2 of 5
 d Forward elimination:
 
1 1 0 1 2
2 1 −1 1 1
 
 
4 −1 −2 2 0
 
 
3 −1 −1 2 −3
 
1 1 0 1 2
R2 − 21 R1 → R2
 
0 −1 −1 −1 −3  
 
∼  R3 − 41 R1 → R3 

0 −5 −2 −2 −8 

R4 − 31 R1 → R4

0 −4 −1 −1 −9
 
1 1 0 1 2 !

0 −1 −1 −1 −3 
 R3 − −5
−1
R 2 → R3
∼ 

0 0 3 3 7  R4 − −4 R → R4

−1 2

0 0 3 3 3
 
1 1 0 1 2
0 −1 −1 −1 −3 
 

∼   R4 − 33 R3 → R4

 0 0 3 3 7 
0 0 0 0 −4

The linear system does not have a unique solution. The linear system has
no solutions since −4 6= 0.

Solution to Problem 2
The number of multiplications and divisions (combined) performed is

n n n n−1
X X X X 2
2(i−1)+ 1=2 (i−1)+n = 2 j+n = (n−1)n+n = n2 −n+n = n2 .
i=2 i=1 i=2 j=1
2

Solution to Problem 3
We start with the augmented matrix

4 −1 1 8
 
 2 5 2 3 .
1 2 4 11

Using 2-digit arithmetic with rounding yields

2 4 = 0.5,

5 0.5 ⊗ (−1) = 5 (−0.5) = 5.5,

2 0.5 ⊗ 1 = 2 0.5 = 1.5,
3 0.5 ⊗ 8 = 3 4 = −1,
1 4 = 0.25,

Page 3 of 5
 2 0.25 ⊗ (−1) = 2 (−0.25) = f l(2.25) = 2.3,
4 0.25 ⊗ 1 = 4 0.25 = f l(3.75) = 3.8
and
11 0.25 ⊗ 8 = 11 2 = 9.
So performing R2 − 24 R1 → R2 and R3 − 14 R1 → R3 using 2-digit arithmetic
with rounding yields
4 −1
 
1 8
 0 5.5 1.5 −1  .
0 2.3 3.8 9
Using 2-digit arithmetic with rounding yields

2.3 5.5 = f l(0.418) = 0.42,

3.8 0.42 ⊗ 1.5 = 3.8 0.63 = f l(3.17) = 3.2,

and
9 0.42 ⊗ (−1) = 9 (−0.42) = f l(9.42) = 9.4.
2.3
So performing R3 − R
5.5 2
→ R3 using 2-digit arithmetic with rounding yields

4 −1
 
1 8
 0 5.5 1.5 −1  .
0 0 3.2 9.4
Performing backward substitution using 2-digit arithmetic with rounding yields

x3 = 9.4 3.2 = f l(2.93) = 2.9,

x2 = (−1 1.5 ⊗ x3 ) 5.5

= (−1 f l(4.35)) 5.5
= (−1 4.4) 5.5
= −5.4 5.5
= f l(−0.981)
= −0.98

and

x1 = (8 ((−1) ⊗ x2 ⊕ 1 ⊗ x3 )) 4
= (8 (0.98 ⊕ 2.9)) 4
= (8 f l(3.88)) 4
= (8 3.9) 4
= 4.1 4
= f l(1.02)
= 1.0.

Page 4 of 5
References

• Burden, Faires & Burden, Numerical Analysis, 10E

– Section 6.1

Page 5 of 5