MATH F112 (MATHEMATICS-II)

Gaurav Dwivedi
Department of Mathematics
BITS Pilani, Pilani Campus.
 Module 2
Real Vector Spaces and Subspaces
Module 2

1 Vector Spaces
2 Subspaces
3 Linear Span
4 Linear Independence

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 3 / 80

 Introduction to Vector
Spaces

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 4 / 80

What is a Vector space?

A set of pointed arrows?

An ordered list of numbers?

A set of functions?

Or something more abstract?

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 5 / 80

R2
Consider the set of all ordered pair of real numbers is
denoted by R2 i.e. R2 = {(u, v) | u, v 2 R}.
Let us call elements of R2 ‘vectors’ and elements of R
‘scalars’.
Define the operations called vector addition ( ) and
scalar multiplication ( ) as follows:
For every a 2 R and u = (x1 , x2 ), v = (y1 , y2 ) 2 R2
Vector Addition u v = (x1 + y1 , x2 + y2 )
Scalar Multiplication a u=a (x1 , x2 ) = (ax1 , ax2 )

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 6 / 80

Properties of and

Let u = (x1 , x2 ), v = (y1 , y2 ) and w = (z1 , z2 ) 2 R2 and

a, b 2 R.
P1 u v 2 R2 (Closure Property of addition)
P2 u v = v u (Commutative Property)
P3 (u v) w = u (v w) (Associative Property)
P4 there exists a vector x 2 R2 such that u x = u
(Existence of additive identity (zero vector))
P5 there exists a vector x 2 R2 such that u x = 0
(Existence of additive inverse ( u))

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 7 / 80

Properties of and

P6 a u 2 R2 (Closure Property of scalar

multiplication)
P7 a (u v) = (a u) (a v) (Distributivity over
vector addition)
P8 (a + b) u = (a u) (b u) (Distributivity
over scalar addition)
P9 (ab) u = a (b u)
P10 1 u = u

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 8 / 80

Vector Space: Definition

Let V be an arbitrary nonempty set of objects, together

with two operations
(vector) addition (denoted as ) and
scalar multiplication (denoted as ),
is said to be a (real) vector space (denoted as V (R)) if
for every u, v, w in V (known as vectors) and for every
a, b 2 R (known as scalars), the following properties
hold:

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 9 / 80

 P1: u v 2 V (Closure property of addition)
P2: u v = v u (Commutativity)
P3: (u v) w = u (v w) (Associativity)
P4: There exists an element 0 2 V , called a zero
vector, such that u 0 = u (Existence of additive
identity)
P5: For each u 2 V , there is an element u 2 V such
that u ( u) = 0 (Existence of additive inverse)
P6: a u 2 V (Closed under scalar multiplication)
P7: a (u v) = (a u) (a v) (Distributivity)
P8: (a + b) u = (a u) (b u) (Distributivity)
P9: (ab) u = a (b u)
P10: 1 u = u.
Trivial vector space

Q:⇧ The set V = {0} is a vector space with respect to

1 vector addition 0 0 = 0
2 scalar multiplication a 0 = 0 for all a 2 R

The vector space V = {0} is called the zero (trivial)

vector space.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 11 / 80

Some Results
Let V be a vector space with vector addition and
scalar multiplication .
Theorem
Prove that the identity element for vector addition in V
is unique.

Theorem
For every u 2 V and k 2 R,
1 k 0V = 0V
2 0 u = 0V
3 ( 1) u = u
4 If k u = 0V , then k = 0 or u = 0V .
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 12 / 80
Example
The set R+ of a positive real numbers form a vector
space with respect to the following operations:
u v = u · v (vector addition)
a u = ua (scalar multiplication)
for all a 2 R and u, v 2 R+ .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 13 / 80

Solution
(1) Since u, v 2 R+ , then u v = u · v 2 R+ . Thus,
R+ is closed with respect to .
(2) Let u, v 2 R+ . Then
u v =u·v
= v · u multilication is commutative in R+
=v u
(3) For u, v, w 2 R+ , we have
u (v w) = u (v · w)
= u · (v · w)
= (u · v) · w multiplication is associative
= (u v) · w = (u v) w
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 14 / 80
Continued
(4) Since 1 2 R+ and u 1 = u · 1 = u. Thus,
0R+ = 1 2 R+ is the additive identity in R+ .
(5) The additive inverse of v 2 R+ is v1 2 R+ because
v v1 = v · v1 = 1 (identity).
(6) For u 2 R+ , a u = ua is a positive real number.
Thus, a u 2 R+ .
(7) Here, we have
a (u v) = a (u · v)
= (u · v)a
= ua · v a
= (a u) · (a v)
= (a u) (a v).
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 15 / 80
Continued
(8) Here, we have
(a + b) u = ua+b
= ua · ub
= (a u) · (b u)
= (a u) (b u).
(9). Here, we have
(ab) u = uab
a
= ub
= (b u)a
= a (b u).
(10) 1 u = u1 = u.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 16 / 80
Example

We have already verified that

R2 = {(x1 , x2 ) | x1 , x2 2 R} is a vector space with
respect to the following vector addition and scalar
multiplication :
(x1 , x2 ) (y1 , y2 ) = (x1 + y1 , x2 + y2 )
a (x1 , x2 ) = (ax1 , ax2 )
for all a 2 R and (x1 , x2 ), (y1 , y2 ) 2 R2 .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 17 / 80

Example

The set R of real numbers is a vector space with respect

to the following operations:
u v = u + v (vector addition)
a u = au (scalar multiplication)
for all a, u, v 2 R.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 18 / 80

Example
Consider the set

Rn = {(x1 , x2 , . . . , xn ) : xi 2 R}.

For any u, v 2 Rn and a 2 R, define

u v = (x1 , x2 , . . . , xn ) (y1 , y2 , . . . , yn )
= (x1 + y1 , x2 + y2 , . . . , xn + yn )
a u = (ax1 , ax2 , . . . , axn ).

Then Rn is a vector space with respect to and .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 19 / 80
Exercises
1 Does R2 form a vector space under the above
defined vector addition and the following scalar
multiplication

a (x1 , x2 ) = (0, ax2 )

for all a 2 R and (x1 , x2 ) 2 R2 . If not, then which

properties are not satisfied?
2 Does the set R+ of positive real numbers form a
vector space under the usual addition and scalar
multiplication? If not, then which properties are not
satisfied?
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 20 / 80
Exercise
Consider the set R2 = {(x1 , x2 ) | x1 , x2 2 R} with
respect to the following vector addition and scalar
multiplication :
(x1 , x2 ) (y1 , y2 ) = (x1 + y1 1, x2 + y2 1)
a (x1 , x2 ) = (ax1 , ax2 )
for all a 2 R and (x1 , x2 ), (y1 , y2 ) 2 R2 .
i. Compute zero vector and additive inverse.
ii. Prove or disprove: R2 is a vector space for defined
vector addition and scalar multiplication.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 21 / 80

Exercise
Consider the set R2 = {(x1 , x2 ) | x1 , x2 2 R} with
respect to the following vector addition and scalar
multiplication :
(x1 , x2 ) (y1 , y2 ) = (x1 + y1 2, x2 + y2 + 3)
a (x1 , x2 ) = (ax1 2a + 2, ax2 + 3a 3)
for all a 2 R and (x1 , x2 ), (y1 , y2 ) 2 R2 .
i. Compute zero vector and additive inverse.
ii. Prove or disprove: R2 is a vector space for defined
vector addition and scalar multiplication.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 22 / 80

Exercise

Prove or disprove: The set R2 = {(x1 , x2 ) | x1 , x2 2 R}

is a vector space with respect to the following vector
addition and scalar multiplication :
(x1 , x2 ) (y1 , y2 ) = (x1 + y1 2, x2 + y2 + 3)
a (x1 , x2 ) = (ax1 2a + 2, ax2 + 3a 3)
for all a 2 R and (x1 , x2 ), (y1 , y2 ) 2 R2 .
What is additive identity, (zero vector), and the additive
inverse of each vector (x1 , x2 ) in V ?

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 23 / 80

Exercise

Show that V = R2 be a vector space with respect to the

following vector addition and scalar multiplication :

(x1 , x2 ) (y1 , y2 ) = (x1 + y1 + 4, x2 + y2 5)

a (x1 , x2 ) = (ax1 + 4a 4, ax2 5a + 5)

for all a 2 R and (x1 , x2 ), (y1 , y2 ) 2 R2 .
What is the additive identity (zero vector), and the
additive inverse of each vector (x1 , x2 ) in V ?

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 24 / 80

Exercise

Let V = {(x1 , x2 ) : x1 , x2 2 R}. Define vector addition

and scalar multiplication as follows:

(x1 , x2 ) (y1 , y2 ) = (x1 + y1 , x2 + y2 )

(
(0, 0) if a = 0
a (x1 , x2 ) =
ax1 , xa2 if a 6= 0
for all a 2 R and (x1 , x2 ), (y1 , y2 ) 2 V . Prove or
disprove that X is a vector space.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 25 / 80

Vector space of Matrices

The set
Mmn = {[aij ]m⇥n | aij 2 R}
of all m ⇥ n matrices with real entries is a vector space
with respect to the following operations:
[aij ]m⇥n [bij ]m⇥n = [aij + bij ]m⇥n (vector addition)

a [aij ]m⇥n = [aaij ]m⇥n (scalar multiplication)

for all a 2 R and [aij ]m⇥n , [bij ]m⇥n 2 Mmn .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 26 / 80

Vector space of functions
Let F ( 1, 1) = {f | f : ( 1, 1) ! R} be the set of
real-valued functions defined on ( 1, 1). For any
f, g 2 F ( 1, 1) and a 2 R, define

f g = f + g (vector addition),

where (f + g)(x) = f (x) + g(x) for all x 2 ( 1, 1)

and a f = af (scalar multiplication),

where (af )(x) = af (x) for all x 2 ( 1, 1).

Then F ( 1, 1) is a vector space with respect to above
defined operations.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 27 / 80
Questions

Q:⇧ Check if the following sets form a vector space?

i. The set of all 2 ⇥ 2 invertible matrices with the
standard matrix addition and scalar multiplication.
ii. The set of real numbers with usual scalar
multiplication but with addition given as
x y = 2(x + y).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 28 / 80

Definition: Subspaces
Let V be a vector space with vector addition and
scalar multiplication .
A nonempty subset W of V is said to be a subspace of
V if W is itself a vector space with respect to the same
operations and .

Note that every vector space V has at least two

subspaces: {0} and V itself. The subspace {0} is known
as zero (trivial) subspace.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 29 / 80

Some Results

Theorem. A nonempty subset W of a vector space V

is a subspace of V if and only if the following conditions
hold:
1 If u and v are vectors in W, then u + v is in W .
2 If k is a scalar and u is a vector in W , then ku is in
W.
Remark. If W is a subspace of a vector space V , then
0 2 W.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 30 / 80

Questions

1. Is
W = (x, y) 2 R2 | y = 0
a subspace of R2 ?
2. Is
W = (x, y) 2 R2 | x 6= y
a subspace of R2 ?

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 31 / 80

Examples
Examine whether the following sets are subspaces of the
vector space R3 .
1 W1 = (x, y, z) 2 R3 | x + y + z = 0 ? Yes
2 W2 = (x, y, z) 2 R3 | x > 0 ? No
3 W3 = (x, y, z) 2 R3 | x = y 2 ? No
4 W4 = (x, y, z) 2 R3 | x + y + z = 2 ? No
5 W5 = (x, y, z) 2 R3 | x2 + y 2 + z 2 = 1 ? No

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 32 / 80

Examples
Examine whether the following sets are subspaces of the
vector space M22 .
1 W1 = {A 2 M22 | A is singular}? No
2 W2 = {A 2 M22 | A is nonsingular}? No
3 W3 = {A 2 M22 | A is in RREF}? No
4 W4 = {A 2 M22 | A is symmetric}? Yes
5 W5 = A 2 M22 | A2 = A ? No

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 33 / 80

Examples
Examine whether the following sets are subspaces of the
vector space F ( 1, 1).
1 W1 = {f 2 F ( 1, 1) | f is continuous}. This
subspace is denoted by C( 1, 1)? Yes
2 W2 = {f 2 F ( 1, 1) | f is discontinuous}? No
3 W3 = {f 2 F ( 1, 1) | f (1) = 0}? Yes
4 W4 = {f 2 F ( 1, 1) | f (1) 6= 0}? No
1
5 W5 = f 2 F ( 1, 1) | f 2 = f (1) ? Yes

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 34 / 80

Example
Consider the vector space V = R2 with operations
(x1 , y1 ) (x2 , y2 ) = (x1 + x2 + 1, y1 + y2 2)
↵ (x, y) = (↵x + ↵ 1, ↵y 2↵ + 2).
Examine whether the following sets are subspaces of the
vector space V .
1 The set of (x, y) 2 R2 , lying on the straight line
y = 0. No
2 The set of (x, y) 2 R2 , lying on the straight line
x = 1. Yes
3 The set of (x, y) 2 R2 , lying on the straight line
x + y = 1. Yes
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 35 / 80
Results
Let W1 and W2 be two subspaces of vector space V .
Then
1 W1 \ W2 is a subspace of V .
2 W1 [ W2 is not necessarily a subspace of V .
3 W1 [ W2 is subspace of V if and only if either
W1 ✓ W2 or W2 ✓ W1 .
4 The sum, defined as

W1 + W2 = {w1 + w2 | w1 2 W1 , w2 2 W2 },

is a subspace of V .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 36 / 80

Subspace of polynomials

Recall that a polynomial is a function that can be

expressed in the form
p(x) = a0 + a1 x + · · · + an xn ,
where a0 , a1 , . . . , an are real constants. If an 6= 0, then
we say p(x) is a polynomial of degree n.
We denote by Pn the set of all polynomials of degree less
than or equal to n, i.e.,

Pn = {p 2 F ( 1, 1) : p(x) = a0 + a1 x + · · · + an xn }

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 37 / 80

Subspace of polynomials

Note that for any fixed natural number n, the set Pn of

all polynomials of degree less than or equal to n is a
subspace of F ( 1, 1).

We consider all constants (including 0) to be polynomials

of degree zero.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 38 / 80

Exercise

Prove that the set S = {p(x) 2 P2 |p(0) = p(2)}

forms a subspace of P2 with respect to usual
addition and scalar multiplication of polynomials.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 39 / 80

Exercise
V = R2 be a vector space with respect to the following
vector addition and scalar multiplication :
(x1 , x2 ) (y1 , y2 ) = (x1 + y1 2, x2 + y2 + 3)
a (x1 , x2 ) = (ax1 2a + 2, ax2 + 3a 3)
for all a 2 R and (x1 , x2 ), (y1 , y2 ) 2 R2 . Which of the
following are subspaces of V :
1 W = {(x, y) 2 R2 : x + y = 0}.
2 W = {(x, y) 2 R2 : x y = 5}.
3 W = {(x, y) 2 R2 : 3x + y = 3}.
4 W = {(x, y) 2 R2 : x = 2}.
Can you identify all possible subspaces of V ?
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 40 / 80
Exercise

Suppose W1 = {f 2 F ( 1, 1) | f is even} and

W2 = {f 2 F ( 1, 1) | f is odd} are subspaces of the
vector space V = F ( 1, 1). Prove that V = W1 + W2
and W1 \ W2 = {0}.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 41 / 80

Exercises

Show that the following sets form vector spaces with

standard operations:

1 W = {(x, y, z) 2 R3 : z = 0}

2 W = {A 2 Mnn : AB = BA for a fixed matrix B}

3 W = {p(x) = a0 + a1 x + a2 x2 + a3 x3 2 P3 :
a0 + a1 + a2 + a3 = 0}

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 42 / 80

 Linear Span

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 43 / 80

Linear Combination
Let V be a vector space and v1 , v2 , . . . , vr 2 V . Then a
vector w 2 V is said to be a linear combination of
v1 , v2 , . . . , vr if

w = k1 v1 + k2 v2 + · · · + kr vr ; ki 2 R, 1 6 i 6 r.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 44 / 80

Example

The vector (3, 4) is a linear combination of (1, 0) and

(0, 1) in R2 because

(3, 4) = 3(1, 0) + 4(0, 1)

Question: Can we write (3, 4) as linear combination of

(1, 1) and (1, 2)? Yes, as

(3, 4) = 2(1, 1) + (1, 2).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 45 / 80

Example
Recall that, R2 = {(x1 , x2 ) | x1 , x2 2 R} is a vector
space with operations
(x1 , x2 ) (y1 , y2 ) = (x1 + y1 2, x2 + y2 + 3)
a (x1 , x2 ) = (ax1 2a + 2, ax2 + 3a 3)
Q:⇧ Is (1, 10) a linear combination of (2, 3) and (1, 4)?

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 46 / 80

Span of a set
Let S be a nonempty subset of a vector space V . Then
the span of S is the set of all possible (finite) linear
combinations of the vectors in S and it is denoted by
span(S) i.e. if S = {v1 , v2 , . . . , vk }, then

span(S) = {a1 v1 + · · · + ak vk | ai 2 R, 1 6 i 6 k}

1 For a subset S = {(1, 0), (0, 1)} of R2 , we have

span(S) = R2 .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 47 / 80

Questions

1 Let V = R3 and S = {(1, 0, 0), (0, 1, 0)}.

1 Find span(S).
2 Do (3, 2, 0) and (2, 5, 1) belong to span(S)?
3 Is span(S) a subspace of R3 ?

2 Let V be a vector space and v1 , v2 2 V. Then show

that W = span{v1 , v2 } is a subspace of V .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 48 / 80

Theorem

Let S be a nonempty subset of a vector space V . Then

1 span(S) is a subspace of V .
2 span(S) is the smallest subspace of V containing S.
3 If S is a subspace, then span(S) = S.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 49 / 80

Question
Determine whether the vectors
v1 = (1, 2, 3), v2 = (2, 0, 0) and v3 = ( 2, 1, 0) span the
vector space R3 .
Solution. Let S = {v1 , v2 , v3 }. Clearly, by definition of
span(S), we have span(S) ✓ R3 . In order to check
span(S) = R3 , we have to check whether R3 is a subset
of span(S) or not.
Let (a, b, c) be an arbitrary element of R3 . We must
check whether (a, b, c) belongs to span(S) or not, i.e.,
whether there exists k1 , k2 , k3 2 R such that

(a, b, c) = k1 (1, 2, 3) + k2 (2, 0, 0) + k3 ( 2, 1, 0)

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 50 / 80

This is equivalent to checking whether the system of
equations 8
>
<k1 + 2k2 2k3 = a
2k1 + k3 =b
>
:3k
1 =c
is consistent for any a, b, c 2 R.
Note that the reduced row echelon form of the
augmented matrix
2 3 2 c
3
1 2 2 a 1 0 0 3
42 0 1 b 5 is 40 1 0 a + b 5c 5 .
2 6
3 0 0 c 0 0 1 b 2c3
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 51 / 80
This implies that the coefficient matrix is invertible and
hence the given system is consistent for any a, b, c 2 R.
Hence, span(S) = R3 .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 52 / 80

Example
Determine whether the vectors v1 = (3, 2, 4),
v2 = ( 3, 1, 0), v3 = (0, 1, 4), and v4 = (0, 2, 8) span
the vector space R3 .
Hint: By a similar argument used in the previous
exercise, one should check whether the system of
equations The system of equations
8
>
<3k1 3k2 =a
2k1 k2 + k3 + 2k4 = b
>
:4k + 4k + 8k
1 3 4 =c

is consistent for any a, b, c 2 R.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 53 / 80
Now, the reduced row echelon form of the augmented
matrix
2 3 2 3
3 3 0 0 a 1 0 1 2 b a3
42 1 1 2 b 5 is 40 1 1 2 b 2a 5.
3
4 0 4 8 c 0 0 0 0 4a 12b + 3c

Since the system is not consistent for all choices of

(a, b, c) 2 R3 , hence, span(S) 6= R3 .
Note that the vector (0, 0, 1) 2 R3 , but it is not in
span(S).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 9, 2024 54 / 78

Theorem
The solution set of a homogeneous linear system Ax = 0
in n unknowns is a subspace of Rn .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 55 / 80

 Introduction to Linear
Independence

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 56 / 80

Linear Independence
If S = {v1 , v2 , . . . , vr } is a nonempty set of vectors in a
vector space V , then the vector equation

k1 v1 + k2 v2 + · · · + kr vr = 0

has atleast one solution, namely

k1 = 0, k2 = 0, . . . , kr = 0.

If this is the only solution, then S is said to be a

linearly independent (LI) set. If there are solutions in
addition to the trivial solution, then S is said to be
linearly dependent (LD) set.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 57 / 80
Example

1 The subset S = {(1, 0), (0, 1)} of R2 is linearly

independent.

2 The subset S = {(1, 2), (5, 10)} of R2 is linearly

dependent.

3 The subset S = {(1, 0, 0), (0, 1, 0), (0, 0, 1)} of R3 is

linearly independent.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 58 / 80

 Any subset of V containing 0 2 V is linearly
dependent.
For v 6= 0 in V , the set {v} is linearly independent.
Let S = {v1 , v2 } be a set of nonzero vectors of V .
Then S is linearly dependent if and only if one
vector is a scalar multiple of the other.
Let S be a finite set of nonzero vectors having at
least two elements. Then S is linearly dependent if
and only if some vector in S can be expressed as a
linear combination of the other vectors in S.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 59 / 80

Example

Show that

S = {(3, 1, 1), ( 5, 2, 2), (2, 2, 1)}

is linearly independent subset of R3 .

Solution: Let a, b, c 2 R such that
a(3, 1, 1) + b( 5, 2, 2) + c(2, 2, 1) = 0
(3a, a, a) + ( 5b, 2b, 2b) + (2c, 2c, c) = (0, 0, 0)
(3a 5b + 2c, a 2b + 2c, a + 2b c) = (0, 0, 0)

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 60 / 80

To find a, b, c 2 R, we need to solve the following
homogenous system:

3a 5b + 2c = 0
a 2b + 2c = 0
a + 2b c = 0

To solve above homogenous system, write augmented

matrix 2 3
3 5 2 0
[A 0] = 4 1 2 2 0 5
1 2 1 0

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 61 / 80

reduced row echelon form of [A 0] is
2 3
1 0 0 0
40 1 0 05
0 0 1 0

Thus, we have a = 0, b = 0, c = 0. Hence, S is linearly

independent subset of R3 .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 62 / 80

Exercises
For a given vector space V and a given subset S of V ,
check the linear independence of S in the following:
1 V = P2 , S = {(x 2)2 , x2 4x, 12}.
2 V = P2 , S = {1 + x, x + x2 , 1 + x2 }.
3 V = Pn , S = {1, x, x2 , . . . , xn }.
4 V = F ( 1, 1), S = {sin2 x, cos2 x, cos 2x}.
5 V = F ( 1, 1), S = {0, cos4 x, sin2 5x}.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 63 / 80

Exercises
For a given vector space V and a given subset S of V ,
check the linear independence of S in the following:
⇢   
1 0 0 1 0 0 0 0
1 V = M22 , S = , , , .
0 0 0 0 1 0 0 1
⇢  
1 1 1 1 2 0
2 V = M22 , S = , , .
0 1 1 0 1 1

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 64 / 80

Example
Consider a vector space R2 with vector addition and
scalar multiplication are defined as

(x, y) (w, z) = (x + w, y + z 1)
and
a (x, y) = (ax, ay a + 1).
Let

S1 = {(0, 0), (1, 0)}, S2 = {(0, 1), (1, 0)}.

Are S1 and S2 linearly independent subsets of R2 .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 65 / 80

Exercise
Consider a vector space R2 with vector addition and
scalar multiplication are defined as

(x, y) (w, z) = (x + w + 1, y + z 2)
and
a (x, y) = (ax + a 1, ay 2a + 2).
Let
S = {(1, 1), (2, 2)}.
Is S a linearly independent subset of R2 .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 66 / 80

Example
Let u, v and w be three linearly independent vectors.
Prove or disprove: u + v, v + w and w + u are linearly
independent.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 67 / 80

 Results related to Linear
Independence

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 68 / 80

Theorem. If S is any subset of Rn containing r distinct
vectors, where r > n, then S is linearly dependent.

Example. Examine the linear independence of a subset

S = {[2, 5, 1], [1, 1, 1], [0, 2, 3], [2, 2, 6]} of R3 .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 69 / 80

 Linear independence of functions

Exercise. Examine the linear independence of the subset

S = {x, sin x} of F ( 1, 1).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 70 / 80

Wronskian
Definition. If f1 = f1 (x), f2 = f2 (x), . . . fn = fn (x) are
functions that are n 1 times di↵erentiable on the
interval ( 1, 1), then the determinant

f1 (x) f2 (x) . . . fn (x)

0
f1 (x) f20 (x) . . . fn0 (x)
W (x) = .. .. .. ..
. . . .
(n 1) (n 1) (n 1)
f1 (x) f2 (x) . . . fn (x)

is called the Wronskian of f1 , f2 , . . . , fn .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 71 / 80

Theorem. If the functions f1 , f2 , . . . , fn have n 1
continuous derivatives on the interval ( 1, 1), and if
the Wronskian of these function is not identically zero on
( 1, 1), then f1 , f2 , . . . , fn are linearly independent.
1 Use the Wronskian to show that the functions x, ex
and e x are linearly independent vectors in
F ( 1, 1).
2 Examine the linear independence of the subset
S = {x, sin x} of F ( 1, 1).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 72 / 80

Exercises
1 If {u, v, w} is a linearly independent set of a vector
space V , then prove or disprove that the set
{u + v, u v, u 2v + w} is a linearly independent
subset of V .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 6, 2024 73 / 80

MATH F112 (MATHEMATICS-II)
Gaurav Dwivedi
Department of Mathematics
BITS Pilani, Pilani Campus.
 Module 3
Basis, dimension of a vector space, Rank of a matrix
Module 3

1 Basis
2 Dimension,
3 Row Space, Column Space, and Null Space of a
matrix
4 Rank, Nullity of a matrix

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 3 / 29
 Basis of a Vector Space

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 4 / 29
Basis of a Vector Space
Definition: A finite subset S = {v1 , v2 , . . . , vn } of a
vector space V is said to be a basis of V if
1 S is LI, and
2 span(S) = V .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 5 / 29
Example
The subset S = {(1, 0), (0, 1)} = {e1 , e2 } is a basis of
R2 as S is LI and span(S) = R2 . The subset S is called
the standard basis of R2 .
Analogously, S = {e1 , e2 , . . . , en } is called the standard
basis of Rn , where ei is a vector of Rn such that its ith
component is 1 and remaining components are 0.

Question: Examine whether S = {(4, 1), (−7, −8)} is a

basis of R2 ?

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 6 / 29
1 A set containing a single vector is linearly
independent. (F)
2 No linearly independent set contains the zero vector.
(T)
3 Every linearly dependent set contains the zero
vector. (F)
4 The span of a single vector in R2 is a line. (F)
5 The span of a nonempty set S of vectors in V is the
smallest subspace of V that contains S. (T)
6 The span of any finite set of vectors in a vector
space is closed under addition and scalar
multiplication. (T)
7 Two subsets of a vector space V that span the
same subspace of V must be equal. (F)
8 Every linearly independent subset of a vector space
V is a basis for V . (F)
Standard bases

1 S = {1, x, x2 , . . . , xn } is a basis of Pn as S is LI
and span(S) = Pn . The set S is called the
standard basis of Pn .
2 The set
       
1 0 0 1 0 0 0 0
S= , , ,
0 0 0 0 1 0 0 1

is a basis of M22 . The set S is called the standard

basis of M22 .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 8 / 29
Questions
1 Find a basis for
W = {(a, b, c, d) ∈ R4 | d = a + 2b, c = 3a − b}.
2 Consider the subspace
W1 = {(a, b, c, d) | b − 2c + d = 0} and
W2 = {(a, b, c, d) | a = d, b = 2c} of R4 . Find a
basis for W1 ∩ W2 .
3 Find a basis for W = {A ∈ M33 | A = AT }.
4 Find a basis for W = {p(x) ∈ P4 |p(2) = p(3) = 0}.
5 Find a basis for
W = {p(x) ∈ P4 | p(3) = p(5) = p(8)}.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 9 / 29
Dimension of a Vector Space

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 10 / 29
Definition. A vector space that can be spanned by
finitely many vectors is said be finite dimensional.
Otherwise, it is called infinite dimensional.

Example: The vector spaces Rn , Pn and Mmn are finite

dimensional, whereas the vector space P∞ is infinite
dimensional.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 11 / 29
Theorem. Let V be a finite dimensional vector space,
and let S = {v1 , v2 , . . . , vn } be any basis
If a subset of V has more than n vectors, then it is
linearly dependent
If a subset of V has fewer than n vectors, then it
does not span V

Theorem. All bases for a finite dimensional vector space

have the same number of elements.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 12 / 29
Definition. The dimension of a finite dimensional vector
space V is the number of elements in a basis of V and it
is denoted by dim(V ).

The dimension of the zero vector space {0} is defined to

be zero.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 13 / 29
 Examples

dim(R2 ) = 2.
dim(R3 ) = 3.
dim(Rn ) = n.
dim(Pn ) = n + 1.
dim(Mmn ) = mn.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 14 / 29
Questions
1 Find a basis and the dimension of a subspace W of
R3 , where

W = {(x, y, z) ∈ R3 | x + 2z = 0}.
2 Find a basis and the dimension of a subspace W of
M33 , where

W = {A ∈ M33 | A = −AT }.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 15 / 29
Questions
1 Find a basis and dimension for the solution space of
the following homogeneous linear system
x + 2y − z = 0
2x − y + 2z = 0
3x + y + z = 0
4x + 3y = 0
 
1 0 1
2 Let A =  0 12 − 21 . Find values of x such that
n −2 0 x o
N = X ∈ R3 : AX = 0 is one dimensional
subspace of R3 .
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 16 / 29
Exercise

Let  
1 1 1
A =  2 2 3 .
x y z
Find the dimension of the subspace

W = {(x, y, z) ∈ R3 | A is singular}.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 17 / 29
Theorem. If S = {v1 , v2 , . . . , vn } is a basis for a vector
space V , then every vector v in V can be expressed in
the form v = c1 v1 + c2 v2 + · · · + cn vn in exactly one way.

Definition. If S = {v1 , v2 , . . . , vn } is a basis for a vector

space V , and

v = c1 v1 + c2 v2 + · · · + cn vn

then the scalars c1 , c2 , . . . , cn are called coordinates of v

relative to the basis S.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 18 / 29
The vector (c1 , c2 , . . . , cn ) ∈ Rn constructed from these
coordinates is called the coordinate vector of v
relative to S; it is denoted by

(v)S = (c1 , c2 , . . . , cn )
Remark: Sometime we shall write a coordinate vector as
column matrix and denoted by [v]S , i.e.,
 
c1
 c2 
[v]S = 
 ... 


cn
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 19 / 29
Example. Find the coordinate vector of the polynomial
p = 3 − x − 2x2 relative to the basis
S = {1 + x, 1 + x2 , x + x2 }.

Sol.

3 − x − 2x2 = c1 (1 + x) + c2 (1 + x2 ) + c3 (x + x2 )
= (c1 + c2 ) + (c1 + c3 )x + (c2 + c3 )x2

This leads to solve the system of equations

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 20 / 29
 c1 + c2 = 3
c1 + c3 = −1
c2 + c3 = −2

On solving, we get c1 = 2, c2 = 1, c3 = −3. Thus,

(p)S = (2, 1, −3).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 21 / 29
Exercise
Find the coordinate vector of A relative to the basis
S = {A
 1 , A2, A3 , A4 }, where  
3 −2 1 −1 0 1
A= , A1 = , A2 = ,
0 1  0 0 1 0
1 0 0 1
A3 = , A4 =
0 0 0 1

Answer. (A)S = (3, 0, 0, 1).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 22 / 29
Theorem. Let S = {v1 , v2 , . . . , vn } be a basis for a
vector space V . Let u, v ∈ V and let c be a scalar. Then

(u + v)S = (u)S + (v)S

(cu)S = c(u)S .

Theorem. Let S = {v1 , v2 , . . . , vn } be a basis for a

vector space V and let u1 , u2 , . . . , uk be vectors in V .
Then {u1 , u2 , . . . , uk } is linearly independent in V if and
only if {(u1 )S , (u2 )S , . . . , (uk )S } is linearly independent
in Rn .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 23 / 29
Exercise

Using the previous theorem, show that

v1 = −1 + x − 2x2 , v2 = 3 + 3x + 6x2 , and v3 = 9
form a linearly independent subset in P2 .
Hint: Since S = {1, x, x2 } is a standard basis of P2 , we
have

(v1 )S = (−1, 1, −2), (v2 )S = (3, 3, 6); , (v3 )S = (9, 0, 0).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 24 / 29
Results Related to Basis and
Dimension

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 25 / 29
Theorem. Let S = {v1 , v2 , . . . , vn } be a nonempty set
of vectors in a vector space V .
If S is a linearly independent and v ∈ V such that
v ̸∈ span(S), then S1 = {v, v1 , v2 , . . . , vn } is a
linearly independent set.
If v ∈ S such that it can be expressed as a linear
combination of other vectors in S, then

span(S) = span(S − {v}).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 26 / 29
Theorem. Let V be an n-dimensional vector space, and
let S be a set in V with exactly n vectors.
S is a basis of V if and only if S spans V .
S is a basis of V if and only if S is linearly
independent.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 27 / 29
Examples. For a given vector space V and a given
subset S of V , determine which of following S form a
basis of the respective vector space V :
1 V = R3 , S = {(3, 1, −1), (−5, −2, 2), (2, 2, −1)}.
2 V = R4 , S = {(7, 1, 2, 0), (8, 0, 1, −1)}.
3 V = P2 , S = {1 + x, x + x2 , 1 + x2 }.
4 V = P2 , S = {1 − x, x − x2 , 1 − x2 }.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 28 / 29
Theorem

Let S be a finite set in a finite-dimensional vector space

V.
1 If S is a linearly independent set but is not a basis
for V , then S can be enlarged to a basis for V by
inserting appropriate vectors in S.
2 If S spans V but is not a basis for V , then S can be
reduced to a basis for V by removing appropriate
vectors from S.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 29 / 29
Exercises
1 Extend the set

S = {(0, 0, 0, 3), (1, 1, 0, 0), (0, 1, −1, 0)}

to form a basis of R4 .
2 Extend the set S = {(1, 1, 0)} to form two different
bases of R3 .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 30 / 29
Exercises
1 Let S = {(4, 2, 1), (2, 6, −5), (1, −2, 3)} be a subset
of vector space R3 . Find dim(span(S)).
2 Find a basis for the subspace of P2 spanned by
v1 = −1 + x − 2x2 , v2 = 3 + 3x + 6x2 and v3 = 9.
3 Let U = {(a, b, c, d, e) ∈ R5 |a + c − 3d + e = 0}
and W = {(a, b, c, d, e) ∈ R5 |b − c − e = 0, a = d}
be the subspaces of R5 under usual operations. Find
a basis for U and for W and for U ∩ W .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 31 / 29
Theorem. Let W be a subspace of a finite dimensional
vector space V . Then
1 W is also finite dimensional and dimW ⩽ dimV .
2 dimW = dimV if and only if W = V .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 32 / 29
 Subspaces associated with
Matrices

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 33 / 29
Definitions
Let A be an m × n matrix.
The column space of A denoted by col(A), is the
subspace of Rm spanned by the column vectors of
A.
The row space of A, denoted by row(A), is the
subspace of Rn spanned by the row vectors of A.

row(A) = col(AT ).

The null space of A, denoted by null(A), is the

subspace of Rn consisting of solutions of the
homogeneous linear system Ax = 0.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 34 / 29
Row Equivalent Matrices
Matrices A and B are said to be row equivalent if A can
be obtained from B (or vice-versa) by a finite sequence
of elementary row operations.
Example. Matrices
   
3 2 7 3 2 7
A = −4 1 6 and B = −2 6 10
2 5 4 2 5 4

are row equivalent.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 35 / 29
 Every matrix is row equivalent to its row echelon
form.
If a matrix A is row equivalent to a matrix B, then
B is row equivalent to A.
Result. Let B be any matrix that is row equivalent to a
matrix A. Then row(B) = row(A). Is col(B) = col(A)?

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 36 / 29
Example. Find a basis for the row space, column space
and null space of
 
1 −3 2 4
0 1 −1 0
A=
0 0

1 3
0 0 0 1

Solution: Since given matrix is in row echelon form.

Hence, the set of row vectors
{(1, −3, 2, 4), (0, 1, −1, 0), (0, 0, 1, 3), (0, 0, 0, 1)}
is LI and forms a basis of row(A).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 37 / 29
The vectors
       
1 −3 2 4
0 1 −1 0
c1 = 
0 , c2 =  0  , c3 =  1  , c4 = 3
      

0 0 0 1

form a basis of col(A).

The null space has only the zero vector, hence it has no
basis and dimension is zero.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 38 / 29
Theorem. If a matrix R is in row echelon form, then
the row vector with the leading 1’s (the nonzero row
vectors) form a basis for the row space of R, and the
column vectors with the leading 1’s of the row vector
form a basis for the column space of R.

Theorem. If A and B are row equivalent matrices,

then:
A given set of column vectors of A forms a basis for
the column space of A if and only if the
corresponding column vectors of B forms a basis for
the column space of B.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 39 / 29
Finding Row(A) and Col(A)

Let A be the given matrix and R be an REF of A. Then

Row vectors with the leading 1’s form a basis for
the row space of A.
Identify the columns of R that contain the leading
1’s. The corresponding column vectors of R form a
basis for Col(A).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 40 / 29
Exercise

Find a basis for the row space, column space and null
space of  
1 4 5 6 9
3 −2 1 4 −1
A= −1

0 −1 −2 −1
2 3 5 7 8

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 41 / 29
Example
Find a basis for the row space, column space and null
space of  
1 4 5 6 9
3 −2 1 4 −1
A= −1

0 −1 −2 −1
2 3 5 7 8
Solution: Let B be the RREF of matrix A. Then
 
1 0 1 2 1
0 1 1 1 2
B= 0 0 0 0 0


0 0 0 0 0
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 42 / 29
Example
Let S = {v1 , v2 , v3 , v4 }, where

v1 = (1, 2, 3, −1, 0), v2 = (3, 6, 8, −2, 0)

v3 = (−1, −1, −3, 1, 1), v4 = (−2, −3, −5, 1, 1)

be a subset of R5 . Find a basis for span(S).
Step 1:  
1 2 3 −1 0
 3 6 8 −2 0
A= −1 −1 −3 1 1


−2 −3 −5 1 1

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 43 / 29
Step 2:
 
1 0 0 2 −2
0 1 0 0 1
RREF(A) =  
0 0 1 −1 0 
0 0 0 0 0

Step 3:

B = {(1, 0, 0, 2, −2), (0, 1, 0, 0, 1), (0, 0, 1, −1, 0)}

is a basis for span(S).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 44 / 29
Example
Let S = {v1 , v2 , v3 , v4 , v5 }, where
v1 = (1, 2, −2, 1), v2 = (−3, 0, −4, 3)
v3 = (2, 1, 1, −1), v4 = (−3, 3, −9, 6)
and v5 = (9, 3, 7, −6)
be a subset of R4 . Find a basis for span(S) comprising
of the elements of S.
Solution.
 
1 −3 2 −3 9
2 0 1 3 3
A= −2 −4 1 −9 7 


1 3 −1 6 −6
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 45 / 29
  
1 0 1/2 3/2 3/2
0 1 −1/2 3/2 −5/2
RREF(A) = 
0

0 0 0 0 
0 0 0 0 0
The set of vectors corresponding to pivot columns is

B = {v1 , v2 } = {(1, 2, −2, 1), (−3, 0, −4, 3)}

forms a basis for the subspace span(S).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 46 / 29
Basis for the Space Spanned by
{v1, . . . , vk }

Step 1. Form the matrix A whose columns are the

vectors in the set S = {v1 , v2 , . . . , vk }.
Step 2. Reduce the matrix A to reduced row echelon
form R.
Step 3. Denote the column vectors of R by
w1 , w2 , . . . , wk .
Step 4. Identify the columns of R that contain the
leading 1’s. The corresponding column
vectors of A form a basis for span(S).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 47 / 29
Example

Find a basis for the null space of

 
1 4 5 6 9
3 −2 1 4 −1
A= −1

0 −1 −2 −1
2 3 5 7 8

We know that

null(A) = {x : Ax = 0}

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 48 / 29
  
1 0 1 2 1
0 1 1 1 2
RREF(A) = 
0

0 0 0 0
0 0 0 0 0

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 49 / 29
null(A) = {(−r − 2s − t, −r − s − 2t, r, s, t) : r, s, t ∈ R}
= span(S), where

S = {(−1, −1, 1, 0, 0), (−2, −1, 0, 1, 0), (−1, −2, 0, 0, 1)}

Also, S is linearly independent (prove by yourself). Thus
S is a basis for null(A). Hence, dim(null(A)) = 3.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 50 / 29
Rank and Nullity of a Matrix
Theorem. The row space and column space of a matrix
have the same dimension.
Definition: The common dimension of row(A) and
col(A) of a matrix A is called the rank of A and is
denoted by rank(A);
dim(null(A)) is called the nullity of A and it is denoted
by nullity(A).
Result. For any matrix A, rank(A) = rank(AT ).
Dimension Theorem. If A is a matrix with n columns,
then
rank(A) + nullity(A) = n

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 51 / 29
Dimension Theorem for Matrices

If A is a matrix with n columns, then

rank(A) + nullity(A) = n

Example. Find the rank and nullity of the matrix

 
1 3 1 4
A =  2 4 2 0
−1 −3 0 5

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 52 / 29
  
1 3 1 4
REF(A) = 0 1 0 4
0 0 1 9

rank(A) = 3 =⇒ nullity(A) = 4 − 3 = 1.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 53 / 29
Theorem
Let A be an n × n matrix. The following statements are
equivalent:
A is invertible.
Ax = b has a unique solution for every b ∈ Rn .
The homogenous system Ax = 0 has only the trivial
solution.
The reduced row echelon form of A is In .
det(A) ̸= 0.
The column vectors of A are linearly independent.
The column vectors of A span Rn .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 54 / 29
Theorem (contd.)
The column vectors of A form a basis of Rn .
The row vectors of A are linearly independent.
The row vectors of A span Rn .
The row vectors of A form a basis of Rn .
A has rank n.
A has nullity 0.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 55 / 29
Conclusion

1 Vector Spaces
2 Subspaces
3 Span
4 Linear Independence
5 Basis and Dimension
6 Row space, Column Space, and Null Space
7 Rank and Nullity of a Matrix

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 15, 2024 56 / 29
MATH F112 (MATHEMATICS-II)
Gaurav Dwivedi
Department of Mathematics
BITS Pilani, Pilani Campus.
 Module 4
Linear Transformation
Recall
The row space and column space of a matrix have
the same dimension.
The common dimension of row(A) and col(A) of a
matrix A is called the rank of A and is denoted by
rank(A);
For any matrix A, rank(A) = rank(AT ).
dim(null(A)) is called the nullity of A and it is
denoted by nullity(A).
If A is a matrix with n columns, then

rank(A) + nullity(A) = n

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 3 / 59
True/False

The zero vector space has dimension zero.

There exists a set of 11 vectors that span R17 .
Every linearly independent set of five vectors in R5
is a basis for R5 .
Every basis of P4 contains at least one polynomial
of degree 3 or less.
There are at least two distinct three-dimensional
subspaces of P2 .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 4 / 59
True/False

The span of v1 , . . . , vn is the column space of the

matrix whose column vectors are v1 , . . . , vn .
If R is the reduced row echelon form of A, then
those column vectors of R that contain the leading
1’s form a basis for the column space of A.
The set of nonzero row vectors of a matrix A is a
basis for the row space of A.
If A and B are n × n matrices that have the same
row space, then A and B have the same column
space.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 5 / 59
True/False

If E is an m × m elementary matrix and A is an

m × n matrix, then the null space of EA is the
same as the null space of A.
The system Ax = b is inconsistent if and only if b is
not in the column space of A.
There is an invertible matrix A and a singular
matrix B such that the row spaces of A and B are
the same.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 6 / 59
Module 4

1 Introduction to Linear Transformation

2 Kernel and Range of a Linear Transformation
3 Rank and Nullity
4 One-to-one and Onto Linear Transformation
5 Isomorphic Vector Sapces
6 Inverse of Linear Transformation

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 7 / 59
 Introduction to Linear
Transformation

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 8 / 59
Linear Transformation
Let V and W be real vector spaces. A map T : V → W
is called a Linear map or Linear transformation (LT)
from V to W if the following two properties hold for all
vectors u and v in V and for every k ∈ R:
1 T (u + v) = T (u) + T (v) (Additivity property)
2 T (ku) = kT (u). (Homogeneity property)

In particular, when V = W , then T is called an

operator on V .
Let T be linear transformation. If v1 , v2 , . . . , vn are
vectors in V and k1 , k2 , . . . , kn are any scalars, then
T (k1 v1 + · · · + kn vn ) = k1 T (v1 ) + · · · + kn T (vn )
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 9 / 59
Zero and Identity Transformations
Let V and W be any two vector spaces.
T : V → W such that T (v) = 0 for every v ∈ V a
linear transformation. It is called zero
transformation.
T : V → V such that T (v) = v for every v ∈ V a
linear transformation. It is called identity operator
on V .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 10 / 59
Example

For A ∈ Mmn , consider the mapping L : Mmn → Mnm

given by L(A) = AT . Check whether L is a LT.
Solution: Let A, B ∈ Mmn and k ∈ R. Note that
L(A + B) = (A + B)T = AT + B T = L(A) + L(B)
L(kA) = (kA)T = kAT = kL(A).
Hence, L is a LT.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 11 / 59
Theorem. If T : V → W is a linear transformation,
then
1 T (0) = 0W .
2 T (u − v) = T (u) − T (v).
3 T (−v) = −T (v).

Example. Determine whether the transformation

T : P2 → P2 defined by

T (a + bx + cx2 ) = (a + 1) + (b + 1)x + (c + 1)x2

is a linear transformation.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 12 / 59
Exercises
Check which of the following maps are LT.
1 T : P2 → R3 given by T (a + bx + cx2 ) = (a, b, c).
2 T : Pn → Pn+1 given by T (p(x)) = xp(x).
3 T : R → R given by T (x) = sin x.
4 T : R → R given by T (x) = x2 .
5 T : Mnn → R given by T (A) = a11 a22 · · · ann .
6 T : Mnn → R given by T (A) = rank(A).

Sol. First two are LT.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 13 / 59
Exercise

Let V = R2 be a vector space with respect to the

following operations:

(x, y) ⊕ (w, z) = (x + w + 1, y + z − 2)
a (x, y) = (ax + a − 1, ay − 2a + 2).

Let T : V → V be a map such that

T (x, y) = 2 (x, y) for all (x, y) ∈ V.

Show that T is a linear transformation.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 14 / 59
Some Examples of LT
1 T : V → W such that T (v) = 0 for every v ∈ V.
2 T : V → V such that T (v) = v for every v ∈ V.
3 T : V → V such that T (v) = cv for any scaler c.
4 T : C 1 (−∞, ∞) → F (−∞, ∞) such that
T (f )(x) = f 0 (x).
5 T : C(−∞, ∞) → C 1 (−∞, ∞) such that
Z x
T (f )(x) = f (t) dt.
0
6 For a matrix A of order m × n, T : Rn → Rm given
by
T (x) = Ax.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 15 / 59
Example
Let T : R3 → R3 be a linear operator such that
T (1, 0, 0) = (−2, 1, 0), T (0, 1, 0) = (3, −2, 1), and
T (0, 0, 1) = (0, −1, 3).
1 Find T (−3, 2, 4).
2 Find T (x, y, z) for all (x, y, z) in R3 .
Solution.
(−3, 2, 4) = −3(1, 0, 0) + 2(0, 1, 0) + 4(0, 0, 1)
T (−3, 2, 4) = T (−3(1, 0, 0) + 2(0, 1, 0) + 4(0, 0, 1))
= −3T (1, 0, 0) + 2T (0, 1, 0) + 4T (0, 0, 1)
= −3(−2, 1, 0) + 2(3, −2, 1) + 4(0, −1, 3)
= (12, 11, 14)
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 16 / 59
Similarly,

T (x, y, z) = T (x(1, 0, 0) + y(0, 1, 0) + z(0, 0, 1))

T (x, y, z) = x(−2, 1, 0) + y(3, −2, 1) + z(0, −1, 3)
T (x, y, z) = (−2x + 3y, x − 2y − z, y + 3z)

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 17 / 59
Theorem. Let T : V → W be a linear transformation,
where V is finite dimensional. If S = {v1 , v2 , . . . , vn } is
a basis for V , then the image of any vector v in V can
be expressed as

T (v) = c1 T (v1 ) + c2 T (v2 ) + · · · + cn T (vn )

where c1 , c2 , . . . , cn are the coordinates of v relative to

the basis S.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 18 / 59
Exercise
Consider the basis S = {v1 , v2 , v3 } for R3 , where

v1 = (1, 1, 1), v2 = (1, 1, 0), v3 = (1, 0, 0)

and let T : R3 → R3 be the linear operator such that

T (v1 ) = (−1, 2, 4), T (v2 ) = (0, 3, 2), T (v3 ) = (1, 5, −1)

find a formula for T (x1 , x2 , x3 ), and use the formula to

find T (2, 4, −1).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 19 / 59
Hint: For (x1 , x2 , x3 ) ∈ R3 , note that if

(x1 , x2 , x3 ) = c1 v1 + c2 v2 + c3 v3

then, on solving above system of equations, we get

c1 = x3 , c2 = x2 − x3 and c3 = x1 − x2 . Thus,

(x1 , x2 , x3 ) = x3 v1 + (x2 − x3 )v2 + (x1 − x2 )v3

T (x1 , x2 , x3 ) = T (x3 v1 + (x2 − x3 )v2 + (x1 − x2 )v3 )

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 20 / 59
Then, using property of linear transformation and
substitute T (v1 ), T (v2 ), T (v3 ), we get

T (x1 , x2 , x3 ) = (x1 −x2 −x3 , 5x1 −2x2 −x3 , −x1 +3x2 +2x3 ).

T (2, 4, −1) = (−1, 3, 8).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 21 / 59
 Kernel and Range of a
Linear Transformation

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 22 / 59
Kernel and Range of a LT
Let T : V → W be a LT. The kernel of T , denoted by
ker(T ) is the subset of all vectors in V that map to 0W ,
i.e.,
ker(T ) = {v ∈ V | T (v) = 0W }.
Range: Let T : V → W be a LT. The range of T ,
denoted by R(T ), is the subset of all vectors in W that
are image of some vector in V , i.e.

R(T ) = {T (v) | v ∈ V }

Thus a vector w ∈ R(T ) implies there exists some vector

v ∈ V such that T (v) = w.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 23 / 59
Kernel and Range
of zero and identity transformation

Let T : V → W be the zero transformation.

ker(T ) = {v ∈ V | T (v) = 0W }
implies
ker(T ) = V.
Moreover,
R(T ) = {0}.
Let I : V → V be the identity operator.
R(I) = V.
Also,
ker(I) = {0}.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 24 / 59
Theorem. If T : V → W is a linear transformation,
then:
1 the kernel of T is a subspace of V .
2 the range of T is a subspace of W .

Rank and Nullity

Let T : V → W be a linear transformation. If the range
of T is finite dimensional, then its dimension is called
rank of T , it is denoted by rank(T );
If the kernel of T is finite dimensional, then its dimension
is called nullity of T , it is denoted by nullity(T ).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 25 / 59
Exercises
Determine ker(T ) and R(T ) of the following linear
transformations:
1 T : P2 → R3 given by T (a + bx + cx2 ) = (a, b, c).
2 T : P3 → P2 given by T (p(x)) = p0 (x).
3 L : M22 → M22 given by L(A) = AT .
R1
4 T : P1 → R given by T (a + bx) = 0 (a + bx)dx.
5 T : R3 → R2 given by T (x, y, z) = (0, y).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 26 / 59
Example
Let T : R3 → R2 be a LT given by

T (x, y, z) = (x − 2y, y + z).

Find ker(T ) and R(T ). Also, find a basis for ker(T ) and
R(T ). Hence, find rank(T ) and nullity(T ).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 27 / 59
Sol.
ker(T ) = (x, y, z) ∈ R3 | T (x, y, z) = 0R2


= (x, y, z) ∈ R3 | (x − 2y, y + z) = (0, 0)



= (x, y, z) ∈ R3 | x = 2y, z = −y


= {(2y, y, −y) | y ∈ R}

Here, ker(T ) = span{(2, 1, −1)}. Since the set

B = {(2, 1, −1)} is LI. Therefore, B = {(2, 1, −1)} is a
basis of ker(T ). Therefore, nullity(T ) = 1.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 28 / 59
 R(T ) = T (x, y, z) | (x, y, z) ∈ R3


= {(x − 2y, y + z) | x, y, z ∈ R}
= {x(1, 0) + y(−2, 1) + z(0, 1) | x, y, z ∈ R}
= span{(1, 0), (−2, 1), (0, 1)}
= span{(1, 0), (0, 1)}(Why?)

Since the set {(1, 0), (0, 1)} is LI. Thus,

{(1, 0), (0, 1)}

is a basis for R(T ) and hence, rank(T ) = 2.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 29 / 59
Example

Let V = R2 be a vector space with respect to the

following operations:

(x, y) ⊕ (w, z) = (x + w + 1, y + z − 2)
a (x, y) = (ax + a − 1, ay − 2a + 2).

Let T : V → V be a linear transformation such that

T (x, y) = 2 (x, y) for all (x, y) ∈ V. Find ker(T ) and
nullity(T ).
Answer: ker(T ) = {0V }, nullity(T ) = 0.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 30 / 59
Dimension Theorem for LT
If T : V → W is linear transformation from an
n-dimensional vector space V to a vector space W , then

rank(T ) + nullity(T ) = n.
Example: Let {e1 , e2 , e3 , e4 } be a standard basis of R4
and let T : R4 → R3 be a linear transformation such that
T (e1 ) = (1, 1, 1), T (e2 ) = (1, −1, 1),
T (e3 ) = (1, 0, 0) and T (e4 ) = (1, 0, 1). Find nullity and
rank of T .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 31 / 59
Exercise
Let A be the matrix
 
1 4 5 0 9
 3 −2 1 0 −1
A= 
−1 0 −1 0 −1
2 3 5 1 8

Consider its matrix transformation T . Then find

1 a basis for the range of T .
2 a basis for the kernel of T .
3 rank(T ) and nullity(T ).
4 rank(A) and nullity(A).
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 32 / 59
Exercise
Let A be the matrix
 
1 −1 3
A = 5 6 −4
7 4 2

What is the linear transformation T, associated with A?

1 a basis for the range of T .
2 a basis for the kernel of T .
3 rank(T ) and nullity(T ).
4 rank(A) and nullity(A).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 33 / 59
Questions

1 Let T : V → R3 be a linear transformation from a

vector space V to R3 . Geometrically, what are the
possibilities for the range of T ?
2 Let T : R3 → W be a linear transformation from R3
to a vector space W . Geometrically, what are the
possibilities for the kernel of T ?

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 34 / 59
One-to-one and Onto Linear
Transformation

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 35 / 59
One-to-One and Onto LT
Let T : V → W be a linear transformation. Then T is
said to be
one-to-one if T maps distinct vectors in V into
distinct vectors in W, i.e., T (u) = T (v) implies that
u = v.
onto if every vector in W is the image of at least
one vector in V, i.e., for each vector w in W , there
exists v ∈ V such that T (v) = w.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 36 / 59
Questions
1 Consider a T : M22 → M22 given by T (A) = AT .
Check if T is one-to-one and onto.
2 T : R2 → R 3 given by T (x, y) = (2x, x − y, 0).
Check if T is one-to-one and onto.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 37 / 59
Example: Consider a LT

T : P3 → P2 given by

T (p) = p0 .
Check if T is one-to-one and onto.
Solution: Consider p1 = x + 2 and p2 = x + 4. Since,
T (p1 ) = T (p2 ) = 1 implies T is not one-to-one.
Let q be an arbitrary element in P2 i.e.
q = a + bx +
cx2 . Note that a + bx + cx2 = p0 , where
p = ax + 2b x2 + 3c x3 so that T (p) = q. Hence, T is


onto.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 38 / 59
Theorem. Let T : V → W be linear transformation.
Then T is one-to-one if and only if ker(T ) = {0V }.

Theorem. Let V and W be finite-dimensional vector

spaces such that dim(V ) = dim(W ) = n. Then the
linear transformation T : V → W is one-to-one if and
only if it is onto.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 39 / 59
Example
Consider a LT T : M22 → M23 given by
   
a b a−b 0 c−d
T =
c d c+d a+b 0
Is T one-to-oneand onto?
a b
Solution: Let ∈ ker(L). Then
c d
     
a b a−b 0 c−d 0 0 0
T = = .
c d c+d a+b 0 0 0 0

We have a − b = c − d = c + d = a + b = 0 implies
a = b = c = d = 0.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 40 / 59
Hence, ker(L) contains only the zero matrix (the zero
vector of M22 ). Thus, T is one-to-one.
Note that
dim(R(T )) = dim(M22 ) − dim(ker(T ))
=4
6= dim(M23 ).

Thus, R(T ) ( M23 . Hence, T is not onto.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 41 / 59
Example
Let A be a fixed n × n matrix, and consider a LT
T : Mnn → Mnn given by T (B) = AB − BA.
Is T one-to-one and onto?
Solution: T (In ) = AIn − In A = 0n×n . Hence,
In ∈ ker(T ) and so, T is not one-to-one. Hence, T is
not onto.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 42 / 59
Exercise
Which of the following transformations are one-to-one?
onto?
1 T : R3 → R4 given by T (x, y, z) = (y, z, −y, 0).
2 T : Pn → Pn+1 given by T (p(x)) = xp(x).
3 T : P3 → R4 given by

T (a + bx + cx2 + dx3 ) = (a, b, c, d).

4 T : R∞ → R∞ given by

T (u1 , u2 , . . . , un , . . .) = (0, u1 , u2 , . . . , un , . . .)

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 43 / 59
Definition: If a linear transformation T : V → W is
both one-to-one and onto, then T is said to be an
isomorphism, and the vector spaces V and W are said
to be isomorphic.
Theorem. Every n-dimensional vector space is
isomorphic to Rn .
Example: Vector spaces Pn−1 and Rn are isomorphic.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 44 / 59
Theorem: Let V and W be two finite dimensional real
vector spaces. Then V is isomorphic to W if and only if
dimV = dimW .

Note that V is isomorphic to W means there exists a

one-one onto linear transformation from V to W .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 45 / 59
Exercise: Let V = {A ∈ M22 : Trace(A) = 0},
W = R2 . Determine whether V and W are isomorphic.
If they are, give an explicit isomorphism T : V → W .
Exercise: Determine whether V and W are isomorphic.

If they are, give an explicit isomorphism T : V → W .

1 V = Mmn , W = Rmn .
2 V = Pn , W = R n .
3 V = P2 , W = {p(x) ∈ P3 : p(0) = 0}.
4 V = {A ∈ M33 : A is diagonal matrix}, W = R3 .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 46 / 59
Recall that if T : V → W is a linear transformation, then
R(T ), the range of T consisting all images under T of
vectors in V . If T is one-to-one, then each vector
w ∈ R(T ) is the image of a unique vector v in V . This
uniqueness allow us to define a function, called the
inverse of T and denoted by T −1 that maps w back
into v, i.e.,
T −1 : R(T ) → V given by
T −1 (w) = v, where T (v) = w
Note that
T −1 (T (v)) = v, T (T −1 (w)) = w
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 47 / 59
Example: Let T : R3 → P2 be a LT given by

T (x, y, z) = x + (x + y − z)t + (x + y + z)t2 .

Find T −1 , if exists.
One How do we know if T −1 exists.
Solution: First show that T is one-to-one. Thus, we can
define T −1 : R(T ) → R3 . Note that R(T ) = P2 (verify!)
Let T −1 : P2 → R3 be defined by

T −1 (a + bt + ct2 ) = (x, y, z)

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 48 / 59
 ⇒ T (x, y, z) = a + bt + ct2
⇒ x + (x + y − z)t + (x + y + z)t2 = a + bt + ct2
⇒ x = a, x + y − z = b, x + y + z = c
b + c − 2a c−b
⇒ x = a, y = ,z = .
2 2
Hence,
 
b + c − 2a c − b
T −1 (a + bx + cx2 ) = a, , .
2 2

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 49 / 59
Exercise
Let T : P2 → R3 be the function defined by
 
p(−1)
T (p(x)) =  p(0) 
p(1)

Find T (x2 + 5x + 6).

Show that T is a linear transformation.
Show that T is one-to-one.
T −1 (0, 3, 0).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 50 / 59
Definition: If T1 : U → V and T2 : V → W are linear
transformations, then the composition of T2 with T1 ,
denoted by T2 ◦ T1 , is defined by

(T2 ◦ T1 )(u) = T2 (T1 (u)) for all u ∈ U.

Note that T2 ◦ T1 : U → W is also a linear

transformation.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 51 / 59
Exercise

Let T1 : P2 → P2 and T2 : P2 → P2 be linear operators

defined as T1 (ax2 + bx + c) = 2ax + b and
T2 (ax2 + bx + c) = 2ax2 + bx, respectively. Compute
T2 ◦ T1 and T1 ◦ T2 .
Answer:
(T2 ◦ T1 )(ax2 + bx + c) = 2ax.
(T1 ◦ T2 )(ax2 + bx + c) = 4ax + b.
Clearly, T2 ◦ T1 6= T1 ◦ T2 .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 52 / 59
Exercise

Let T1 : Pn → Pn and T2 : Pn → Pn be linear operators

defined as T1 (p(x)) = p(x + 2) and T2 (p(x)) = p(x − 2),
respectively. Compute T2 ◦ T1 and T1 ◦ T2 .
Answer:
(T2 ◦ T1 )(p(x)) = p(x).
(T1 ◦ T2 )(p(x)) = p(x).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 53 / 59
Theorem

If T1 : U → V and T2 : V → W are one-to-one linear

transformations, then
T2 ◦ T1 is one-to-one.
(T2 ◦ T1 )−1 = T1−1 ◦ T2−1 .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 54 / 59
Reflection

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 55 / 59
Projection

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 56 / 59
Rotation

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 57 / 59
Compression/Expansion

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 58 / 59
Shear

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 59 / 59
MATH F112 (MATHEMATICS-II)
Gaurav Dwivedi
Department of Mathematics
BITS Pilani, Pilani Campus.
 Module 5
Change of Basis and Matrix of a Linear
Transformation
 Module 5

Change of Basis, Transition Matrix

Matrices of Linear Transformations

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 3 / 30
 Change of Basis

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 4 / 30
Coordinate Vector

Recall that if S = {v1 , v2 , . . . , vn } is a basis for a vector

space V , and
v = c1 v1 + c2 v2 + · · · + cn vn
then the scalars c1 , c2 , . . . , cn are called coordinates of v
relative to the basis S.
The vector (c1 , c2 , . . . , cn ) ∈ Rn constructed from these
coordinates is called the coordinate vector of v
relative to S; it is denoted by
(v)S = (c1 , c2 , . . . , cn )
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 5 / 30
Remark: We shall also write a coordinate vector as
column matrix and in that case it will be denoted by
[v]S , i.e.,  
c1
 c2 
[v]S = 
 ... 


cn

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 6 / 30
Change of Basis Problem
Q: Let V be a vector of a finite dimensional vector
space and v ∈ V . If we change the basis for V from a
basis B to a basis B ′ , how are the coordinate vectors
[v]B and [v]B ′ are related?

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 7 / 30
We discuss a solution to the change of basis problem for
a two dimensional vector space V. Let B = {u1 , u2 } and
B ′ = {u′1 , u′2 } be the old and new bases, respectively.
Now let v ∈ V such that
 
k
[v]B ′ = 1
k2
i.e. v = k1 u′1 + k2 u′2 . In order to find [v]B , we must
express v as a linear combinations of vectors in B. This,
yields
v = k1 (au1 + bu2 ) + k2 (cu1 + du2 )
 
k1 a + k2 c
Thus, [v]B =
k1 b + k2 d
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 8 / 30
   
a c k1
[v]B =
b d k2
 
a c
[v]B = [v]B ′
b d
Thus, the old coordinate vector [v]B results when we
multiply the new coordinate vector [v]B ′ to the right of
the matrix
 
a c
P = = [ [u′1 ]B | [u′2 ]B ]
b d
The matrix P is called the transition matrix from B ′ to
B.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 9 / 30
In general, if we change the basis for a vector space V
from an old basis B = {u1 , u2 , . . . , un } to new basis
B ′ = {u′1 , u′2 , . . . , u′n }, then for each vector v ∈ V , the
old coordinate vector [v]B is related to the new
coordinate vector [v]B ′ by the equation

[v]B = P [v]B ′

where
P = [ [u′1 ]B | [u′2 ]B | · · · | [u′n ]B ]
P is called the transition matrix from B ′ to B. Also,
we shall denote it by PB ′ →B .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 10 / 30
Similarly, the transition matrix from B to B ′ is

PB→B ′ = [ [u1 ]B ′ | [u2 ]B ′ | · · · | [un ]B ′ ]

Thus, we have
[v]B = PB ′ →B [v]B ′
and
[v]B ′ = PB→B ′ [v]B

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 11 / 30
Example
Consider the bases B = {u1 , u2 } and B ′ = {u′1 , u′2 },
where
       
1 0 3 2
u1 = , u2 = , u′1 = , u′2 =
0 1 4 3

1 Find PB→B ′ and PB ′ →B .  

3
2 Compute [w]B , where w = .
−5
3 Compute [w]B ′ , using PB→B ′ .
4 Check your work by computing [w]B ′ directly.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 12 / 30
Solution:
1 To find the transition matrix PB→B ′ form B to B ′ ,
we have
PB→B ′ = [ [u1 ]B ′ | [u2 ]B ′ ]
To obtain [u1 ]B ′ and [u2 ]B ′ , we must solve

u1 = au′1 + bu′2
u2 = cu′1 + du′2

On solving, we get a = 3, b = −4, c = −2 and

d = 3.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 13 / 30
Therefore
   
3 −2
[u1 ]B ′ = and [u2 ]B ′ =
−4 3

Thus,  
3 −2
PB→B ′ =
−4 3

1 To find the transition matrix PB ′ →B form B to B ′ ,

we have
PB ′ →B = [ [u′1 ]B | [u′2 ]B ]

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 14 / 30
To obtain [u′1 ]B and [u′2 ]B , we must solve
u′1 = au1 + bu2
u′2 = cu1 + du2
Since u1 and u2 are standard basis vector, it is easy to
note a = 3, b = 4, c = 2 and d = 3.
Therefore
   
3 2
[u1 ]B ′ = and [u2 ]B ′ =
4 3
Thus,  
3 2
PB→B ′ =
4 3
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 15 / 30
 1 Since B is a standard basis, we have

w = 3u1 − 5u2

implies  
3
[w]B =
−5

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 16 / 30
 1 Since
[w]B ′ = PB→B ′ [w]B ,
we get
    
3 −2 3 19
[w]B ′ = =
−4 3 −5 −27
2 To obtain [w]B ′ directly, we have to solve

w = au′1 + bu′2

On substituting w, u′1 , u′2 , in the above we have

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 17 / 30
 3a + 2b = 3
4a + 3b = −5

On solving, we get a = 19 and b = −27 so that

 
19
[w]B ′ =
−27

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 18 / 30
Theorem. If P is the transition matrix from a basis B ′
to basis B for a finite dimensional vector space V , then
P is invertible and P −1 is the transition matrix from B
to B ′ .

Theorem. Let B = {u1 , u2 , ..., un } be any basis for Rn

and let S = {e1 , e2 , ..., en } be the standard basis for Rn .
If the vectors in these bases are written in column form,
then
PB→S = [u1 | u2 | · · · | un ]

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 19 / 30
A Procedure for Computing PB→B ′ between bases of
Rn

1 Write the matrix [B ′ |B].

2 Find RREF of the matrix written in Step 1.
3 The resulting matrix will be [I | PB→B ′ ].
4 Extract the matrix PB→B ′ from the right side of the
matrix in Step 3.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 20 / 30
Exercise
Consider the bases B = {u1 , u2 , u3 } and
B ′ = {u′1 , u′2 , u′3 } for R3 , where
     
−3 −3 1
u1 = 0 , u2 = 2 , u3 = 6 
    
−3 −1 −1
     
−6 −2 −2
′ ′ ′
u1 = −6 , u2 = −6 , u3 = −3
    
0 4 7

1 Find the transition matrix (PB→B ′ ) from B to B ′ .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 21 / 30
 1 Compute [w]B , where
 
−5
w = 8 .

−5
2 Compute [w]B ′ , using PB→B ′ .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 22 / 30
 1 Step 1: Construct
 
−6 −2 −2 −3 −3 1
 −6 −6 −3 0 2 6 
0 4 7 −3 −1 −1

Step 2: (Show that) the RREF of the above matrix

is  3 3 1 
1 0 0 4 4 12
 0 1 0 − 34 − 17 17
12 − 12
 

2 2
0 0 1 0 3 3

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 23 / 30
Step 3: Thus,
 3 3 1 
4 4 12
PB→B ′ =  − 43 17
− 12 17
− 12
 

2 2
0 3 3

1 In order to find [w]B , we have to solve

w = au1 + bu2 + cu3 , i.e.,

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 24 / 30
 −3a − 3b + c = −5
2b + 6c = 8
−3a − b − c = −5

The RREF of the augmented matrix of the corresponding

system is  
1 0 0 1
0 1 0 1
0 0 1 1

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 25 / 30
Thus, a = b = c = 1. Therefore, the coordinate vector
 
1
[w]B = 1 .

1

1 Since
[w]B ′ = PB→B ′ [w]B
 3 3 1    19 
4 4 12 1 12
[w]B ′ = − 43 − 17 17     43 
− 12  1 = − 12 

12
0 2 2 1 4
3 3 3
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 26 / 30
Exercise
Consider the bases B = {p1 , p2 } and B ′ = {q1 , q2 },
where

p1 = 1 + 2x, p2 = 3 − x, q1 = 2 − 2x, q2 = 4 + 3x

1 Compute the transition matrices PB ′ →B and PB→B ′ .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 27 / 30
Sol. Since P1 is isomorphic to R2 , any arbitrary element
a + bx of P1 can be written as (a, b) an element of R2 so
that

p1 = (1, 2), p2 = (3, −1), q1 = (2, −2), q2 = (4, 3)

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 28 / 30
To find PB ′ →B , find RREF of
 
1 3 2 4
2 −1 −2 3
(Show that) RREF of the above matrix is
" #
1 0 − 47 137
.
0 1 7 75
6

Therefore, " #
− 47 13
7
PB ′ →B = 6 5
.
7 7
Similarly, workout for PB→B ′ .
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 29 / 30
Exercise
Let  
1 0 0
P =0 2 4
0 1 3
be the transition matrix from basis B to the basis
B ′ = {v1 = (1, 1, 1), v2 = (1, 1, 0), v3 = (0, 1, 0), } for
R3 . Then find B?

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 30 / 30
 Matrices for General Linear
Transformations

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 31 / 30
Result: Let V and W be non-trivial vector spaces, with
dim(V ) = n and dim(W ) = m. B = {v1 , . . . , vn } and
B ′ = {w1 , . . . , wm } are ordered bases for V and W ,
respectively and T : V → W be a LT. Then there is a
unique m × n matrix A defined by
A = [ [T (v1 )]B ′ | [T (v2 )]B ′ | · · · |[T (vn )]B ′ ]
satisfies A[v]B = [T (v)]B ′ , for all v ∈ V .
We call A as the matrix for T relative to the bases
B and B ′ and it will be denoted by [T ]B ′ ,B .
Thus,
[T ]B ′ ,B = [ [T (v1 )]B ′ | [T (v2 )]B ′ | · · · | [T (vn )]B ′ ]
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 32 / 30
We call A as the matrix for T relative to the bases
B and B ′ and it will be denoted by [T ]B ′ ,B .
Thus,

[T ]B ′ ,B = [ [T (v1 )]B ′ | [T (v2 )]B ′ | · · · | [T (vn )]B ′ ]

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 33 / 30
In the case of linear operator, i.e. V = W , it is usual to
take B = B ′ when we construct matrix of T . In this case
the resulting matrix is called the matrix for T relative
to the basis B, i.e.,

[T ]B = [ [T (v1 )]B | [T (v2 )]B | · · · | [T (vn )]B ]

satisfies

[T ]B [v]B = [T (v)]B for all v ∈ V

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 34 / 30
Example
Consider the LT T : P1 → P2 , given by

T (p(x)) = xp(x)

with bases B = {v1 = x, v2 = 1} and

B ′ = {x2 , x − 1, x + 1} of P1 and P2 , respectively.
Compute [T ]B ′ ,B .

Sol.  
1 0
[T ]B ′ ,B = 0 1/2 .
0 1/2

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 35 / 30
We know that [T ]B ′ ,B = [ [T (v1 )]B ′ | [T (v2 )]B ′ ]
T (x) = x2 = 1(x2 ) + 0(x − 1) + 0(x + 1) so that
 
1
[T (v1 )]B ′ = 0 .
0

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 36 / 30
Similarly, T (1) = x = 0(x2 ) + 12 (x − 1) + 12 (x + 1) implies
 
0
[T (v2 )]B ′ = 1/2 .
1/2

Hence,  
1 0
[T ]B ′ ,B = 0 1/2 .
0 1/2

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 37 / 30
Exercise: Let T : P2 → P2 be the linear transformation
defined by T (p(x)) = p(2x − 1). For B = {1, x, x2 },
compute [T ]B (matrix of T with respect to basis B).
Hence, compute T (3 + 2x − x2 ), using part 1.
 
1 −1 1
Sol. [T ]B = 0 2 −4 and
0 0 4
T (3 + 2x − x ) = 8x − 4x2 .
2

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 38 / 30
Method for computing [T ]B ′ ,B : Let B = {v1 , . . . , vn }
and B ′ = {w1 , . . . , wm } be ordered bases for Rn and
Rm , respectively. Also, let T : Rn → Rm be a LT.
Compute T (vi ) for all i = 1, 2, . . . , n.
Form the augmented matrix

[w1 w2 . . . wm | T (v1 )| T (v2 )| . . . |T (vn )]

Find RREF of

[w1 w2 . . . wm | T (v1 )| T (v2 )| . . . |T (vn )].

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 39 / 30
 It will be of the form

[Im |[T (v1 )]B ′ | [T (v2 )]B ′ | . . . |[T (vn )]B ′ ].

Extract the matrix [T ]B ′ ,B from the right hand side

of the matrix in Step 4.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 40 / 30
Exercise. Consider a LT

T : R2 → R2 given by
   
x1 x1 − x2
T =
x2 x1 + x2
and let B = {u1 , u2 } be the basis for which
   
1 0
u1 = , u2 =
−1 1

Compute [T ]B .
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 41 / 30
Solution: Note that
   
2 −1
T (u1 ) = , T (u2 ) =
0 1

In order to find [T (u1 )]B and [T (u2 )]B , construct

 
1 0 2 −1
−1 1 0 1

The RREF of the above matrix is

 
1 0 2 −1
0 1 2 0
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 42 / 30
Hence,  
2 −1
[T ]B =
2 0
Example: Consider the LT T : R2 → P2 , given by

T (a, b) = (−a + 5b)x2 + (3a − b)x + 2b

with ordered bases B = ((5, 3), (3, 2)) and

B ′ = (3x2 − 2x, −2x2 + 2x − 1, x2 − x + 1)

of R2 and P2 , respectively. Compute [T ]B ′ ,B .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 43 / 30
Solution: Since T (5, 3) = 10x2 + 12x + 6 and
T (3, 2) = 7x2 + 7x + 4. Consider
 
3 −2 1 10 7
 −2 2 −1 12 7 
0 −1 1 6 4

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 44 / 30
RREF of the above matrix is
 
1 0 0 22 14
 0 1 0 62 39 
0 0 1 68 43

so that  
22 14
[T ]B ′ ,B = 62 39
68 43

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 45 / 30
Exercise: Consider the LT T : R3 → R2 , given by
T (x, y, z) = (x + y, y − z). Compute [T ]B ′ ,B with
respect to bases B = ((1, 0, 1), (0, 1, 1), (1, 1, 1)) and
B ′ = ((1, 2), (−1, 1)).
 
0 1/3 2/3
Answer: [T ]B ′ ,B =
−1 −2/3 −4/3

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 46 / 30
Exercise: Consider the LT T : P3 → M22 , given by
 
3 2 −3a − 2c −b + 4d
T (ax + bx + cx + d) = .
4b − c + 3d −6a − b + 2d

Compute [T ]B ′ ,B with respect to standard bases for P3

and M22 .
Answer:  
−3 0 −2 0
0 −1 0 4
[T ]B ′ ,B =  
0 4 −1 3
−6 −1 0 2

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 47 / 30
Example: Let the matrix of LT T : P1 → P1 with

2 3
respect to basis B = (x + 1, x − 1) be . Find
−1 −2
the matrix of T with respect
 to basis C = (x, 1).
2 3
Solution: Since [T ]B = , we have
−1 −2

T (x + 1) = 2(x + 1) − 1(x − 1) = x + 3
T (x − 1) = 3(x + 1) − 2(x − 1) = x + 5

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 48 / 30
  
a+b a−b
T (ax + b) = T (x + 1) + (x − 1)
2 2
 
a+b a−b
T (ax + b) = (x + 3) + (x + 5)
2 2

so that T (x) = x + 4 and T (1) = −1. Hence,

 
1 0
[T ]C = .
4 −1

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 49 / 30
Exercise: Let B = ((1, 2), (2, −1)) and
B ′ = ((1, 0), (0, 1)) be ordered bases for R2 . If 
4 3
T : R2 → R2 be a LT such that [T ]B ′ ,B = . Find
2 −4
T (5, 5). Also, find T (x, y) for all (x, y) ∈ R2 .
Answer: T (5, 5) = (15, 2).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 50 / 30
Exercise: Let
B = ((1, 1, 0, 0), (0, 1, 1, 0), (0, 0, 1, 1), (0, 0, 0, 1)) and
B ′ = ((1, 1, 1), (1, 2, 3), (1, 0, 0))
be ordered bases for R4 and R3 , respectively.
 If 
1 1 0 0
T : R4 → R3 be a LT such that [T ]B ′ ,B = 0 1 1 0.
0 1 0 1
Find T ?

Answer:
T (x1 , x2 , x3 , x4 ) = (−2x1 + 3x2 + x4 , x2 + 2x3 , x2 + 3x3 ).
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 51 / 30
Theorem. Let U, V and W be finite dimensional vector
spaces with bases B, B ′′ and B ′ , respectively. Let
T1 : U → V and T2 : V → W are linear transformations.
Then
[T2 ◦ T1 ]B ′ ,B = [T2 ]B ′ ,B ′′ [T1 ]B ′′ ,B

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 52 / 30
Exercise
Let T1 : P1 → P2 be the linear transformation defined by
T1 (c0 + c1 x) = 2c0 − 3c1 x
and let T2 : P2 → P3 be the linear transformation defined
by
T2 (c0 + c1 x + c2 x2 ) = 3c0 x + 3c1 x2 + 3c2 x3 .
Let B = {1, x}, B ′′ = {1, x, x2 } and B ′ = {1, x, x2 , x3 }.
Then
Compute [T2 ]B ′ ,B ′′ and [T1 ]B ′′ ,B , and hence
Compute [T2 ◦ T1 ]B ′ ,B .
Compute T2 ◦ T1 .
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 53 / 30
Answer:
 
  0 0 0
2 0 3 0 0
[T1 ]B ′′ ,B = 0 −3 and [T2 ]B ′ ,B ′′ =
0

3 0
0 0
0 0 3
 
0 0
6 0 
[T2 ◦ T1 ]B ′ ,B = [T2 ]B ′ ,B ′′ [T1 ]B ′′ ,B = 
0 −9
0 0

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 54 / 30
Since T2 ◦ T1 : P1 → P2 and let v = a + bx ∈ P1 . We
know that for any v ∈ P1 , we have

[T2 ◦ T1 ]B ′ ,B [v]B = [T2 ◦ T1 (v)]B ′

   
0 0   0
6 0 a =  6a 
 
[T2 ◦ T1 (v)]B ′ =  0 −9 b −9b
0 0 0
so that
(T2 ◦ T1 )(v) = 6ax − 9bx2
(T2 ◦ T1 )(a + bx) = 6ax − 9bx2
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 55 / 30
Theorem. If T : V → V is a linear operator on a finite
dimensional vector space V , and B is a basis of V , then
the following are equivalent.
T is one-to-one.
[T ]B is invertible.
Moreover,
[T −1 ]B = [T ]−1
B

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 56 / 30
Exercise
Let the matrix of a linear transformation T : R3 → R3
relative to the ordered basis
 
0 1 1
B = {(1, 0, 0), (0, 1, 0), (0, 0, 1)} be  1 0 −1  .
−1 −1 0

Find the matrix of T relative to the ordered basis

B ′ = {(0, 1, −1), (−1, 1, 0), (1, −1, 1)}.
Answer:  
1 0 0
[T ]B ′ = 0 −1 0
0 0 0
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 29, 2024 57 / 30
MATH F112 (MATHEMATICS-II)
Gaurav Dwivedi
Department of Mathematics
BITS Pilani, Pilani Campus.
 Module 6
Eigenvalues and Eigenvectors
Definition: Let A be n × n matrix. A nonzero vector
x ∈ Rn is called eigenvector of A if

Ax = λx

for some scalar λ. The scalar λ is an eigenvalue of A and

x ∈ Rn is called eigenvector corresponding to eigenvalue
λ.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 3/1
Example
For the matrix
 
−4 8 −12
A =  6 −6 12  .
6 −8 14
   
4 4
Note that A 3 = 2 3 implies λ = 2 is an
0 0  
4
eigenvalue of A and x = 3 is an eigenvector

0
corresponding to 2.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 4/1
Q: Is the eigenvector corresponding to an eigenvalue
unique?
If x is an eigenvector of A corresponding to an
eigenvalue λ i.e. Ax = λx. Then for c ∈ R, we have

A(cx) = c(Ax) = c(λx) = λ(cx).

Thus, if x is an eigenvector of A corresponding to an

eigenvalue λ then, for c ∈ R, cx is also an eigenvector
corresponding to λ. Hence, there are infinitely many
eigenvectors corresponding to an eigenvalue.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 5/1
Question: How to find eigenvalues and eigenvector of a
given matrix?
Since Ax = λx = λIn x implies (λIn − A)x = 0. Thus,
eigenvector x corresponding to λ is a nontrivial solution
of the homogeneous system whose coefficient matrix is
λIn − A. Therefore, |λIn − A| = 0.
Theorem: Let A be n × n matrix. Then λ is an
eigenvalue of A if and only if

|λIn − A| = 0 (characteristic equation of A).

The polynomial p(λ) = |λIn − A| of degree n is called

the characteristic polynomial of A.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 6/1
Note that The eigenvectors are the nontrivial solutions
of the homogeneous system

(λIn − A)x = 0.

For an n × n matrix A, the polynomial

p(λ) = |λIn − A|

of degree n is called the characteristic polynomial of

A.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 7/1
Example
Find the characteristic polynomial and eigenvalues of
 
1 0 1
A = 0 2 −3
0 0 −5

Solution: The characteristic polynomial

λ−1 0 −1
p(λ) = |λI3 − A| = 0 λ−2 3
0 0 λ+5

= (λ − 1)(λ − 2)(λ + 5).

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 8/1
Thus, p(λ) = 0 gives the eigenvalues of A. Hence,
eigenvalues are λ1 = 1, λ2 = 2 and λ3 = −5.

Exercise: Find all eigenvalues of the matrix A, where

 
4 0 −2
A = 6 2 −6
4 0 −2

Answer: λ1 = 0 and λ2 = 2

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 9/1
Theorem

If A is an n × n matrix, the following statements are

equivalent.
1 λ is an eigenvalue of A.
2 The homogeneous system (λI − A)x = 0 has
nontrivial solutions.
3 There is a nonzero vector x such that Ax = λx.
4 λ is a solution of the characteristic equation
|λI − A| = 0.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 10 / 1
Note that the eigenvectors of a matrix A corresponding
to an eigenvalue λ are the nonzero vectors in the null
space of the matrix λI − A.

Definition: The eigenspace of A corresponding to

eigenvalue λ, denoted by Eλ , is the solution space of the
homogenous system (λI − A)x = 0 i.e.

Eλ = {x ∈ Rn : (λI − A)x = 0}
Verify that Eλ is a subspace of Rn .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 11 / 1
Example
Find all the eigenvalues and corresponding eigenspaces of
the matrices
 
  4 0 −2
1 3
A= and B = 6 2 −6
0 1
4 0 −2

Solution: The characteristic polynomial of A is

λ − 1 −3
p(λ) = |λI − A| = = (λ − 1)2 .
0 λ−1

Since eigenvalues are roots of characteristic equation

p(λ) = 0. Hence, eigenvalues are λ = 1, 1.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 12 / 1
To compute eigenspace E1 for λ = 1, we need to solve
the homogeneous system (λI − A)x = 0, i.e.,
(I − A)x = 0. The augmented matrix is
 
0 −3 0
[I − A|0] =
0 0 0

which reduces to  
0 1 0
.
0 0 0
So the associated system is x2 = 0. Since column 1 is
not a pivot column, x1 is an independent variable. Let
x1 = a ∈ R.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 13 / 1
Then
E1 = {(a, 0) | a ∈ R} = {a(1, 0) | a ∈ R}.
For the matrix B, the characteristic polynomial
λ−4 0 2
p(λ) = |λI − B| = −6 λ − 2 6 = λ(λ − 2)2 .
−4 0 λ+2
Hence, eigenvalues are λ1 = 0 and λ2 = 2.
To compute eigenspace E0 for λ = 0, we need to solve
the homogeneous system
(λI − B)x = 0 implies − Bx = 0.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 14 / 1
The augmented matrix is
 
−4 0 2 0
[−B|0] =  −6 −2 6 0 
−4 0 2 0

which reduces to
 
1 0 −1/2 0
 0 1 −3/2 0  .
0 0 0 0

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 15 / 1
The associated system is
1 3
x1 − x3 = 0 and x2 − x3 = 0
2 2
Since column 3 is not a pivot column, x3 is an
independent variable. Let x3 = 2c so that
x1 = c, x2 = 3c. Then
E0 = {(c, 3c, 2c) | c ∈ R} = {c(1, 3, 2) | c ∈ R}.
Note that
E0 = span{(1, 3, 2)} = span(B),
where B = {(1, 3, 2)}. Since, B is LI, it is a basis for E0 .
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 16 / 1
To compute eigenspace E2 for λ = 2, we need to solve
the homogeneous system (λI − B)x = 0, i.e.,
(2I − B)x = 0. The augmented matrix is
 
−2 0 2 0
[2I − B|0] =  −6 0 6 0 
−4 0 4 0

which reduces to
 
1 0 −1 0
 0 0 0 0 .
0 0 0 0

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 17 / 1
The associated system is x1 − x3 = 0
Since columns 2 and 3 are not pivot columns, x2 and x3
are independent variables. Let x2 = b and x3 = c so that
x1 = c. Then

E2 = {(c, b, c) | b, c ∈ R}
= {b(0, 1, 0) + c(1, 0, 1) | b, c ∈ R}
= span(B), where B = {(0, 1, 0), (1, 0, 1)}.

Since, B is LI ((verify it)), it is a basis for E2 .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 18 / 1
Theorem: Let A be a square matrix with eigenvalue λ
and corresponding eigenvector x.
If λ is an eigenvalue of a matrix A, then for any
positive integer n, λn is an eigenvalue of An with
corresponding eigenvector x.
If A is nonsingular, then 1/λ is an eigenvalue of
A−1 with corresponding eigenvector x.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 19 / 1
Example
 
1
Let A be a 2 × 2 matrix with eigenvectors v1 =
  −1
1
and v2 = corresponding to eigenvalues
2  
5
λ1 = −1, λ2 = 2. Let x = . Find A10 x without
1
computing the matrix A.

Solution: 
2051
A10 x =
4093 2×1

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 20 / 1
Exercise
Find all the eigenvalues and an associated eigenvector of
the A25 , where
 
−1 −2 −2
A= 1 2 1 
−1 −1 0

Solution: Note that eigenvalues of A are λ1 = λ2 = 1

and λ3 = −1. For the eigenvalue 1, the basis for E1 is:
   
 −1 −1 
 0 , 1 
1 0
 

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 21 / 1
For the eigenvalue −1, the basis for E−1 is:
 
 2 
−1
1
 

Therefore, eigenvalues of A25 are 125 = 1 and

(−1)25 = −1. Note that A25 = A, thus the bases for
eigenspaces corresponding to these eigenvalues are
       
 −1 −1   −1 −1 
 0  ,  1  , and  0  ,  1 
1 0 1 0
   

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 22 / 1
Exercise

Find all the eigenvalues and bases of eigenspaces of

A and A2 , where
 
1 0
A=
0 −1

Also, compare the eigenspaces of A and A2 .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 23 / 1
Theorem 1 A square matrix A is invertible if and only if
λ = 0 is not an eigenvalue of A.

Theorem 2 Let λ1 , λ2 , . . . , λn be n eigenvalues

(repetitions included) of an n × n matrix A. Then

det(A) = λ1 λ2 · · · λn

and
Trace(A) = λ1 + λ2 + . . . + λn .

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 24 / 1
Proof of Theorem 2
a11 · · · a1n
 

Let A =  ... . . . ...  and λ1 , . . . , λn be eigenvalues

an1 · · · ann
of A. Then the characteristic polynomial of A
p(λ) = |λI − A| can be factorized as
p(λ) = (λ − λ1 ) . . . (λ − λn ).
The constant term of p(λ) is given by p(0), which can
be calculated in two ways: Firstly,
p(0) = (0 − λ1 ) . . . (0 − λn ) = (−1)n λ1 . . . λn . Secondly,
p(0) = |0I − A| = | − A| = (−1)n |A|. Therefore
|A| = λ1 . . . λn . Similarly, we can also prove that
Trace(A) = λ1 + λ2 + . . . + λn by looking at the sum of
roots in p(λ) and |λI − A|.
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 25 / 1
Theorem
Let A be an n × n matrix. The following statements are
equivalent:
A is invertible.
Ax = b has a unique solution for every b ∈ Rn .
The homogenous system Ax = 0 has only the trivial
solution.
The reduced row echelon form of A is In .
A is expressible as a product of elementary matrices.
det(A) ̸= 0.
The column vectors of A are linearly independent.
The column vectors of A span Rn .
Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 26 / 1
Theorem. (contd.)
The column vectors of A form a basis of Rn .
The row vectors of A are linearly independent.
The row vectors of A span Rn .
The row vectors of A form a basis of Rn .
A has rank n.
A has nullity 0.
The range of TA is Rn .
TA is one-to-one.
λ = 0 is not an eigenvalue of A.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 27 / 1
Exercises
 
3 2
Find the eigenvalue and eigenspaces of
5 0
Hence, find the eigenvalues and eigenspaces of A−1 ,
A − 2I, and A + 2I.
A and AT have the same characteristic polynomial.
Do A and AT have the same eigenvectors in
general?
If x is an eigenvector of A corresponding to an
eigenvalue λ, then for a scalar c, show that x is an
eigenvector of A − cI with corresponding eigenvalue
λ − c.

Gaurav Dwivedi (BITS Pilani) MATH F112 (MATHEMATICS-II) February 24, 2024 28 / 1