spanda@maths.iitkgp.ac.

in 1

Solution of Tutorial Problems set-I

Note: All these problems can be solved using the results of Chapter-I.

[0.0.1] Exercise Check that each of the following sets are vector space with respect to usual addition
and scalar multiplication.
(i) The set of all real sequences over the field F = R.
(ii) The set of all bounded real sequences over the filed R.
(iii) The set of all convergent real sequences over the field R.
(iv) {(an ) | an ∈ R, an → 0} over the field R.
(v) The set of all eventually 0 sequences over the field R.We call (xn ) eventually 0 if ∃k s.t. xn = 0
for all n ≥ k.
(vi) P(x)= {p(x) | p(x) is a real polynomial in x} over the field R.
(vii) P5 (x)= {p(x) ∈ R[x] | degree of p(x) ≤ 5} over the field R.
(viii) {An×n | aij ∈ R, A upper triangular} over the field R.

[0.0.2] Exercise Consider Pn (x) and P(x) over R. Check that each of the following sets is subspace or
not.
(i) {P (x) ∈ P3 (x) | P (x) = ax + b, a, b ∈ R}.
(ii) {P (x) ∈ P | P (0) = 0}.
(iii) {P (x) ∈ P | P (0) = 1}.
(iv) {P (x) ∈ P | P (−x) = P (x)}.
(v) {P (x) ∈ P | P (−x) = −P (x)}.

[0.0.3] Exercise Fix A ∈ Mn (R). Let U = {B ∈ Mn (R) : AB = BA}.

a) Show that U is a subspace of Mn (R).
b) Let W = {a0 I + a1 A + · · · an Am |m ∈ N0 := N ∪{0}, ai ∈ R}. Show that W is a subspace of U.

Sol. a) Let B, C ∈ U and α, β ∈ R. To show that αB + βC ∈ U. A[αB + βC] = αAB + βAC =

αBA + βCA = [αB + βC]A. Therefore αB + βC ∈ U. So U is a subspace.

b) Note that if r(A) is any polynomial, then Ar(A) = r(A)A. Hence r(A) ∈ U. Thus W ⊆ U.
Let p(A) and q(A) be two polynomials in A. Then αp(A) + βq(A) is also polynomial in A.
So αp(A) + βq(A) ∈ W. So W is a subspace.

[0.0.4] Exercise Find basis and dimension for each of the following vector spaces.
(i) Mn (C) over R.
(ii) Hn (C), n × n Hermitian matrices, over R.
(iii) Sn (C), n × n Skew-Hermitian matrices, over R.
spanda@maths.iitkgp.ac.in 2

Sol. (i) 2n2 .

(ii) n2 .
(iii)n2 .

[0.0.5] Exercise Check whether the following vector space is finite dimensional or infinite dimensional.
(i) The set of all real sequences over the field F = R.
(ii) The set of all bounded real sequences over the field R.
(iii) The set of all convergent real sequences over the field R.
(iv) {(an ) | an ∈ R, an → 0} over the field R.
(v) The set of all eventually 0 sequences over the field R. We call (xn ) eventually 0 if ∃k s.t. xn = 0
for all n ≥ k.
(vii) P(x)= {p(x) | p(x) is a real polynomial in x} over the field R.
(viii) P5 (x)= {p(x) ∈ R[x] | degree of p(x) ≤ 5} over the field R.

[0.0.6] Exercise Write 4 proper subspaces of R4 (a subspace W is called proper if neither W is trivial
and nor W equal to whole vector space).

 t
Sol. a) Put v1 = 1 1 1 1 . Then ls(v1 ) is a proper subspace.
 t
b) Put v2 = 0 1 1 1 . Then ls(v1 , v2 ) is a proper subspace.
 t
c) Put v3 = 0 0 1 1 . Then ls(v1 , v2 , v3 ) is a proper subspace.
Therefore ls v1 , ls v2 , ls v3 , ls(v1 , v2 ), ls(v1 , v2 , v3 ) are proper subspace of R3 . There are
many proper subspace of R3 .

The following is an extra information:

Claim Any proper subspace of R4 must be a span of either a) one nonzero vector or b) two
linearly independent vectors or c) three linearly independent vectors. Why?
Let W be a nontrivial subspace. So ∃ a nonzero v1 ∈ W. If W = ls(v1 ), it is of type a)
and we are done. So let v2 ∈ W − ls(v1 ). By our theorem, v1 , v2 are linearly independent. If
W = ls(v1 , v2 ), it is of type b) and we are done. So let v3 ∈ W − ls(v1 , v2 ). By our theorem,
v1 , v2 , v3 are linearly independent. If W = ls(v1 , v2 , v3 ), it is of type c) and we are done. So let v4 ∈
W−ls(v1 , v2 , v3 ). By our theorem, v1 , v2 , v3 , v4 are linearly independent. Then ls(v1 , v2 , v3 , v4 ) =
R4 . Hence W is

[0.0.7] Exercise Show that u1 , · · · , uk ∈ Rn are linearly independent iff Au1 , · · · , Auk are linearly
independent for any invertible An .
spanda@maths.iitkgp.ac.in 3

Sol. First we assume that Au1 , · · · , Auk are linearly independent. To show u1 , · · · , uk ∈ Rn are
linearly independent.
Suppose u1 , · · · , un are linearly dependent. Then there exits α1 , . . . , αk in F not all zero s.t.
P P P
αi ui = 0. Multiplying both side by A. So A αi ui = αi (Aui ) = 0. So Au1 , . . . , Aun
are linearly dependent. A contradiction because Au1 , · · · , Auk are linearly independent. Hence
u1 , · · · , uk ∈ Rn are LI.
We now assume that u1 , · · · , uk ∈ Rn are linearly independent. To show Au1 , · · · , Auk are
linearly independent.
P
Suppose Au1 , · · · , Aun are linearly dependent. Then ∃α 6= 0 s.t. αi (Aui ) = 0. So 0 =
P P P P
αi (Aui ) = A αi ui . So A−1 0 = A−1 A αi ui = αi ui . So u1 , . . . , un are linearly dependent.
A contradiction because u1 , · · · , uk ∈ Rn are linearly independent. Hence Au1 , · · · , Auk are LI.

k
P k
P
[0.0.8] Exercise Show that u1 , · · · , uk ∈ V is linearly independent iff ai1 ui , · · · , aik ui are linearly
i=1 i=1
independent for any invertible Ak×k . Show that {u, v} is linearly independent iff {u + v, u − v} is
linearly independent.

Sol. We first assume that u1 , . . . , uk are LI.

Pk
Put wr = air ui .
i=1
      
w1 a11 · · · ak1 u1 u1
 ..  .. .. .. = At  ...
. .
Then  . =  . . To show that w1 , . . . , wk are LI.
   

wr a1k · · · akk uk uk
Suppose w1 , . .. , wk are linearly dependent. Then exist α1 , . . . , αk not all zero s.t.
w1
 . 
α1 · · · αk  .. = 0. So


wk
    
a11 · · · ak1 u1 u1
 . .. .. = β · · ·
αk  .. βk  ...,
  
0 = α1 · · · . . 1
a1k · · · akk uk uk

αk At = β1 · · ·
   
where α1 · · · βk =6 0. Thus u1 , . . . , uk are linearly dependent.

We now assume that w1 , . . . , wk are LI. To show that u1 , . . . , uk are LI.

Suppose u1 , . . . , uk are LD. There there exists c1 , . . . , ck ∈ F not all zero such that
c1 u1 + c2 u2 + · · · + ck uk = 0.

Multiplying both side by At . At (c1 u1 + c2 u2 + · · · + ck uk ) = 0

spanda@maths.iitkgp.ac.in 4

  
a11 · · · ak1 u1
αk  ... .. .. = 0.
 
α1 · · · . .
a1k · · · akk uk
 k 
P
 ai1 ui 
i=1
 .. = 0.
 
α1 · · · αk 

 .

P k 
aik ui
i=1

k
P k
P k
P
This implies that ai1 ui , ai2 ui , . . . , aik ui are LI. A contradiction. Hence u1 . . . , uk are
i=1 i=1 i=1
LI.

[0.0.9] Exercise Let V be a vector space over F. Let A and B be two non-empty subsets of V. Prove
or disprove: ls(A) ∩ ls(B) 6= {0} =⇒ A ∩ B 6= φ.

Sol. Disprove. Consider R2 and A = {(1, 0), (0, 1)} and B = {(1, 2), (2, 1)}. Notice that
ls(A) = R2 and ls(B) = R2 . Hence ls(A) ∩ ls(B) 6= {0}. But A ∩ B = φ

[0.0.10] Exercise Show that a vector space V over F has a unique basis if and only if either dim (V) = 0
or dim (V) = 1 and |F| = 2.

Sol. First assume that V(F) has a unique basis.

There are two cases. Either V is trivial or non-trivial.

Case I. If V is trivial, then we are done.

Case II. V is non-trivial. First we show that dim V = 1. Suppose that dim V ≥ 2. There
are two cases.

Case II(a). V is finite dimensional. Let {x1 , . . . , xk } be a basis of V where k ≥ 2.

Consider B = {x1 , x2 , . . . , xi , . . . , xj−1 , xj + xi , xj+1 , . . . , xk }. To show that B is also a basis
of V. Using Problem 0.0.8 in tutorial sheet, you can easily show that B is LI. Since B is LI and
|B| = dim (V). Hence B is basis. A contradiction that V has unique basis. Hence dim (V) = 1.

Case II(b). V is infinite dimensional. Let S = {xα : α ∈ I} be a basis of V where I is an

index set. Take xγ , xβ ∈ S. Then xγ + xβ ∈ V as V is a vector space. But xγ + xβ ∈
/ S as S is
LI, otherwise there is a vector xγ + xβ which is a linear combination of xγ and xβ .
spanda@maths.iitkgp.ac.in 5

 
Consider B = {xα : α ∈ I} − {xγ } ∪ {xγ + xβ }.

To show that B is a basis of V. Let A be a finite subset of B. There are two cases.

Case I. A = {xα1 , xα2 , . . . , xαk } that means xγ + xβ is not in A. Therefore A ⊆ S. Since S

is LI, then A is LI.

Case II. A = {xα1 , xα2 , . . . , xγ + xβ , . . . , xαk }. Then applying the same techniques as of finite
case, we have A is LI. Therefore each finite subset of B is LI. Hence B is LI.

To show ls(B) = V. It is enough to show ls(B) = ls(S). B − {xγ + xβ } = S − {xγ } and you
know that xγ + xβ is in B. Using this you can easily show that ls(B) = ls(S).

Hence B is a basis of V. Therefore V has two basis a contradiction. Hence dim (V ) = 1.

We now show that |F| = 2. Suppose that |F| > 2. Then |F | has at least one element which
is other than additive identity and multiplicative identity. Let {x} be a basis of V and let α ∈ F
such that α is neither 0 nor 1. Then {x} and {αx} both are basis of V. A contradiction that V
has unique basis. Hence |F| = 2.

Assume that either dim (V) = 0 or dim (V) = 1 and |F| = 2. If dim (V) = 0, then V = {0}.
Hence its has unique basis.
If dim (V) = 1 and |F| = 2. Suppose that B has two bases they are {x} and {y}. Since y ∈ V
and {x} is basis, then y = αx where α ∈ F. Since F has two elements, then α = 1 (multiplicative
identity). Therefore y = 1(x) = x. Hence V has unique basis.

[0.0.11] Exercise Let V be an n dimensional vector space over F and let F has exactly p elements.
Then show that |V| = pn .

Sol. Let B = {u1 , u2 , . . . , un } be a basis of V. Any element in V can be written as a unique

linear combination of u1 , u2 , . . . , un . That is c1 u1 + c2 u2 + c3 u3 + · · · + cn un where ci ∈ F. For
each ci we have p choices and the choice of ci does not depend on the choice of cj for i 6= j. Hence
|V| = pn .

[0.0.12] Exercise Check whether vector space R (set of real numbers) over the field Q (set rational
number) is infinite dimensional or finite dimensional.

Sol. If we are able to show that for each n ∈ N there exists a LI subset of R(Q) containing n+1 el-
ements. Then we are done. Let α be a transcendental number. To show that {1, α, α2 , α3 , . . . , αn }
is LI.Take a0 + a1 α + a2 α2 + · · · + an αn = 0. We never have non-trivial solution of this equation
spanda@maths.iitkgp.ac.in 6

as α is transcendental. Hence {1, α, α2 , α3 , . . . , αn } is LI. For each n we have a LI set of n + 1

vectors. Therefore R(Q) is infinite dimensional.

     
4 a 4
[0.0.13] Exercise Let S = {5,  2, 3}. Find the values of a for which ls(S) 6= R3 .
6 4 2

Sol. We find the values of a for which ls(S) = R3 . Since S contains exactly three and
dim (S) = R3 , then S is a basis of R3 . This implies S is LI.

 
4 a 4
Consider A = 5 2 3. Since S is LI, then A is invertible. Hence det (A) 6= 0 and
6 4 2
det (A) = 8a. Therefore we have 8a 6= 0. This implies a 6= 0.
We have seen that if a ∈ R − {0}, then ls(S) = R3 .
Then So ls(S) 6= R3 iff a = 0.

[0.0.14] Exercise Give 2 bases for the trace 0 real symmetric matrices of size 3 × 3. Extend these bases
to bases of the real symmetric matrices of size 3 × 3. Extend these bases to bases of the real matrices
of size 3 × 3.

[0.0.15] Exercise Consider W = {v ∈ R6 |v1 + v2 + v3 = 0, v2 + v3 + v4 = 0, v4 + v5 + v6 = 0}. Supply a

basis for W and extend it to a basis of R6 .

   
1 1 1 0 0 0 0 1 0 0 0 1 1 0
Sol. For the system: 0 1  1 1 0 0 0. Get the rref: 0 1 1 0 −1 −1 0.
0 0 0 1 1 1 0 0 0 0 1 1 1 0
Free variables: v3 , v5 , v6 . Obtain linearly independent solutions by putting a free variable 1
and other free variables 0:
0 −1 −1
     
−1  1  1
     
 1  0  0
{
 0, −1, −1}.
    
     
 0  1  0
0 0 1
This is a basis for W. Note that e1 cannot be a linear combination of these three vectors.!!
To extend add e1 , e2 , e4 . These three with the first basis vector will give e3 . Similarly..... So
these 6 will span R6 . They are linearly independent.

[0.0.16] Exercise For what values α are the vectors (0, 1, α), (α, 1, 0) and (1, α, 1) in R3 linearly inde-
pendent?
spanda@maths.iitkgp.ac.in 7

[0.0.17] Exercise If S and T are two subspaces of a vector spaces having a common complement set
W , does it follow that S = T ?

Sol. Not necessarily. Consider R2 (R). Take S = {(x, 0) : x ∈ R} and T = {(0, y) : y ∈ R}.
Consider W = ls({(1, 1)}). It is easy to check W is complement of S and T . But S 6= T .

[0.0.18] Exercise In the vector space R4 , find two different complements of the subspace S =
{(x1 , x2 , x3 , x4 ) : x3 − x4 = 0}

Sol. We first find a basis of W. Let (x1 , x2 , x3 , x4 ) ∈ W. Then x3 − x4 = 0. Therefore

(x1 , x2 , x3 , x4 ) = (x1 , x2 , x3 , x3 ) = x1 (1, 0, 0, 0) + x2 (0, 1, 0, 0) + x3 (0, 0, 1, 1) This implies that
ls({(1, 0, 0, 0), (0, 1, 0, 0), (0, 0, 1, 1)}). One can easily prove that {(1, 0, 0, 0), (0, 1, 0, 0), (0, 0, 1, 1)}
is LI. Hence {(1, 0, 0, 0), (0, 1, 0, 0), (0, 0, 1, 1)} is basis of W.
We now extend {(1, 0, 0, 0), (0, 1, 0, 0), (0, 0, 1, 1)} to a basis of R4 . Take (0, 0, 0, 1) ∈ R3 − W.
Using basis extension theorem {(1, 0, 0, 0), (0, 1, 0, 0), (0, 0, 1, 1), (0, 0, 0, 1)} is LI. It has exactly
four elements. Then {(1, 0, 0, 0), (0, 1, 0, 0), (0, 0, 1, 1), (0, 0, 0, 1)} is a basis of R3 . Therefore
S1 = ls({(0, 0, 0, 1)}) is a complement of V.

Take (1, 1, 1, 0) ∈ R3 − W.
Using basis extension theorem {(1, 0, 0, 0), (0, 1, 0, 0), (0, 0, 1, 1), (1, 1, 1, 0)} is LI. It has exactly
four elements. Then {(1, 0, 0, 0), (0, 1, 0, 0), (0, 0, 1, 1), (1, 1, 1, 0)} is a basis of R3 . Therefore S2 =
ls({(1, 1, 1, 0)}) is a complement of V.

[0.0.19] ExerciseShow that a non-trivial subspace S of a finite dimensional vector space V has two
virtually disjoint complements iff dim (S) ≥ dim2 (V) .

Sol. We first assume that dim (S) ≥ dim2 (V) . To show S has two virtually disjoint complements.
n
Let dim (V) = n and dim (S) = k. Then k ≥ 2. Let {x1 , . . . , xk } be a basis of S. We extend
it to basis of V that is {x1 , . . . , xk , xk+1 , . . . , xn }. Take S1 = ls(xk+1 , . . . , xn }). Then S1 is a
complement of S.
n
Consider {x1 , . . . , xk , xk+1 + x1 , xk+2 + x2 , . . . , xn + xn−k },. This is possible as k ≥ 2. To
show this set is basis of V. Take S2 = ls({xk+1 + x1 , xk+2 + x2 , . . . , xn + xn−k }). This is also a
complement of S. To show S1 ∩ S2 = {0}.
Let x ∈ S1 ∩ S2 = {0}. Then x ∈ S1 and x ∈ S2 .

x = c1 xk+1 + c2 xk+2 + · · · + cn−k xn and

spanda@maths.iitkgp.ac.in 8

x = b1 (xk+1 + x1 ) + b2 (xk+2 + x2 ) + · · · + bn−k (xn + xn−k )

b1 x1 + · · · bn−k xn−k + (b1 − c1 )xk+1 + · · · + (bn−k − cn−k )xn = 0

{x1 , . . . , xn−k , xk+1 , . . . , xn } is LI

Then bi = ci = 0 for i = 1, . . . , n − k. Hence x = 0.

We now assume that S has two virtually disjoint complements.

To show dim (S) ≥ dim2 (V) . Suppose that dim (S) < dim2 (V) . Let dim (V) = n and dim (S) =
k. Let S1 and S2 be two virtually disjoint complements. Then dim (S1 ) = n − k = dim (S2 ).

dim (S1 + S2 ) = dim (S1 ) + dim (S2 ) = 2(n − k) > 2(n − n2 ) = n, a contradiction. Hence
dim (S) ≥ dim (V) .
2

n
xi = 0} in Rn .
P
[0.0.20] Exercise Find a complement of the subspace {(x1 , x2 , . . . , xn } :
i=1

Sol. Same as Exercise 0.0.18