MAT3701

May/June 2011
LINEAR ALGEBRA

Duration : 2 Hours 100 Marks

EXAMINERS :
FIRST : PROF JD BOTHA
SECOND : PROF TA DUBE
EXTERNAL : PROF LM PRETORIUS (PRETORIA - UP)

This paper consists of 3 pages

ANSWER ALL THE QUESTIONS.

QUESTION 1
Let V be the vector space C 2 with scalar multiplication over the real numbers R. It is given that

β = {(1, 0) , (i, 0) , (0, 1) , (0, i)}

is a basis for V. The mapping T : V → V is defined by

T (z1 , z2 ) = (z1 − z 1 , z2 + z 2 ) ,

where z1 and z2 are complex numbers and z denotes the complex conjugate of z.

(a) Show that T is a linear operator. (6)

(b) Find a basis for N (T ) . (6)
(c) Find a basis for R (T ) . (6)
(d) Determine whether V = N (T ) ⊕ R (T ) . (2)

[20]

QUESTION 2
Let T : P2 (C) → P2 (C) be the linear operator defined by

T a + bx + cx2 = (a + c) + (b + c) x + 2cx2 .

(a) Show that T satisfies the test for diagonalisability. (10)

(b) Find a basis τ for P2 (C) consisting of eigenvectors of T, and write down [T ]τ . (7)
(c) Determine whether T − IP2 (C) is a projection. (3)

[20]

Open Rubric
 MAT3701/102

QUESTION 3
Let
1
 
0 0 

 2 
 
 1 1 
A= .
1 

 2 2 
 
 1 
0 0
2
(a) Show that T is a regular transition matrix. (4)

(b) Find lim Am . (8)

m→∞

(c) Describe the Gerschgorin discs in which the eigenvalues of A lie. (4)

[16]

QUESTION 4
Let P denote the orthogonal projection of C 3 on
 
1 1
W = span √ (1, 0, i) , √ (i, 0, 1)
2 2

(a) Find the formula for P (z1 , z2 , z3 ) . (7)

(b) Find the eigenvalues and corresponding eigenspaces of P. (4)

(c) Find the vector in W closest to (1, 1, 1) ∈ C 3 . (2)

[13]

QUESTION 5
It is given that A ∈ M3×3 (C) is a self-adjoint matrix with eigenvalues 1 and 2, and corresponding
eigenspaces
 
1
E1 = span (0, 1, 0) , √ (i, 0, 1)
2
 
1
E2 = span √ (1, 0, i)
2
Find the spectral decomposition of A.
[15]

3
 QUESTION 6
Let  
−1 3
A= .
3 −1

(a) Find kAk , kA−1 k and cond (A) . (9)

(b) Suppose x and x e are vectors such that Ax = b, kbk = 1, and kb − Ae xk ≤ 0.001. Use (a)
to determine upper bounds for ke x − A−1 bk (the absolute error) and ke
x − A−1 bk / kA−1 bk (the
relative error). (7)

[16]

TOTAL: [100]

Memorandum: May/June 2011 Exam

Question 1

(a) T ((z1 , z2 ) + (z3 , z4 )) = T (z1 + z3 , z2 + z4 )

= (z1 + z3 − (z1 + z3 ) , z2 + z4 + (z2 + z4 ))
= (z1 + z3 − z 1 − z 3 , z2 + z4 + z 2 + z 4 )
= (z1 − z 1 , z2 + z 2 ) + (z3 − z 3 , z4 + z 4 )
= T (z1 , z2 ) + T (z3 , z4 )

T (a (z1 , z2 )) = T (az1 , az2 )

= (az1 − az 1 , az2 + az 2 )
= (az1 − az1 , az2 + az2 ) , a real
= a (z1 − z 1 , z2 + z 2 )
= aT (z1 , z2 )

Thus T is linear. (6)

(b) T (z1 , z2 ) = (z1 − z 1 , z2 + z 2 ) = 0 ⇔ z1 = z 1 , z2 = −z 2

∴ z1 = a real and z2 = ib imaginary
∴ (z1 , z2 ) = (a, ib) = a (1, 0) + b (0, i)
∴ N (T ) = span {(1, 0) , (0, i)} with basis α = {(1, 0) , (0, i)} (6)

(c) T (z1 , z2 ) = (z1 − z 1 , z2 + z 2 )

= (2ia, 2b) , a, b real
= 2a (i, 0) + 2b (0, 1)
∴ R (T ) = span {(i, 0) , (0, 1)} with basis β = {(i, 0) , (0, 1)} (6)

4
 MAT3701/102

(d) Since α ∪ β is a basis for V, it follows that V = N (T ) ⊕ R (T ) (2)

[20]

Question 2

(a) Let β = {1, x, x2 } .

T (1) = 1
T (x) = x
T (x2 ) = 1 + x + 2x2
 
1 0 1
∴ [T ]β =  0 1 1 
0 0 2
∴ c (T ) = (x − 1)2 (x − 2)
 
0 0 1
[T ]β − I3 =  0 0 1  ,
0 0 1
 
hence rank [T ]β − I3 = 1 = 3 − (multiplicity of λ = 1)
∴ T is diagonalisable. (10)

(b) From (a),

 
E1 [T ]β = span {(1, 0, 0) , (0, 1, 0)}
∴ E1 (T ) = span {1, x}
 
−1 0 1
λ = 2 [T ]β − 2I3 =  0 −1 1 
0 0 0
 
E2 [T ]β = span {(1, 1, 1)}
∴ E2 (T ) = span {1 + x + x2 }
 
1 0 0
Let τ = {1, x, 1 + x + x2 } , then Mτ (T ) =  0 1 0  (7)
0 0 2

(c) T − IP2 (C) is a projection since

 
0 0 0
[T − IP2 (C)]τ =  0 0 0  = B and B 2 = B. (3)
0 0 1

[20]

5
Question 3

(a) All three column sums of A are one, hence it is a transition matrix. Further,
1 1 1 3 1 1
   
 4 4 2   8 4 4 
   
   
2
 3 1 1   3
 1 5 1 
A = 
, A =  2 8 2 
 
 4 2 2   
   
 1   1 1 1 
0 0
4 8 8 4
3
Since all entries of A are positive, A is regular. (4)

1
 
 −1 2
0 
 
 
 1 1 
(b) A − I = 
 2 − 1  , so that
 2 

 
 1 
0 −1
2
   
1 1
E1 (A) = span 1, 2, = span (2, 4, 1)
2 7
1
Since (2, 4, 1) is a probability vector,
7
 
2 2 2
1
lim Am =  4 4 4 
m→∞ 7
1 1 1

(8)

 
1
(c) D1 = z : |z| ≤
2
 
1 3
D2 = z : z − ≤
2 2
 
1
D3 = z : |z| ≤ = D1 (4)
2

[16]

6
 MAT3701/102

Question 4
 
1 1
(a) P (z1 , z2 , z3 ) = (z1 , z2 , z3 ) , √ (1, 0, i) √ (1, 0, i)
2 2
 
1 1
+ (z1 , z2 , z3 ) , √ (i, 0, 1) √ (i, 0, 1)
2 2

1 1
= (z1 − iz3 ) (1, 0, i) + (−iz1 + z3 ) (i, 0, 1)
2 2
1
= (2z1 , 0, 2z3 )
2

= (z1 , 0, z3 )

(7)

(b) E1 (P ) = span {(1, 0, 0) , (0, 0, 1)}

E0 (P ) = N (P ) = span {(0, 1, 0)} (4)

(c) P (1, 1, 1) = (1, 0, 1) (2)

[13]

Question 5
Let  i 1 
0 √ √

 2 2 

 
P = 1 0 0 ,
 
 
 
 1 i 
0 √ √
2 2
then
 
1 0 0
A = P  0 1 0 P∗
0 0 2
   
1 0 0 0 0 0
= P  0 1 0  P ∗ + 2P  0 0 0  P ∗
0 0 0 0 0 1

7
 i
 
0 1 0
  1  0 1 0

0 √ 0 0 0 √
 2     2 
  
  i 1   i 1 
  −√ 0 √    −√ 0 √
     
=  1 0 0  + 2 0 0 0
  
  2 2  



 2 2 

     
 1  1 i   i  1 i 
0 √ 0 √ 0 −√ 0 0 √ √ 0 −√
2 2 2 2 2 2
   
1 i 1 i
 2 0 2   2 0 − 
   2 
   
=  0 1 0  + 2 0 0 0 
   
   
   
 i 1   i 1 
− 0 0
2 2 2 2
[15]

Question 6
   
10 −6 5 −3
(a) A∗ A = =2
−6 10 −3 5
(5 − λ)2 − 32 = (5 − λ − 3) (5 − λ + 3) = (2 − λ) (8 − λ)
The eigenvalues of A∗ A are 16 and 4.
1
∴ kAk = 4, kA−1 k = and cond (A) = 2 (9)
2

1
x − A−1 bk = kA−1 (Ae
(b) ke x − b)k ≤ kA−1 k kAe
x − bk ≤ 0.001 = 0.0005
2
−1
ke
x − A bk kAe
x − bk
≤ cond (A) ≤ 0.001cond (A) = 0.002 (7)
kA−1 bk kbk

[16]

TOTAL [100]

8
 MAT3701/102

MAT3701
October/November 2011
LINEAR ALGEBRA

Duration : 2 Hours 100 Marks

EXAMINERS :
FIRST : PROF JD BOTHA
SECOND : PROF TA DUBE
This examination question paper remains the property of the University of South Africa
and may not be removed from the examination venue.

This paper consists of 2 pages

ANSWER ALL THE QUESTIONS.

QUESTION 1
Let V denote the vector space C 2 with scalar multiplication over the real numbers R, and let T :
V → V be the mapping defined by

T (z1 , z2 ) = (z1 − z 1 , z 2 ) .

(a) Show that T is a linear operator. (6)

(b) Find a basis for N (T ) , the null space of T. (5)

(c) Find a basis for R (T ) , the range of T. (5)

(d) Is λ = 0 an eigenvalue of T ? State your reason. (3)

(e) Given that λ = 1 is an eigenvalue of T, find an eigenvector of T corresponding to λ = 1. (5)

[24]

QUESTION 2
Let T : M2×2 (C) → M2×2 (C) be the linear operator defined by T (A) = At .

(a) Show that T satisfies the test for diagonalisability. (11)

(b) Find a basis τ for M2×2 (C) consisting of eigenvectors of T , and write down [T ]τ . (10)

(c) Show that T = T −1 . (4)

[25]

9
 QUESTION 3
Let
1 1
 
2 2
0
1
A= 0 0 2
.
1 1 1
2 2 2
(a) Show that A is a regular transition matrix. (3)
(b) Find lim Am . (8)
m→∞

(c) Give an example of a 2 × 2 transition matrix which is not regular. (4)

[15]

QUESTION 4
For the set of data
{(−2, 0) , (−1, 4) , (0, 2) , (1, 4)} ,
find the line that provides the best least-squares fit. Calculate the estimation error E. [13]

QUESTION 5
Let  
−1 −2 2
A =  −2 −2 2 .
2 2 −4
(a) Find a nonsingular matrix Q such that Qt AQ = D, where D is diagonal. (9)
(b) Write down the rank, index, and signature of A. (2)
[11]

QUESTION 6
Consider the inner product space C 2 over the scalar field C with the standard inner product h , i .
Let T : C 2 → C 2 be the linear operator defined by
1
T (z1 , z2 ) = √ (z1 + iz2 , iz1 + z2 ) .
2
(a) Show directly from the definition of T that
hT (z1 , z2 ) , T (z1 , z2 )i = h(z1 , z2 ) , (z1 , z2 )i for all (z1 , z2 ) ∈ C 2 . (10)

(b) In general, what is an operator called that satisfies the property in (a)? (2)
[12]
TOTAL: [100]

10
 MAT3701/102

Memorandum: October/November 2011 Exam

QUESTION 1

(a)
T [(z1 , z2 ) + (z3 , z4 )] = T (z1 + z3 , z2 + z4 )
= (z1 + z3 − z1 + z3 , z2 + z4 )
= (z1 + z3 − z 1 − z 3 , z 2 + z 4 )
= (z1 − z 1 , z 2 ) + (z3 − z 3 , z 4 )
= T (z1 , z2 ) + T (z3 , z4 )

and
T [a (z1 , z2 )] = T (az1 , az2 )
= (az1 − az 1 , az 2 )
= (az1 − az 1 , az 2 ) , since a is real
= a (z1 − z 1 , z 2 )
= aT (z1 , z2 )

(6)

(b)
T (z1 , z2 ) = (0, 0) ⇔ (z1 − z 1 , z 2 ) = (0, 0)
⇔ z1 = z 1 , z 2 = 0
⇔ z1 is real and z2 = 0
∴ N (T ) = {(a, 0) : a is real}
= span {(1, 0)}
Thus, {(1, 0)} is a basis.

(5)

(c)
T (a + ib, c + id) = (a + ib − (a − ib) , c − id)
= (2ib, c − id)
= 2b (i, 0) + c (0, 1) − d (0, i)
∴ R (T ) = span {(i, 0) , (0, 1) , (0, i)}
with {(i, 0) , (0, 1) , (0, i)} as basis.

(5)

(d) Yes, since N (T ) 6= {0} . (3)

(e) T (z1 , z2 ) = (z1 , z2 ) ⇔ (z1 − z 1 , z 2 ) = (z1 , z2 ) ⇔ z1 − z 1 = z1 and z 2 = z2 ⇔ z1 = 0 and z2 is

real.
Thus, v = (0, 1) is an eigenvector corresponding to λ = 1. (5)

[24]

11
 QUESTION 2
        
1 0 0 1 0 0 0 0
(a) Let β = E11 = ; E12 = ; E21 = ; E22 =
0 0 0 0 1 0 0 1

T (E11 ) = E11 ; T (E12 ) = E21 ; T (E21 ) = E12 ; T (E22 ) = E22

 
1 0 0 0
 0 0 1 0 
∴ [T ]β =  0

1 0 0 
0 0 0 1

λ−1 0 0 0
  0 λ −1 0
= (λ − 1)2 λ2 − 1


det λI − [T ]β =
0 −1 λ 0
= (λ − 1)3 (λ + 1)
0 0 0 λ−1
The eigenvalues of T are λ1 = 1, with multiplicity three, and λ2 = −1.
Now,  
0 0 0 0
I − [T ]β = 
 0 1 −1 0 
.
 0 −1 1 0 
0 0 0 0
 
Since rank I − [T ]β = 1, dim(E1 ) = 3 = multiplicity (λ1 ) , hence T is diagonalisable. (11)

(b) From the previous system, x1 = t1 , x2 = x3 = t2 , x4 = t3 , so that

        

 t1 
 
 1 0 0 

 t2  : t1 , t2 , t3 ∈ C  0 ,
      1   0 
E1 =  t2  = span  0 
 , 
 1   0
 ,


 
 
 
t3 0 0 1
   

hence      
1 0 0 1 0 0
τ1 = , ,
0 0 1 0 0 1
is a basis for E1 (T ) .
 
−2 0 0 0
 0 −1 −1 0 
For λ2 = −1, the system becomes  ,
 0 −1 −1 0 
0 0 0 −2
so that x1 = 0, x2 = t, x3 = −t and x4 = 0.
Thus,     

 0 
 
 0 

t  : t ∈ C = span  1
    
E−1 =  −t   −1
 ,


 
 
 
0 0
   

12
 MAT3701/102

hence  
0 1
τ2 = is a basis for E−1 (T ) .
−1 0
The required basis is therefore
       
1 0 0 1 0 0 0 1
τ = τ1 ∪ τ2 = , , , ,
0 0 1 0 0 1 −1 0

and  
1 0 0 0
 0 1 0 0 
[T ]τ = 
 0

0 1 0 
0 0 0 −1
(10)
(c) Since [T ]τ = [T ]−1
τ , it follows that T = T
−1
. (4)
[25]

QUESTION 3
(a) A is a transition matrix since all its entries are nonnegative and all three column sums are equal
to one.
Since
1 1 1
 
 4 4 4 
 
 
2
 1 1 1 
Since A = 
 4
 has only positive entries, A is regular as well.
 4 4 

 
 1 1 1 
2 2 2
(3)
(b)
1 1 1 1 1
     
− 0   2 −2 0   

 2 2    1  4 
 
       
 1   1  1   1 
I −A= 0 1 − ∼ 0 1 −   1
∴  = 
2  4
2 
     4 
     
     
 1 1 1   1  2  1 
− − 0 −1
2 2 2 2 2
R3 →R3 +R1

is an eigenvector corresponding to λ = 1, which is also a probability vector.

 
1 1 1
m 1
∴ lim A = 1 1 1 .
m→∞ 4
2 2 2

(8)

13
  
0 1
(c) E.g. . (4)
1 0
[15]

QUESTION 4

  
1 −2 0
 1 −1   4 
A=
 1
, B =   , y = a + bx
0   2 
1 1 4
      
∗ a ∗ 4 −2 a 10
AA =A B ∴ =
b −2 6 b 0
" . # " .. #
2 −1 .. 5 0 5 . 5 b = 1, a = 3
. ∼ .. ∴
−1 3 .. 0 −1 3 . 0 y =3+x

(11)

E = (y (−2) − 0)2 + (y (−1) − 4)2 + (y (0) − 2)2 + (y (1) − 4)2

= 12 + (−2)2 + 12 + 02 = 6

(2)
[13]

QUESTION 5

(a)
. ..
   
−1 −2 2 .. 1 0 0 −1 −2 2 . 1 0 0
 .   .. 
 −2 −2
 2 .. 0 1 0 
 →  0
 2 −2 . −2 1 0  
. ..
2 2 −4 .. 0 0 1 2 2 −4 . 0 0 1
R2 →R2 −2R1

. .
   
−1 2 ..
0 1 0 0 −1 2 ..
0 1 0 0
 .   . 
→
 0 2 −2 .. −2 1 0 
 →

 0 2 −2 .. −2 1 0 

.. ..
2 −2 −4 . 0 0 1 0 −2 0 . 2 0 1
C2 →C2 −2C1 R3 →R3 +2R1

.. ..
   
−1 0 0 . 1 0 0 −1 0 0 . 1 0 0
 .   . 
→ 0
 2 −2 .. −2 1 0  →  0 2 −2 .. −2 1 0 


.. ..
0 −2 0 . 2 0 1 0 0 −2 . 0 1 1
C3 →C3 +2C1 R3 →R3 +R2

14
 MAT3701/102

..
 
−1 0 0 . 1 0 0

→ .. 
 0 2 0 . −2 1 0  
..
0 0 −2 . 0 1 1
C3 →C3 +C2

Thus
 t    
1 0 0 1 −2 0 −1 0 0
Q =  −2 1 0  =  0 1 1  and D =  0 2 0 
0 1 1 0 0 1 0 0 −2

(9)

(b) Index (A) = 1; rank(A) = 3; signature (A) = −1. (2)

[11]

QUESTION 6

(a)
1


hT (z1 , z2 ) , T (z1 , z2 )i = (z1 + iz2 ) z1 + iz2 + (iz1 + z2 ) iz1 + z2
2
1
= [(z1 + iz2 ) (z 1 − iz 2 ) + (iz1 + z2 ) (−iz 1 + z 2 )]
2
1
= [z1 z 1 − iz1 z 2 + iz2 z 1 + z2 z 2 + z1 z 1 + iz1 z 2 − iz2 z 1 + z2 z 2 ]
2
1
= [2z1 z 1 + 2z2 z 2 ]
2

= z1 z 1 + z2 z 2

= h(z1 , z2 ) , (z1 , z2 )i

(10)

(b) Unitary. (2)

[12]

[100]

15