co

Differentiability

Mathematical analysis 2
Chapter 1: Multivariable and vectorial functions

n
Part 3: Differentiability

te
es R. KECHKAR
il
2023/2024
R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function
 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Plan

1 Differentiability

n
First partial derivatives
Diractionnal derivation
Partial derivatives of numerical function

te
Partial derivatives of vectorial function
Functions of class C1
Higher order partial derivatives
Second order partial derivatives
Functions of class Ck
es
Differentiability
Differentiability of numerical functions
Differentiability of vectorial functions
Differentiation of composite functions (The chain rule)
Taylor’s formula
Implicit derivation
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Plan

1 Differentiability

n
First partial derivatives
Diractionnal derivation
Partial derivatives of numerical function

te
Partial derivatives of vectorial function
Functions of class C1
Higher order partial derivatives
Second order partial derivatives
Functions of class Ck
es
Differentiability
Differentiability of numerical functions
Differentiability of vectorial functions
Differentiation of composite functions (The chain rule)
Taylor’s formula
Implicit derivation
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Diractionnal derivation

n
Definition
Let f be a function defined on D ⊆ Rn → R and v = (v1 , . . . , vn ) ∈ Rn a

te
non null vector. We say f is derivable at a = (a1 , . . . , an ) ∈ D in the
diriction of v if
f (a + tv) − f (a)
lim exist and is finite.
es t →0 t
This limit is denoted by Dv f (a)·
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Diractionnal derivation

Example.
Find the derivatives of f in the direction of v1 = (2, 1) et v2 = (1, 0) at point

n
a = (0, 0) if there exist , where :f (x, y) = x2 − 2xy + | sin(y)|.
• In the direction of v1 :

te
f ((0, 0) + t(2, 1)) − f (0, 0) f (2t, t) − f (0, 0) | sin(t)|
lim = lim = lim Ø.
t →0 t t →0 t t →0 t
Then f is not derivable at the point (0, 0) in the direction of v1 .
• In the direction of v2 :
es
f ((0, 0) + t(1, 0)) − f (0, 0) f (t, 0)) − f (0, 0) t2
lim = lim = lim = 0.
t →0 t t →0 t t →0 t

Then f is derivable at (0, 0) in the direction of v2 and we have

il
Dv2 f (0, 0) = 0

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Plan

1 Differentiability

n
First partial derivatives
Diractionnal derivation
Partial derivatives of numerical function

te
Partial derivatives of vectorial function
Functions of class C1
Higher order partial derivatives
Second order partial derivatives
Functions of class Ck
es
Differentiability
Differentiability of numerical functions
Differentiability of vectorial functions
Differentiation of composite functions (The chain rule)
Taylor’s formula
Implicit derivation
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Partial derivatives

Definition
Let e1 , . . . , ei , . . . , en the canonical base of Rn and f : D ⊆ Rn −→ R.
¡ ¢

n
¡ ¢
• We call first partial derivative of f at the point a = a1 , . . . , an ∈ U with respect
∂f
to xi variable and we denote ∂x (a), the derivative of f at a in the direction of

te
i
the vector ei , that is to say
∂f
¡ ¢
f a + tei − f (a)
(a) = Dei f (a) = lim
∂xi t →0 t
¡ ¢ ¡ ¢
f a1 , . . . , ai + t, . . . , an − f a1 , . . . , ai , . . . , an
= lim
t →0 t
es ¡ ¢ ¡
f a1 , . . . , xi , . . . , an − f a1 , . . . , ai , . . . , an
¢
= lim .
x i →a i xi − ai
¡ ¢
• In particular, for n = 2 the first partial derivatives of f at the point x0 , y0 ∈ U
are written as follow¡
∂f ¡ ∂f ¡
¢ ¡ ¢ ¡ ¢ ¡ ¢
¢ f x, y0 − f x0 , y0 ¢ f x0 , y − f x0 , y0
x0 , y0 = lim , x0 , y0 = lim .
∂x x→x0 ∂y y→y0
il
x − x0 y − y0

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Gradient

Definition

n
Let f : D ⊆ Rn −→ R, a ∈ D. Suppose that all first partial derivatives of
f exist. The vector denoted by ∇f (a) ∈ R defined by

te
∂f
 
 ∂x ( a) 
 1 
 . 
∇f (a) =  .. 
 
es  
 ∂f
 

(a)
∂xn

is called gradient of f at a.
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Gradient: Example

Example.

n
Find the first partial derivatives of f at all points of f R2 . if there exist, where
f the function defined by ( xy
si (x, y) , (0, 0) ,

te
f (x, y) = x + y2
2
0 si (x, y) = (0, 0) .
1. On R2∗ the partial derivatives of first order exist since f is the quotient of
tow polynomial and we have ∀(x, y) ∈ R2∗ ,
es ∂f y(x2 + y2 ) − xy(2x) y2 − x2
(x, y) = =y 2 .
∂x 2
(x + y )2 2 (x + y2 )2

∂f x(x2 + y2 ) − xy(2y) x 2 − y2
(x, y) = = x .
∂y (x 2 + y 2 )2 (x 2 + y 2 )2
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Gradient: Example
Then on R2∗ the gradient of f is y2 − x 2
 
y 2
 (x + y2 )2 

∇f (x, y) = 
 x2 − y2 

n

x 2 22
(x + y )

te
• The first partial derivatives at (0, 0)

∂f f (x, 0) − f (0, 0) 0
(0, 0) = lim = lim 3 = 0 ∈ R
∂x x →0 x−0 x →0 x

∂f
es (0, 0) = lim
f (0, y) − f (0, 0) 0
= lim 3 = 0 ∈ R.
∂y y →0 y−0 y →0 y

Then at (0, 0) the gradient of f is

 
0
∇f (0, 0) =  
il
0

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Gradient: Important remark

n
Remark

te
• Note that for the pervious example, limit of f at (0, 0) does not
exist then f is not continuous at (0, 0) however it is derivable.
• Generally, the existence of partial derivatives at a point a does
es
not imply continuity at a.
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Tangent plane

n
Definition

te
Suppose f has continuous partial derivatives. An equation of the
tangent plane to the surface z = f (x, y) at the point (x0 , y0 , f (x0 , y0 )) is
∂f ∂f
z= (x0 , y0 ) · (x − x0 ) + (x0 , y0 ) · (y − y0 ) + f (x0 , y0 )
∂x ∂y
es
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Plan

1 Differentiability

n
First partial derivatives
Diractionnal derivation
Partial derivatives of numerical function

te
Partial derivatives of vectorial function
Functions of class C1
Higher order partial derivatives
Second order partial derivatives
Functions of class Ck
es
Differentiability
Differentiability of numerical functions
Differentiability of vectorial functions
Differentiation of composite functions (The chain rule)
Taylor’s formula
Implicit derivation
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Partial derivatives of vectorial function

n
Definition
Let f : D ⊆ Rn → Rm (m ≥ 2) and a = (a1 , · · · an ) ∈ D. Then the

te
derivative with respect to the variable xi at the point a of f exist if and
only if the derivatives with respect to the variable xi at the point a of
all fj : D ⊆ Rn → R for j = 1, · · · , m exist. In this case we have

∂f ∂f1 ∂fm
es (a ) =
µ
(a), · · · , (a )
¶

∂xi ∂xi ∂xi

il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Jacobian Matrix

Definition
Let f : D ⊆ Rn → Rm (m ≥ 2) that all partial derivatives at the point a

n
exist. The matrix
∂f1 ∂f1 ∂f1

te
 
 ∂x (a) ∂x2
(a ) · · ·
∂xn
(a) 
 1 
 ∂f2 ∂f2 ∂f2
 

 ∂x (a) (a ) · · · (a) 

1 ∂x2 ∂xn 
Jf (a) = 
 
es 
 ..
.
..
.
..
.
..
.



 
 
 ∂fm ∂fm ∂fm
 

(a ) (a) · · · (a)
∂x1 ∂x2 ∂xn
il
of size (m, n) is called Jacobian matrix of f at a.

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Plan

1 Differentiability

n
First partial derivatives
Diractionnal derivation
Partial derivatives of numerical function

te
Partial derivatives of vectorial function
Functions of class C1
Higher order partial derivatives
Second order partial derivatives
Functions of class Ck
es
Differentiability
Differentiability of numerical functions
Differentiability of vectorial functions
Differentiation of composite functions (The chain rule)
Taylor’s formula
Implicit derivation
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Function of class C1

n
Definition
• Let f : D ⊆ Rn → R be a function. We say that f is of calss C1 on

te
∂f
D if and only if all first partial derivatives exist and are
∂xi
continuous on D.
• Let f : D ⊆ Rn → Rm (m ≥ 2) be a vectorial function we have
es
f is of class C1 on D ⇐⇒ ∀i = 1, · · · , m fi is of class C1 on D
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Plan

1 Differentiability

n
First partial derivatives
Diractionnal derivation
Partial derivatives of numerical function

te
Partial derivatives of vectorial function
Functions of class C1
Higher order partial derivatives
Second order partial derivatives
Functions of class Ck
es
Differentiability
Differentiability of numerical functions
Differentiability of vectorial functions
Differentiation of composite functions (The chain rule)
Taylor’s formula
Implicit derivation
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Plan

1 Differentiability

n
First partial derivatives
Diractionnal derivation
Partial derivatives of numerical function

te
Partial derivatives of vectorial function
Functions of class C1
Higher order partial derivatives
Second order partial derivatives
Functions of class Ck
es
Differentiability
Differentiability of numerical functions
Differentiability of vectorial functions
Differentiation of composite functions (The chain rule)
Taylor’s formula
Implicit derivation
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Second order partial derivatives

Definition

n
Let f : D ⊆ Rn → R be a function. If the first partial derivative of f at
the point a with respect to the variable xi exist and if the derivative of
∂f

te
at the point a with respect to the variable xj exist. Then we denote
∂xi
by  2
∂ f
if j , i


∂ ∂f  ∂x ∂x
µ ¶ 
j i
es ∂xj ∂xi
(a) = 2
∂ f

 if j = i
∂x2
 2

the second order partial derivative of f with respect to the variables xi

and xj .
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Schwarz theorem

n
Theorem
∂2 f

te
Let f : D ⊆ Rn → R be a function. If the partial derivatives and
∂xi ∂xj
∂2 f
exist on an open containing a and are continuous at a then
∂xj ∂xi
es ∂2 f
(a) =
∂2 f
(a)
∂xi ∂xj ∂xj ∂xi
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Plan

1 Differentiability

n
First partial derivatives
Diractionnal derivation
Partial derivatives of numerical function

te
Partial derivatives of vectorial function
Functions of class C1
Higher order partial derivatives
Second order partial derivatives
Functions of class Ck
es
Differentiability
Differentiability of numerical functions
Differentiability of vectorial functions
Differentiation of composite functions (The chain rule)
Taylor’s formula
Implicit derivation
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Function of class Ck

Proposition

n
• Let f : D ⊆ Rn → R be a function. We say that f is of calss Ck on D
if and only if all partial derivatives up to order k exist and are

te
continuous on D.
• Let f : D ⊆ Rn → Rm (m ≥ 2) be a vectorial function we have

f is of class Ck on D ⇐⇒ ∀i = 1, · · · , m fi is of class Ck on D
es
Remark
The function f is said to be C∞ if f is of class Ck for all k ∈ N.
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Hessian matrix

Definition

n
Let f : D ⊆ Rn → R such that f ∈ C2 on D. We call Hessian matrix of f
at the point a the matrix defined by

te
∂2 f ∂2 f ∂2 f
 
···
 ∂x12
 ∂x1 ∂x2 ∂x1 ∂xn 

 
 ∂2 f ∂2 f 2
∂ f 
···
 
Jf (a) =  ∂x2 ∂x1 ∂x22 ∂x2 ∂xn 
 
es 

 .. .. .. .. 
. . . .
 
 
 
 ∂2 f ∂2 f 2
∂ f
 

···
∂xn ∂x1 ∂xn ∂x2 ∂xn2
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Second order derivatives: Example

n
Example.
Let us consider the function defined on R2 by

te
2 2

xy x − y

if (x, y) , (0, 0),
f (x, y) = x2 + y2

0 if (x, y) = (0, 0).
es
∂2 f ∂2 f
Find and at the point (0, 0). What can we conclude.
∂y∂x ∂x∂y
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Solution

• Partial derivatives of f at (0, 0) :

n
∂f f (x, 0) − f (0, 0) 0
(0, 0) = lim = lim = 0 ∈ R
∂x x →0 x−0 x →0 x
∂f

te
f (0, y) − f (0, 0) 0
(0, 0) = lim = lim = 0 ∈ R.
∂y y →0 y−0 y →0 y

• On R2∗ , f is of class C1 sice f is quotient of tow polynomials, we have

es ∂f
¡ 2
3x − y2 x2 + y2 − 2x x3 − xy2
¢¡ ¢ ¡ ¢
∀(x, y) ∈ R2∗ , (x, y) = y .
∂x
¡ ¢2
x2 + y2
¡ 2
∂f 3y − x2 x2 + y2 − 2y y3 − yx2
¢¡ ¢ ¡ ¢
∀(x, y) ∈ R2∗ , (x, y) = −x .
∂y
¡ ¢2
x2 + y2
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Solution
∂2 f ∂2 f
• Now, let’s find and at (0, 0) :
∂y∂x ∂x∂y

n
y 5
∂f ∂f − 4 −0
(0, y) − (0, 0)
∂2 f
(0, 0) = lim ∂x ∂x y

te
= lim = −1 ∈ R.
∂y∂x y→0 y−0 y→0 y
∂f ∂f x5
2 (x , 0 ) − (0 , 0 ) −0
∂ f ∂y ∂y 4
(0, 0) = lim = lim x = 1 ∈ R.
∂x∂y x →0 x−0 x →0 x
es
∂2 f ∂2 f
• Since (0 , 0 ) , (0, 0), we conclude using Schwartz
∂y∂x ∂x∂y
∂2 f ∂2 f
theorem that at least or is not continuous at (0, 0).
il
∂y∂x ∂x∂y
Then f is not C2 on R2 .
R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function
 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Plan

1 Differentiability

n
First partial derivatives
Diractionnal derivation
Partial derivatives of numerical function

te
Partial derivatives of vectorial function
Functions of class C1
Higher order partial derivatives
Second order partial derivatives
Functions of class Ck
es
Differentiability
Differentiability of numerical functions
Differentiability of vectorial functions
Differentiation of composite functions (The chain rule)
Taylor’s formula
Implicit derivation
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Plan

1 Differentiability

n
First partial derivatives
Diractionnal derivation
Partial derivatives of numerical function

te
Partial derivatives of vectorial function
Functions of class C1
Higher order partial derivatives
Second order partial derivatives
Functions of class Ck
es
Differentiability
Differentiability of numerical functions
Differentiability of vectorial functions
Differentiation of composite functions (The chain rule)
Taylor’s formula
Implicit derivation
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Differentiability definition

Definition
Let f : D ⊆ Rn → R.

n
• We say that f is differentiable at the point a ∈ Rn if there exist a linear
mapping L : Rn → R such that

te
f (a + h) − f (a) − L(h)
lim =0
h →0 ||h||

• The linear mapping L is called differential of f at the point a, denoted

es
by df (a) or dfa .
• The function f is said to be differentiable on D if it is differentiable at
every point of D.

Proposition
il
Let f : D ⊆ Rn → R be a differentiable function then L is unique.

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Differentiability results

n
Theorem

te
Let f : Rn → R be a differentiable function at point a. Then:
• The function f has partial derivatives at point a with respect to
all its variables.
∂f ∂f ∂f
es
• dfa (h) = h1
∂x1
(a) + h2
∂x2
(a ) + . . . + h n
∂xn
(a).
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Example

Example.
Let f a function defined by

n
 xy

q if (x, y) , (0, 0)

te
f (x, y) = x2 + y2

0 if (x, y) = (0, 0)


We have
∂f f (h, 0) − f (0, 0) ∂f f (0, h) − f (0, 0)
∂x
(0, 0) = lim
h →0
es h
=0 and
∂y
(0, 0) = lim
h →0 h
=0

Also,
∂f ∂f
f (0 + h, 0 + k) − f (0, 0) − (0, 0) · h − (0, 0) · k
∂x ∂y hk
lim = lim
h2 + k2
p
(h,k)→(0,0) h2 + k2 (h,k)→(0,0)
il
does not exist. Thus, f has partial derivatives with respect to x and y at the
point (0, 0), but f is not differentiable at (0, 0).
R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function
 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Differentiability

n
Theorem
Let f : D ⊂ Rn → R. Let f is differentiable at point a if and only if the

te
following two conditions are satisfied:
• f has partial derivatives at a with respect to all its variables.
• and
∂f ∂f
es
f (a1 + h1 , . . . , an + hn ) − f (a) − h1
∂x1
(a) − . . . − hn
∂xn
(a)
lim q = 0.
h→0
h21 + h22 + . . . + h2n
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Plan

1 Differentiability

n
First partial derivatives
Diractionnal derivation
Partial derivatives of numerical function

te
Partial derivatives of vectorial function
Functions of class C1
Higher order partial derivatives
Second order partial derivatives
Functions of class Ck
es
Differentiability
Differentiability of numerical functions
Differentiability of vectorial functions
Differentiation of composite functions (The chain rule)
Taylor’s formula
Implicit derivation
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Differential

Definition

n
Let f : D ⊂ Rn → Rm . (m ≥ 2) and a = (a1 , . . . , an ) ∈ D. Then, f is
differentiable at a if and only if each of the m component functions

te
fj : U ⊂ Rn → R is differentiable at a for j = 1, . . . , m. In this case, the
differential of f at a is given by:

dfa = (df1 (a), df2 (a), . . . , dfm (a))

es
Proposition
Let f : D ⊂ Rn → Rm . (m ≥ 1). If f is differentiable at a ∈ U, then f is
continuous at a.
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Relation between function of class C1 and and function differentiable

n
Proposition

te
Let f : D ⊂ Rn → Rm . (m ≥ 1). If all the partial derivatives of f exist in
a neighborhood of a ∈ U and are continuous at a, then f is
differentiable at a, i.e.

f is C1 on U =⇒ f is differentiable on U .
es
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Example

n
Example.

te
Study the differentiability of f in R2 where
 3
x + xy3
if (x, y) , (0, 0)


q
f (x, y) = x2 + y2
es 


0 if (x, y) = (0, 0)
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Solution

• On R2∗ , f is of class C1 as it is a quotient and composition of two

n
functions of class C1 .
• Study of the differentiability of f at (0, 0): If f is differentiable at

te
(0, 0), then the differential of f is expressed using partial derivatives at
(0, 0). Let’s calculate the partial derivatives of f at (0, 0):

∂f f (x, 0) − f (0, 0) x3
(0, 0) = lim = lim =0∈R
∂x
es x→0 x−0 x →0 x | x |

∂f f (0, y) − f (0, 0) 0
(0, 0) = lim = lim = 0 ∈ R
∂y y→0 y−0 y→0 y
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Solution
If f is differentiable at (0, 0), then its differential at (0, 0) is given by:
∂f ∂f
d(0,0) f (h1 , h2 ) = h1 (0, 0) + h2 (0, 0) = 0.

n
∂x ∂y
According to the definition, f is differentiable at (0, 0) if there exists a linear
transformation d(0,0) f such that:

te
f (0 + h1 , 0 + h2 ) − f (0, 0) − d(0,0) f (h1 , h2 )
lim = 0.
(h1 ,h2 )→(0,0) k(h1 , h2 )k

Let’s choose the Euclidean norm:

es f (h1 , h2 ) − f (0, 0) − d(0,0) f (h1 , h2 )
lim
(h1 ,h2 )→(0,0) k(h1 , h2 )k2
h31 + h1 h32
= lim
(h1 ,h2 )→(0,0) h21 + h22
h21 h22
il
à !
= lim h1 + h1 h2 2 = 0.
(h1 ,h2 )→(0,0) h21 + h22 h1 + h22
R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function
 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Solution

n
te
We can conclude that f is differentiable at (0, 0) with a zero
differential.
Conclusion: f is differentiable over R2 .
es
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Plan

1 Differentiability

n
First partial derivatives
Diractionnal derivation
Partial derivatives of numerical function

te
Partial derivatives of vectorial function
Functions of class C1
Higher order partial derivatives
Second order partial derivatives
Functions of class Ck
es
Differentiability
Differentiability of numerical functions
Differentiability of vectorial functions
Differentiation of composite functions (The chain rule)
Taylor’s formula
Implicit derivation
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Chain rule (case 1)

n
Proposition

te
Let f : D ⊂ R2 → R be a differentiable function on D, where x = x(t)
and y = y(t) are both differentiable functions with respect to t. Then f
is a differentiable function with respect to t
df ∂f dx ∂f dy
es = +
dt ∂x dt ∂y dt
.
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Chain rule: Example

n
Example.

te
Let consider the function f defined by

f : D = R+ × R∗+ −→ R
x
(x, y) 7−→ f (x, y) = .
y
es df
We set x = t2 y = ln t. Find using two different methods.
dt
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Solution

n
Note that f ∈ C1 on D
• Using direct method: we substitute the value of x and y in the

te
expression of f :
t2
2 t2 df 2t ln t − t 2t ln t − t
f (x, y) = f (t , ln t) = =⇒ = =
ln t dt ln2 t ln2 t
Then
df 2t ln t − t
es =
dt ln2 t
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Solution

• Using chain rule: We have

df ∂f dx ∂f dy

n
= + .
dt ∂x dt ∂y dt

te
∂f 1 1

dx

 dx = y = ln t

 = 2t
(  
x = t2
 
dt
=⇒ and
y = ln t  dy = 1
  ∂f x t2
 =− =− .
 

dt t dy y2 ln2 t
Then
es
df 1 t2 1 2t ln t − t
= · 2t − 2 · =
dt ln t ln t t ln2 t
Consequently
df 2t ln t − t
il
=
dt ln2 t

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Chain rule (case 2)

n
Proposition
Let f : D ⊂ R2 → R be a differentiable function on D x and y, where

te
x = x(s, t) and y = y(s, t) are differentiable functions with respect to s
and t. Then
∂f ∂f ∂x ∂f ∂y
= +
∂s ∂x ∂s ∂y ∂s
and
es ∂f ∂f ∂x ∂f ∂y
= + .
∂t ∂x ∂t ∂y ∂t
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Chain rule: Example

Example.

n
∂f ∂f
Given f : R2 → R of class C2 . Let u = x − y and v = x + y. Express , and
∂x ∂y

te
in terms of the partial derivatives of f with respect to u and v.
Solution.
u+v u−v
Considering that x = and y = , we have:
2 2
∂f ∂f ∂x ∂f ∂y ∂f 1 ∂f ∂f  ∂f ∂f ∂f
  µ ¶

 = +
es
 ∂u ∂x ∂u ∂y ∂u



 = +
 ∂u 2 ∂x ∂y


 = +
 ∂x ∂u ∂v


=⇒ =⇒
∂f ∂f ∂x ∂f ∂y ∂f 1 ∂f ∂f ∂f ∂f ∂f
  µ ¶ 
  
 =

 +  =

 −  =
 −
∂v ∂x ∂v ∂y ∂v ∂v 2 ∂x ∂y ∂y ∂u ∂v
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Plan

1 Differentiability

n
First partial derivatives
Diractionnal derivation
Partial derivatives of numerical function

te
Partial derivatives of vectorial function
Functions of class C1
Higher order partial derivatives
Second order partial derivatives
Functions of class Ck
es
Differentiability
Differentiability of numerical functions
Differentiability of vectorial functions
Differentiation of composite functions (The chain rule)
Taylor’s formula
Implicit derivation
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Taylor’s formula of first order

Proposition

n
• Let f : D ⊆ Rn → R be a function of class C1 on D. For
p = (α1 , . . . , αn ) ∈ D fixed, there exists a function ε defined on D

te
with lim ε(u) = 0 such that for all u = (u1 , . . . , un ) ∈ D, we have
u→p

∂f ∂f
f (u) = f (p) + (u1 − α1 ) (p) + . . . + (um − αm ) (p) + ||u − p||ε(u).
∂x1 ∂xn
es
• In particular for f : D ⊆ R2 → R we have

∂f ∂f
f (u) = f (p) + (u1 − α1 ) (p) + (u2 − α2 ) (p) + ||u − p||ε(u, v).
∂x1 ∂x2
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Taylor’s formula of second order

n
Proposition

te
Let f : D ⊆ Rn → R be a function of class C2 on D. For p ∈ D, there
exists a function ε defined on D with lim ε(u) = 0. Then, for all
u→p
u = (u1 , . . . , un ), we have
∂f 1 X ∂2 f
(p)+ku − pk2 ε(u − p).
X
f (u) = f (p)+
es
1≤i≤n
(ui − pi )
∂xi
(p)+
2 1≤i,j≤n
(ui − pi )(uj − pj )
∂xi ∂xj
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Plan

1 Differentiability

n
First partial derivatives
Diractionnal derivation
Partial derivatives of numerical function

te
Partial derivatives of vectorial function
Functions of class C1
Higher order partial derivatives
Second order partial derivatives
Functions of class Ck
es
Differentiability
Differentiability of numerical functions
Differentiability of vectorial functions
Differentiation of composite functions (The chain rule)
Taylor’s formula
Implicit derivation
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Implicit Function Theorem (2D version)

Theorem
Let F : D ⊂ Rn → R be a Ck function on D with k ≥ 1. Consider (a, b) ∈ R2

n
such that
∂F
F (a, b) = 0 and (a, b) , 0.
∂y

te
Then, there exist neighborhoods V and W of a and b and a Ck function
ϕ : V → W such that V × W ⊂ D and

∀x ∈ V , ∀y ∈ W , F (x, y) = 0 ⇐⇒ y = ϕ(x).
es
Furthermore,we have for all x ∈ V, the derivative ϕ0 (x) is given by

∂F
(x, ϕ(x))
ϕ (x) = − ∂x
0
.
∂F
(x, ϕ(x))
il
∂y

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Important remark

Remark

n
∂F
• If (a, b) , 0 then, there exist also neighborhoods V and W of a
∂y
and b and a Ck function ψ : W → V such that V × W ⊂ D

te
∀x ∈ V , ∀y ∈ W , F (x, y) = 0 ⇐⇒ x = ψ(y).

Furthermore,we have for all y ∈ W, the derivative ψ0 (y) is given

by
es
∂F
(ψ(y), y
0 ∂y
ψ (y) = − .
∂F
(ψ(y), y
∂x
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Important remark

Remark
• The result of the previous proposition can be generalized to a function

n
with n variables . If F : D ⊂ Rn → R be a Ck function on D with k ≥ 1.
and p = (α1 , . . . , αm ) ∈ D ⊂ Rm such that

te
∂f
f (p) = 0, and (p) , 0.
∂xm
Then there exists a neighborhood V of (α1 , . . . , αm−1 ) and an interval J
centered at αm such that V × J ⊂ D, and a function ϕ : V → J satisfying
∀(x1 , . . . , xm ) ∈ V × J : f (x1 , . . . , xm ) = 0 ⇐⇒ xm = ϕ(x1 , . . . , xm−1 ),
es
Furthermore ∀(x1 , . . . , xm ) ∈ V × J, ∀j ∈ {1, . . . , m − 1}
∂f
(x1 , . . . , xm−1 , ϕ(x1 , . . . , xm−1 ))
∂ϕ ∂xj
(x1 , . . . , xm−1 ) = −
∂xj ∂f
(x1 , . . . , xm−1 , ϕ(x1 , . . . , xm−1 ))
∂xm
il
.
R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function
 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Implicit derivation: example

n
Example.

te
Let consider the function f defined by

f : R2 −→ R
(x, y) 7−→ f (x, y) = x4 + x3 y2 − y + y2 + y3 − 1.
es
Apply the implicit function theorem at point p = (−1, 1).
il

R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function

 First partial derivatives

co
Differentiability Higher order partial derivatives
Differentiability

Solution
We have f ∈ C∞ (R2 ) f (−1, 1) = 0,
∂f
(x, y) = 2x3 y − 1 + 2y + 3y2 ,

n
∂y
∂f
(−1, 1) = 2 , 0.

te
∂y
Then there exist neighborhoods V of -1 and W of 1 and a function ϕ : V → W
of class C1 such that V × W ⊂ R2 and
∀x ∈ V , ∀y ∈ W , F (x, y) = 0 ⇐⇒ y = ϕ(x).
We have
es ∂f
(−1, 1)
g0 (−1) = − ∂x .
∂f
(−1, 1)
∂y
but
∂f ∂f
(x, y) = 4x3 + 3x2 y2 =⇒ (−1, 1) = −1
il
∂x ∂x
Then 1
ϕ0 (−1) = .
2
R. KECHKAR Mathematical analysis 2 Chapter 1: Multivariable and vectorial function