SULTAN QABOOS UNIVERSITY-COLLEGE OF SCIENCE

DEPARTMENT OF MATHEMATICS AND STATISTICS

MATH 3171- LINEAR ALGEBRA AND MULTIVARIATE CALCULUS FOR ENGINEERS
SPRING 2016 (MARCH 13, 2016)
MIDTERM EXAM I
TIME ALLOWED: 1h 15. TOTAL MARKS: 40
SOLUTION

Problem 1. 1 1 2 0
(8 marks) Let A = ;u= ; and v = :
3 5 1 2

(a) Evaluate AuvT :

(b) Find the size of a matrix B if BAu is a 3 1 matrix? Justify your answer.
T 0 12
(c) Find a 2 1 vector x such that xT A vT = :
0 4
(d) Is the expression Au + vAT well de…ned? Justify your answer.

SOLUTION. 1 1 2 0
Let A = ;u= ; and v = :
3 5 1 2

(a)
T
1 1 2 0
AuvT =
3 5 1 2
3
= 0 2
1
0 6
= :
0 2

(b) Suppose that B is an m n matrix. Since A is a 2 2 matrix, the product BA is well

de…ned if n = 2: Moreover, the matrix BAu is of size m 1: So if BAu is a 3 1 matrix,
then m = 3: Hence the matrix B is a 3 2 matrix.
x
(c) Let x = : Then
y
T
xT A vT = AT xvT
!
T T
1 1 x 0
=
3 5 y 2
T
1 3 x 0
=
1 5 y 2
x 3y
= 0 2
x + 5y
0 2x 6y
=
0 2x + 10y

1
 So the equation

T 0 1 0 2x 6y 0 12
xT A vT = is equivalent to =
0 5 0 2x + 10y 0 4

T 0 1 0 2x 6y 0 12
xT A vT = () =
0 5 0 2x + 10y 0 4
2x 6y = 12
()
2x + 10y = 4
x 3
() = :
y 1

(d) Since vAT is not well de…ned (v is a 2 1 and AT is a 2 2), the expression Au + vAT
is not well de…ned.
Problem 2. (7 marks) Solve by Cramer’s rule the following system
8
< 3x + 2y = 0
2y 3z = 0 (1)
:
2x + z = 6

SOLUTION. The coe¢ cient matrix of the system (1) is

2 3
3 2 0
A=4 0 2 3 5:
2 0 1

Then
D1 D2 D2
x= ; y= ; and z = ;
D D D
where
3 2 0
D = det A = 0 2 3 = 18;
2 0 1
0 2 0 3 0 0 3 2 0
D1 = 0 2 3 = 36; D2 = 0 0 3 = 54; and D3 = 0 2 0 = 36
6 0 1 2 6 1 2 0 6
Hence
36 54 36
x= = 2; y = = 3; and z = = 2:
18 18 18
Actually it is enough to …nd one of the variables and deduce the two others from the system
equations.
2 3 2 3 2 3
0 4 2 2 x1
Problem 3. (9 marks) A = 4 6 2 1 5 ; b = 4 29 5 ; and x = 4 x2 5 :
4 8 4 24 x3

(a) Use Gauss elimination method to solve the system Ax = b:

(b) Find the rank of the matrix A (Hint: use part (a)).
(c) Are the row vectors of the matrix A linearly independent? Justify your answer.

2
SOLUTION. (a) From Gauss method, we get
2 3 2 3
0 4 2 2 interchange R1 and 1
R
1 2 1 6
4 6 4 3
2 1 j 29 5 ! 4 6 2 1 j 29 5
4 8 4 24 0 4 2 2
1
2 3 7 R2 2 3 2 3
1 2 1 6 1 1 2 1 6 1 2 1 6
R2 6R 2 R2 R R
!14 0 14 7 j 7 5 ! 4 0 2 1 j 1 5 3 !2 4 0 2 1 j 1 5:
0 4 2 2 0 2 1 1 0 0 0 0
So the system Ax = b is equivalent to
x1 + 2x2 x3 = 6
:
2x2 x3 = 1
So we can take x2 = t is a free variable, hence
x3 = 2x2 1
= 2t 1
x1 = 6 2x2 + x3
=6 2t + 2t 1
= 5:
Hence the system has an in…nite number of solution, given by
2 3 2 3
x1 5
x = 4 x2 5 = 4 t 5 ; t arbitrary real number.
x3 2t 1

(b) From part (a) ; the rank of the matrix A is equal to 2:

(c) Since rankA = 2; the rows of the matrix A are linearly dependent.
2 3
3 2 6
Problem 4. (7 marks) Let A be a 3 3 matrix with det A = 2; and B = 4 1 12 1 5:
4 14 7

(a) Evaluate det 2AT B2016 .

(b) Evaluate det A 3 + det 2AT .

SOLUTION. (a) Since the second row of the matrix B is the sum of the …rst and third rows, we deduce
that det B = 0: Hence
det 2AT B2016 = 23 det AT det B2016 (since for a 3 3 matrix C; we have det (2C) = 23 det (C) )
= 8 det (A) (det (B))2016 (since det AT = det (A) and det B2016 = (det (B))2016 )
= 0:

(b)
3 3
det A + det 2AT = (det (A)) + 23 det (A)
= ( 2) 3 + 8 ( 2)
129
= :
8

3
 2 3
1 0 1
Problem 5. (9 marks) Let A = 4 0 3 4 5:
0 3 5

(a) Find the eigenvalues of A:

(b) Find the eigenvectors of A:
(c) Find the defect of each eigenvalue.
2 3
1 0 1
SOLUTION. Let A = 4 0 3 4 5:
0 3 5

(a) The characteristic polynomial DA of the matrix A is given by

DA ( ) = det (A I3 )
1 0 1
= 0 3 4
0 3 5
3 4
= (1 )
3 5
2
= (1 ) +2 3
2
= ( 1) ( + 3) :
So the eigenvalues of the matrix A are

1 = 1 with algebraic multiplicity = 2

2 = 3 with algebraic multiplicity = 1:

(b) Denote by E1 the eigenspace corresponding to the eigenvalue 1 = 1: Then

82 3 2 32 3 2 39
< x 0 0 1 x 0 =
E1 = 4 y 5 2 R3 j 4 0 2 4 54 y 5 = 4 0 5
: ;
z 0 3 6 z 0
Using Gauss method, we deduce that
82 3 9
< x =
y = 0
E1 = 4 y 5 2 R3 j
: z = 0 ;
z
82 3 9
< x =
= 4 0 5 2 R3 j x arbitrary real number
: ;
0
82 39
< 1 =
= span 4 0 5 :
: ;
0
So the geometric multiplicity of 1 = 1 is equal to 1:
Similarly, by E2 the eigenspace corresponding to the eigenvalue 2 = 3: Then
82 3 2 3 2 3 2 39
< x 4 0 1 x 0 =
E1 = 4 y 5 2 R3 j 4 0 6 4 54 y 5 = 4 0 5
: ;
z 0 3 2 z 0

4
 Using Gauss method, we deduce that
82 3 9
< x =
4x + z = 0
E 1 = 4 y 5 2 R3 j
: 3y 2z = 0 ;
z
82 1 3 9
< 4z =
= 4 23 z 5 2 R3 j z arbitrary real number
: ;
z
82 39
< 3 =
= span 4 8 5 :
: ;
12

So the geometric multiplicity of 2 = 3 is equal to 1:

(d) Denote by ( 1 ) (resp. ( 2 )) the defect of the eigenvalue 1 = 1 (resp. 2 = 3):
From part (b) ; we deduce that

( 1) = algebraic mult. ( 1) geometr. mult. ( 1)

=2 1
= 1:

Similarly

( 2) = algebraic mult. ( 2) geometr. mult. ( 2)

=1 1
= 0:

END OF SOLUTION