Indian Institute of Technology Kanpur

Department of Mathematics and Statistics

Lecture 2: Notes on several variables calculus
Semester: 2024-2025 (II)
By
Dr. Bidyut Sanki

Contents

1. Norm and distance function 2

1.1. Euclidean distance function 2
1.2. Norm 2
2. Inner Product 6
3. Matrix of a linear transformation 8

1
2

1. Norm and distance function

1.1. Euclidean distance function. Here, we study Euclidean norm and distance
function in Rn . Euclidean space. Rn = {(x1 , x2 , . . . , xn )|xi ∈ R, 1 ≤ i ≤ n} is
called the n-dimensional Euclidean space when equipped with the Euclidean distance
function. In particular,

• n = 1, R1 (= R) is called the real line.

• n = 2, R2 is caled the plane.
• n = 3, R3 is called the space, etc.

Distance function. For x = (x1 , . . . , xn ) and y = (y1 , . . . , yn ) ∈ Rn , the distance

between them is
n
!1
X 2

d(x, y) = (xi − yi )2 .
i=1
The pair (Rn , d)
is called the Euclidean space of dimension n. The distance function
d, defined above, is called Euclidean distance.
Note. The distance function d has the following fundamental properties:

(1) Positive definite: d(x, y) ≥ 0, for all x, y ∈ Rn .

(2) Symmetric: d(x, y) = d(y, x), for all x, y ∈ Rn .
(3) Triangle inequality: d(x, z) ≤ d(x, y) + (y, z), for all x, y, z ∈ Rn .

1.2. Norm.

Definition 1. Let V be a vector space over a field F (⊂ C). A norm || · || is a

function
|| · || : V → R
satisfying the following:

(1) Positive definiteness: kxk ≥ 0 for all x ∈ V , with equality if and only if
x = 0.
(2) Triangle inequality: kx + yk ≤ kxk + kyk, for all x, y ∈ V .
(3) Absolute homogeneity: kaxk = |a|kxk, for all a ∈ F and x ∈ V .

Euclidean norm: Let x = (x1 , x2 , . . . , xn ) ∈ Rn .

Definition 2. The Euclidean norm of x is given by:

n
!1
X 2
2
kxk = xi .
i=1

Example 1. For xθ = (cos θ, sin θ) ∈ R2 , we have kxθ k = 1.

 3

Exercise 1. Prove that the Euclidean norm satisfies all the properties in the defi-
nition of norm.

Vector space operations.

(1) Addition: For x = (x1 , x2 , . . . , xn ) and y = (y1 , y2 , . . . , yn ) ∈ Rn , the addi-

tion of x and y is given by
x + y = (x1 + y1 , x2 + y2 , . . . , xn + yn ).
(2) Scalar multiplication: For x ∈ Rn as above and a scalar r ∈ R, we define
rx = (rx1 , rx2 , . . . , rxn ).

With the abovementioned operations, Rn is a vector space of dimension n over the

field R. Next, we study the relation between the norm and vector space structure
of Rn below:
Properties of norm.

Theorem 1. If x, y ∈ Rn and a ∈ R, then

(1) Positive definiteness: kxk ≥ 0 and kxk = 0 if and only if x = 0.

n
P
(2) Cauchy-Schwarz inequality: xi yi ≤ kxk × kyk with equality if and only
i=1
if x and y are linearly dependent.
(3) Triangle inequality: kx + yk ≤ kxk + kyk.
(4) Absolute homogeneity: kaxk = |a|kxk.

Proof. (1) The inequality kxk ≥ 0 follows from the fact that the sum of (finitely
many) non-negative reals is non-negative.
Now,
kxk = 0
⇐⇒ kxk2 = 0
⇐⇒ x21 + · · · + x2n = 0
⇐⇒ x1 = 0, . . . , xn = 0
⇐⇒ x = 0.
(2) Case 1: Suppose x and y are linearly dependent. Then it is straightforward to
see that equality holds. To see this, suppose x and y are linearly dependent,
which implies that
x = λy, or y = λx, for some λ ∈ R.
4

Assume x = λy. Then we have

n
X n
X
xi yi = |λ| yi2
i=1 i=1
= |λ| × kyk2
= (|λ| × kyk) × kyk
= kλyk × kyk
= kxkkyk.

Case 2: Let x and y are linearly independent. Then λy − x is nonzero, for

every λ ∈ R. This implies for all λ ∈ R, we have:

n
X
2
kλy − xk = (λyi − xi )2 > 0
i=1
n n n
! ! !
X X X
=⇒ λ2 yi2 − 2λ xi yi + x2i > 0.
i=1 i=1 i=1

The left-hand side of the inequality above is a quadratic equation in λ, which

has no real root and hence (discriminant < 0)

n
!2 n
! n
!
X X X
4 xi yi −4 x2i yi2 <0
i=1 i=1 i=1
n
!2 n
! n
!
X X X
=⇒ x i yi < x2i yi2
i=1 i=1 i=1
n
!2
X
=⇒ x i yi < kxk2 kyk2 .
i=1
n
X
=⇒ xi yi < kxkkyk.
i=1

Now, the inequality follows.

 5

(3) For x, y ∈ Rn , we have

n
X
2
kx + yk = (xi + yi )2
i=1
Xn
= (x2i + 2xi yi + yi2 )
i=1
Xn n
X n
X
= x2i +2 x i yi + yi2
i=1 i=1 i=1
Xn n
X n
X
≤ x2i + 2 x i yi + yi2
i=1 i=1 i=1
2 2
≤ kxk + 2kxkkyk + kyk
=⇒ kx + yk2 ≤ (kxk + kyk)2
=⇒ kx + yk ≤ kxk + kyk,

which is what we wanted to show.

(4) For a ∈ R and x ∈ R, we have

n
!1 n
!1
2 2
X 1 X
kaxk = (axi )2 = (a2 ) 2 x2i = |a| × kxk.
i=1 i=1

Exercise 2. When does equality hold in (3)?

√
Note 1.3. Key definition used here is the following: For a ∈ R, we have a2 = |a|.

Exercise 3. Let f and g be integrable on [a, b].

(1) Prove that

 b  21  b  12
Zb Z Z
2
f · g ≤  f  ·  g2 .
a a a

(2) If equality holds, must f = λg for some λ ∈ R? What if f and g are

continuous?
6

2. Inner Product

Definition 3. Let x = (x1 , . . . , xn ) and y = (y1 , . . . , yn ) ∈ Rn . The Euclidean inner

product of x and y is denoted and defined by
n
X
hx, yi := x i yi .
i=1

Exercise 4. A linear transformation T : Rn → Rn is called norm preserving if

kT (x)k = kxk, for all x ∈ Rn ; and inner product preserving if hT (x), T (y)i =
hx, yi, for all x, y ∈ Rn .

(1) Prove that T is norm preserving if and only if T is inner product preserving.
(2) Prove that such a linear transformation is 1 − 1 and T −1 is of the same sort.

Properties of inner product

Theorem 2. Suppose X, X1 , X2 , Y, Y1 , Y2 ∈ Rn and a ∈ R. Then

(1) Commutativity: hX, Y i = hY, Xi

(2) Bi-linearity:

haX, Y i = ahX, Y i
hX1 + X2 , Y i = hX1 , Y i + hX2 , Y i

Ex. Show that hX, aY i = ahX, Y i and hX, Y1 + Y2 i = hX, Y1 i + hX, Y2 i.

(3) Positive definiteness:
p hX, Xi ≥ 0, and hX, Xi = 0 ⇐⇒ X = 0.
(4) Norm: kXk = hX, Xi.
2 k2
(5) Polarization identity: hX, Y i = kX+Y k −kX−Y
4 .

Proof. (1) Let X = (x1 , x2 , . . . , xn ) and Y = (y1 , y2 , . . . , yn ). Then

n
X
hX, Y i = xi yi (as multiplication of reals is commutative)
i=1
Xn
= yi xi = hY, Xi.
i=1

Note. The key fact used here is the following: The multiplication of reals
is commutative, i.e., ab = ba, for all a, b ∈ R.
 7

n
P n
P
(2) (a) haX, Y i = axi yi = a xi yi = ahX, Y i.
i=1 i=1
Now, we have

hX, aY i = haY, Xi (by (1))

= ahY, Xi (by the previous part)
= ahX, Y i (by (1)).

Thus, we conclude that haX, Y i = hX, aY i = ahX, Y i.

(b) Let X1 = (a1 , . . . , an ) and X2 = (b1 , . . . , bn ). Then
n
X
hX1 + X2 , Y i = (ai + bi )yi
i=1
n
X n
X
= ai yi + bi yi
i=1 i=1
= hX1 , Y i + hX2 , Y i.

(c)

hX, Y1 + Y2 i = hY1 + Y2 , Xi (by (1))

= hY1 , Xi + hY2 , Xi (by the previous item)
= hX, Y1 i + hX, Y2 i (by (1)).

(3) Note that hX, Xi = kXk2 . Therefore, the proof of

hX, Xi ≥ 0, and hX, Xi = 0 ⇐⇒ X = 0

is exactly similar to the proof of (1) in Theorem 1.

n 1  n 1
P 2 2 P 2 1
(4) kXk = xi = xi · xi = hX, Xi 2 .
i=1 i=1
(5) Now, we prove the polarisation identity. We start with the right-hand side
here:
kX + Y k2 − kX − Y k2
4
1
= [hX + Y, X + Y i − hX − Y, X − Y i]
4
1
= [(hX, Xi + hX, Y i + hY, Xi + hY, Y i) − (hX, Xi − hX, Y i − hY, Xi + hY, Y i)]
4
1
= × 4hX, Y i = hX, Y i.
4


8

Exercise 5. If x, y ∈ Rn are nonzero, the angle between x and y, denoted by ∠(x, y),
is defined by
 
hx, yi
∠(x, y) = arc cos
kxkkyk
The linear transformation T is called angle preserving if
∠(T (x), T (y)) = ∠(x, y), ∀x, y ∈ Rn .

(1) Prove that if T is norm preserving, then T is angle preserving.

(2) If there is a basis {xi | i = 1, . . . , n} of Rn and real numbers λi , i = 1, . . . , n,
such that T (xi ) = λi x. If |λi |’s are equal, does it imply that T is angle
preserving?

Exercise 6. Let h−, −i be an inner product and k·k be the norm induced by the
inner product on the vector space V . Prove that
ku + vk2 + ku − vk2 = 2(kuk2 + kvk2 ), for all u, v ∈ V.

3. Matrix of a linear transformation

First, we fix some notations:

0 = (0, 0, . . . , 0) ∈ Rn
ei = (0, . . . , 0, 1, 0, . . . , 0), only ith place is 1
B = {ei | 1 ≤ i ≤ n} is the standard ordered basis for Rn .

Definition 4. A map T : Rn → Rm is a linear transformation if

T (aX + Y ) = aT (X) + T (Y ),
for all a ∈ R and X, Y ∈ Rn .

Let us consider T : Rn → Rm be a linear transformation. Suppose,

T (ej ) = (a1,j a2,j . . . am,j )t ,
which is the jth column of the matrix
 
a1,1 a1,2 . . . a1,n
 a2,1 a2,2 . . . a2,n 
A= . ..  .
 
.. ..
 .. . . . 
am,1 am,2 . . . am,n
Definition 5. The matrix A is called the matrix representation of the linear trans-
formation T with respect to the standard ordered basis.
 9

Example 2. Let T : R3 → R4 be defined by

T (x, y, z) = (3x + 2y + z, x + y + z, x − 3y, 2x + 3y + z).
Then  
3 2 1
1 1 1 
[T ] =  .
 
1 −3 0
2 3 1

Exercise 7. If T : Rn → Rm is a linear transformation, show that ∃ M ∈ R such

that
kT (h)k ≤ M khk, ∀h ∈ Rn .