MTH 102: Linear Algebra

Department of Mathematics and Statistics Indian Institute of Technology - Kanpur

Problem Set 1

Problems marked (T) are for discussions in Tutorial sessions.

1. If A is an m × n matrix, B is an n × p matrix and D is a p × s matrix, then show that

A(BD) = (AB)D (Associativity holds).

Solution: Entry by entry for 1 ≤ i ≤ m and 1 ≤ j ≤ s, we have

n n p p
n X
!
X X X X
[A(BD)]ij = [A]ik [BD]kj = [A]ik [B]kl [D]lj = [A]ik [B]kl [D]lj
k=1 k=1 l=1 k=1 l=1
p Xn p n
!
X X X
= [A]ik [B]kl [D]lj = [D]lj [A]ik [B]kl
l=1 k=1 l=1 k=1
p
X p
X
= [D]lj [AB]il = [AB]il [D]lj = [(AB)D]ij .
l=1 l=1

Hence the result.

2. If A is an m × n matrix, B and C are n × p matrices and D is a p × s matrix, then show that

(a) A(B + C) = AB + AC (Distributive law holds).

Solution: Entry by entry for 1 ≤ i ≤ m and 1 ≤ j ≤ p, we have
n
X n
X
[A(B + C)]ij = [A]ik [B + C]kj = [A]ik ([B]kj + [C]kj )
k=1 k=1
Xn n
X n
X
= ([A]ik [B]kj + [A]ik [C]kj ) = [A]ik [B]kj + [A]ik [C]kj
k=1 k=1 k=1
= [AB]ij + [AC]ij = [AB + AC]ij .

Hence the result.

(b) (B + C)D = BD + CD (Distributive law holds).
Solution: Similar to part (a) with appropriate modifications.
   
x x
3. (T) Let A and B be 2 × 2 real matrices such that A =B for all (x, y) ∈ R2 . Prove that
y y
A = B.
   
a11 a12 b11 b12
Solution: Let A = and B = . The given equation imply
a21 a22 b21 b22
       
a11 a12 b11 b
x +y =x + y 12 (1)
a21 a22 b21 b22
 2

Now, by substituting x = 1 and y = 0 in (1), we get

   
a11 b
= 11 (2)
a21 b21
Similarly, by substituting x = 0 and y = 1 in (1), we get
   
a12 b
= 12 (3)
a22 b22
Equations (2) and (3) together imply the result.
4. Let A and B be m × n real matrices such that Ax = Bx for all x ∈ Rn . Then, A = B
5. (A + B)∗ = A∗ + B ∗ and (AB)∗ = B ∗ A∗ whenever A + B and AB are defined.

Solution: Let A and B be m × n matrices. Then, for 1 ≤ i ≤ m and 1 ≤ j ≤ n, we have

[(A + B)T ]ij = [A + B]ji = [A]ji + [B]ji = [AT ]ij + [B T ]ij = [AT + B T ]ij .
Hence the result.

Solution: Let A be an m × n matrix and B be an n × p matrix. Then, entry by entry for

1 ≤ i ≤ p and 1 ≤ j ≤ m, we have
n
X n
X n
X
T T T
[(AB) ]ij = [AB]ji = [A]jk [B]ki = [A ]kj [B ]ik = [B T ]ik [AT ]kj = [B T AT ]ij .
k=1 k=1 k=1

Hence the result.

6. Let A ∈ Mn (C). Then A = S + T , where S ∗ = S (Hermitian matrix) and T ∗ = −T (skew-
Hermitian matrix).
7. Give examples of 3 × 3 non zero matrices A and B such that A2 = 0 and B 3 = B.
8. Show by an example that if AB 6= BA then (A + B)2 = A2 + 2AB + B 2 need not hold.
       
1 0 0 1 0 1 0 0
Solution: Consider A = and B = . Clearly, AB = 6= = BA.
0 0 −1 0 0 0 −1 0
A straightforward calculation shows that
   
2 0 1 0 2
(A + B) = 6= = A2 + 2AB + B 2 .
−1 −1 0 −1

9. Let A, B ∈ Mn (C) be invertible matrices. Then (AB)−1 = B −1 A−1 .

Solution: Let D = B −1 A−1 . Then

(AB)D = (AB)(B −1 A−1 ) = A(BB −1 )A−1 = AIA−1 = AA−1 = I
and
D(AB) = (B −1 A−1 )(AB) = B −1 (A−1 A)B = B −1 IB = B −1 B = I
imply that D is the inverse of AB.
 3

10. Let A ∈ Mn (C) be a nilpotent matrix. Then show that I + A is invertible.

Solution: As A is nilpotent, there exists an N > 0 such that AN = 0. Define

N
X −1
B= (−1)n An .
n=0

We have
N −1 N −1 N −1
!
X X X
n n
(I + A)B = (I + A) (−1) A = (−1)n An + (−1)n An+1
n=0 n=0 n=0
N
X −1 N
X −1
= (−1)n An − (−1)n An = I
n=0 n=1

and
N −1 N −1 N −1
!
X X X
n n
B(I + A) = (−1) A (I + A) = (−1)n An + (−1)n An+1
n=0 n=0 n=0
N
X −1 N
X −1
= (−1)n An − (−1)n An = I
n=0 n=1

and, therefore, B is the inverse of I + A.

n
P
11. (T) Let A, B ∈ Mn (C). Define Tr (A) = aii . Then show that Tr (AB) = Tr (BA). Hence
i=1
or otherwise, show that if A is invertible then Tr (ABA−1 ) = Tr (B). Furthermore, show that
there do not exist matrices A and B such that AB − BA = cI, for any c 6= 0.

Solution: Tr (AB) = Tr (BA) follows from a straightforward calculation shown below:

n
X n X
X n n X
X n n X
X n n
X
[AB]ii = aij bji = aij bji = bji aij = [BA]jj .
i=1 i=1 j=1 j=1 i=1 j=1 i=1 j=1

Now let D = BA−1 . We have,

Tr(ABA−1 ) = Tr(AD) = Tr(DA) = Tr(BA−1 A) = Tr(B).

12. Let A ∈ Mn (C). If AA∗ = 0 then show that A = 0.

Solution:
n
X n X
X n n X
X n
AA∗ = 0 ⇒ Tr(AA∗ ) = 0 ⇒ [AA∗ ]ii = 0 ⇒ [A]ij [A∗ ]ji = 0 ⇒ [A]ij [A]ij = 0.
i=1 i=1 j=1 i=1 j=1

We, therefore, have [A]ij = 0 for all 1 ≤ i ≤ n, 1 ≤ j ≤ n and thus A = 0.

 4

13. (T) The parabola y = a + bx + cx2 goes through the points (x, y) = (1, 4), (2, 8) and (3, 14).
Find and solve a matrix equation for the unknowns (a, b, c).

Solution: As the parabola passes through point (1, 4), we have 4 = a + b · 1 + c · 12 leading to
the equation a + b + c = 4.
Similarly for points (2, 8) and (3, 14), we get a + 2b + 4c = 8 and a + 3b + 9c = 14.
We can obtain a, b and c as a solution to
    
1 1 1 a 4
1 2 4   b = 8 .
 
1 3 9 c 14

Carry out Gauss-elimination as follows:

     
1 1 1 4 1 1 1 4 1 1 1 4
R2 ←R2 −R1 R ←R −2R2
 1 2 4 8 − −−−−−−→  0 1 3 4  −−3−−−3−−−→  0 1 3 4 
R3 ←R3 −R1
1 3 9 14 0 2 8 10 0 0 2 2

We can thus obtain the solution to the given linear system by solving the equivalent system

a+b+c = 4
b + 3c = 4
2c = 2

The solution is a = 2, b = 1 and c = 1.

14. (T) Let J = 11∗ . Then each entry of J equals 1. Determine condition(s) on a and b such that
bJ + (a − b)In is invertible. Find α and β in terms of a and b such that the inverse has the form
αJ + βI.

Solution: Check that each entry of J equals 1 and J 2 = nJ. The symmetry of the matrix
bJ + (a − b)I motivates us to try to assume that αJ + βI may be the inverse for some choice of
1 b
α and β. Verify β = and α = .
a−b (b − a)((n − 1)b + a)

15. (T) Let x ∈ M3,1 (R). Then find y, z ∈ M3,1 (R) such that xT y = 0 and xT z = 0.
 
x1
Solution: Let x = x2 . Choose y and z such that their dot product with x is zero.
x3

16. (T) Let A be an upper triangular matrix. If AA∗ = AA∗ then A is a diagonal matrix.
n n
Solution: (AA∗ )11 = a1i a1i and (A∗ A)11 = a11 a11 . Thus,
P P
a1i a1i = 0 and hence a1i = 0
i=1 2=1
for all i 6= 1. Now, consider the (2, 2)-entry of both sides and continue the above argument.