Derivation

CHAPTER 5

1
Syllabus
1. Principles of Mathematical Logic and Set Theory.
2. Real Numbers and Topology on ℝ and ℝ𝑛 .
3. Sequences.
4. Functions.
5. Derivation.

2
Derivatives
1. Derivatives of real functions of a real variable.
2. Derivatives of real functions of a vector variable.
3. Derivatives of vector-valued functions.
4. Derivatives of higher order
5. Differentiable functions: some properties.
6. The Chain Rule

3
Additional resources
▪Khan academy - Derivatives of multivariable functions
▪MIT Single Variable Calculus - Differentiation
▪MIT Multivariable Calculus – Partial derivatives
▪MIT Multivariable Calculus - Lecture 1
▪MIT Multivariable Calculus – Partial derivatives Lecture 1

4
Average Rate of Change

𝑓(𝑥1 )−𝑓(𝑥0 )
▪𝐴𝑅𝐶 𝑥0 ,𝑥1 =
𝑥1 −𝑥0

▪The ARC is the slope of the secant

line that goes through (𝑥0 , 𝑓 𝑥0 )
and 𝑥1 , 𝑓 𝑥1 .

5
Differentiation in ℝ
▪ What happens when we move 𝑥1 closer and closer to 𝑥0 ?

These lines are still secant to the graph!

6
Differentiation in ℝ
▪ At the limit, we will obtain a tangent line to the graph at 𝑥0 . The
slope of that line is called the derivative of the function at 𝑥0 .

7
Definition of derivative at a certain point
𝑓(𝑥0 + ℎ) − 𝑓(𝑥0 ) 𝑓(𝑥) − 𝑓(𝑥0 )
𝑓′(𝑥0 ) = lim 𝑓′(𝑥0 ) = lim
ℎ→0 ℎ 𝑥→𝑥0 𝑥 − 𝑥0

8
Derivatives using the definition
▪ Example 1: 𝑓 𝑥 = 4𝑥 2
▪ According to the rules of differentiation: 𝑓 ′ 𝑥 = 8𝑥
Using the definition:

𝑓(𝑥0 + ℎ) − 𝑓(𝑥0 ) 4(𝑥0 + ℎ)2 − 4𝑥02

𝑓′(𝑥0 ) = lim = lim =
ℎ→0 ℎ ℎ→0 ℎ

8𝑥0 ℎ + ℎ2
= lim = lim 8𝑥0 + ℎ = 8𝑥0
ℎ→0 ℎ ℎ→0

9
Derivatives using the definition
▪ Example 2: 𝑔 𝑥 = ln 𝑥
1
▪ According to the rules of differentiation: 𝑔′ 𝑥 =
𝑥

Using the definition:

𝑓(𝑥0 +ℎ)−𝑓(𝑥0 ) ln(𝑥0 +ℎ)−ln 𝑥0 ln 𝑥0 +ℎ /𝑥0

▪𝑔′(𝑥0 ) = lim = lim = lim =
ℎ→0 ℎ ℎ→0 ℎ ℎ→0 ℎ

ln ℎ/𝑥0 +1 1 ln ℎ/𝑥0 +1 1
= lim = lim =
ℎ→0 ℎ 𝑥0 ℎ→0 ℎ/𝑥0 𝑥0

10
Derivatives using the definition
▪Example 3: w 𝑥 = 𝑥
1
▪ According to the rules of differentiation: 𝑤′ 𝑥 =
2 𝑥

Using the definition:

𝑤 𝑥 −𝑤 𝑥0 𝑥− 𝑥0 𝑥− 𝑥0
▪ 𝑤′ 𝑥0 = lim = lim = lim =
𝑥→𝑥0 𝑥−𝑥0 𝑥→𝑥0 𝑥−𝑥0 𝑥→𝑥0 𝑥− 𝑥0 𝑥+ 𝑥0
1 1
▪ = lim =
𝑥→𝑥0 𝑥+ 𝑥0 2 𝑥0

11
Derivatives using the definition
▪ In the previous examples we found the
derivative at point 𝑥0 using one limit only. Will
it always be like this?
▪ No! At points of the domain that are
problematic points (angular points, points
where we shift to another branch of the
function, … ) we must compute the right and
left derivative separately, using two different
limits.

12
One-sided derivatives using the
definition
▪ Left-hand derivative: Right-hand derivative:

𝑓(𝑥0. + ℎ) − 𝑓(𝑥0 ) 𝑓(𝑥0. + ℎ) − 𝑓(𝑥0 )

𝑓 ′ (𝑥0 − ) = lim− 𝑓 ′ (𝑥0 + ) = lim+
ℎ→0 ℎ ℎ→0 ℎ

▪ If both one-sided derivatives are equal and finite, the function is said to
be differentiable at 𝑥0 .

13
Differentiable functions
▪ If 𝑓 ′ (𝑥0 ) is finite, we say that 𝑓 is differentiable at 𝑥0 .

▪ If 𝑓 is differentiable at every point of a set 𝐷, we say that 𝑓 is

differentiable on 𝐷.

14
Differentiable functions
▪ If a function is differentiable at point 𝑥0 ,
then we can approximate the function, in a
neighbourhood of 𝑥0 , by the line tangent
to the function at that point 𝑥0 .
▪ The equation of the tangent line is

𝑦 = 𝑓 𝑥0 + 𝑓′ 𝑥0 𝑥 − 𝑥0

15
Differentiable functions
▪ Write the equation of the tangent lines to
the curve 𝑦 = 𝑓 𝑥 with expression
−𝑥 2 +2𝑥
𝑓 𝑥 = 𝑒
where 𝑦 = 1.

16
Differentiable functions in ℝ
▪ We say that 𝑓 is differentiable at 𝑥 = 𝑎 if the tangent line to the graph at
(𝑎, 𝑓(𝑎)), with equation
𝑇 𝑥 =𝑓 𝑎 +𝑚 𝑥−𝑎
approximates 𝑓 so well near 𝑥0 that we can write 𝑓(𝑥) as

𝑓 𝑥 =𝑇 𝑥 +𝐸 𝑥−𝑎 approximation
error
such that
𝐸 𝑥−𝑎 𝑓 𝑥 −𝑇 𝑥 𝑓 𝑥 − 𝑓 𝑎 +𝑚 𝑥−𝑎
lim = lim = lim =0
𝑥→𝑎 𝑥 − 𝑎 𝑥→𝑎 𝑥−𝑎 𝑥→𝑎 𝑥−𝑎

17
Differentiable functions
▪ The following functions are not differentiable at 𝑥 = 𝑎:

18
Exercise
▪ Let 𝑓: ℝ → ℝ be the function defined by:

𝑒 𝑥−1 𝑥≤1
𝑓 𝑥 =ቐ
1 + ln 𝑥 𝑥>1
Show that f is differentiable in ℝ, using the
previous definition of a differentiable
function.

19
Differentiability
In ℝ,
▪ 𝑓 is differentiable at 𝑥 = 𝑎 if and only if 𝑓′ 𝑎 exists and is finite.
▪ If 𝑓 is differentiable at 𝑥 = 𝑎, then 𝑓 is continuous at 𝑥 = 𝑎.

In ℝ𝑛 ,
▪ The existence of directional (partial) derivatives at 𝒙𝟎 does not imply that 𝑓
is differentiable at 𝒙𝟎 .
▪ The existence of directional (partial) derivatives at 𝒙𝟎 does not imply that 𝑓
is continuous at 𝒙𝟎 .

20
Derivative function and derivation rules
▪ The derivative function of a function 𝑓 is the function which gives
that derivative for each point of the domain of 𝑓 in which it exists.

21
Derivative function and derivation rules

22
Cauchy Rule
▪ If 𝑓 and 𝑔 are differentiable on an open interval ]𝛼, 𝛽[ and 𝒂 is
one of the end points of ]𝛼, 𝛽[ (either finite or infinite);
▪ if 𝑔′ 𝑥 ≠ 0 ∀𝑥 ∈ ]𝛼, 𝛽[ ;
▪ If 𝑓(𝑥) → 0 and 𝑔(𝑥) → 0, or 𝑓(𝑥) → ∞ and 𝑔(𝑥) → ∞, when
𝑥 → 𝑎 (with 𝑥 ≠ 𝑎).
𝑓(𝑥) 𝑓′ (𝑥)
▪ Then, lim = lim ′ (as long as the second limit exists).
𝑥→𝑎 𝑔(𝑥) 𝑥→𝑎 𝑔 (𝑥)

23
Cauchy Rule
∞ 1
ln 𝑥 𝐶𝑅 ∞ 1
▪ lim 5 = lim 𝑥 4 = lim =0
𝑥→+∞ 𝑥 𝑥→+∞ 5𝑥 𝑥→+∞ 5𝑥 5

0 1 0 1
𝑒 𝑥 −ln(1+𝑥)−1 𝐶𝑅 0 𝑒 𝑥 −1+𝑥 𝐿𝑅 0 𝑒 𝑥+
(1+𝑥)2
▪ lim = lim = =1
𝑥→0 𝑥2 𝑥→0 2𝑥 2 𝑥=0

24
Other types of indeterminate forms
∞ 1
0×∞ ln 𝑥 𝐶𝑅 ∞
▪ lim+ 𝑥 . ln 𝑥 = lim+ 1 = lim 𝑥
1 = lim+ −𝑥 = 0
𝑥→0 𝑥→0 𝑥
𝑥→0+ − 2 𝑥→0
𝑥

0 0
1 1 ∞−∞ sin 𝑥−𝑥 𝐶𝑅 0 cos 𝑥−1 𝐶𝑅 0
▪ lim − = lim = lim =
𝑥→0 𝑥 sin 𝑥 𝑥→0 𝑥.sin 𝑥 𝑥→0 sin 𝑥+𝑥 cos 𝑥

− sin 𝑥
= lim =0
𝑥→0 2 cos 𝑥−𝑥 sin 𝑥

25
Other types of indeterminate forms
0
𝑥 =
0
𝑥 lim ln sin 𝑥 𝑥 lim 𝑥 ln sin 𝑥
▪ lim+ sin 𝑥 lim+ 𝑒 ln sin 𝑥 =𝑒 𝑥→0+ =𝑒 𝑥→0+ =
𝑥→0 𝑥→0
cos 𝑥
ln sin 𝑥
lim 1 𝐶𝑅 ∞
∞ lim sin1𝑥
𝑥→0+ 𝑥→0+ − 2
=𝑒 𝑥 = 𝑒 𝑥 =

𝑥2 cos 𝑥 0 2𝑥 cos 𝑥−𝑥2 sin 𝑥

− lim sin 𝑥 𝑅𝐿 0 −
=𝑒 𝑥→0+ = 𝑒 cos 𝑥 𝑥=0 = 𝑒0 = 1

26
Rolle’s Theorem
▪ Let 𝑓 be a continuous function on
[𝑎, 𝑏] and differentiable in 𝑎, 𝑏 .
𝑓
▪ Let 𝑓(𝑎) = 𝑓(𝑏) = 𝑘.
𝑘

▪Then ∃𝑐 ∈ 𝑎, 𝑏 , such that 𝑓´(𝑐) = 0.

𝑎 𝑏 𝑥

27
Rolle’s Theorem

▪It is not enough that the function is 𝑦

continuous! It must be differentiable on

𝑓
𝑎, 𝑏 .
𝑘
▪If it only respects the continuity restriction,
we cannot guarantee that there is a point
in the interval where the derivative is zero. 𝑎 𝑏 𝑥

28
Rolle’s Theorem
▪ Corollary 1 of Rolle’s Theorem
Between two consecutive zeros of a differentiable function there is at
least one zero of the derivative.

▪ Corollary 2 of Rolle’s Theorem

▪ Between two consecutive zeros of the derivative there is, at most, one
zero of the function.

29
Exercise
▪ Let 𝑓 be a function defined on 𝑎, 𝑏 ,
where 𝑓 𝑎 = 𝑓 𝑏 = 0, that satisfies the
conditions of Rolle’s Theorem.
▪ Show that Rolle’s theorem may be used
on a function 𝑔 defined on 𝑎, 𝑏 with
expression 𝑔(𝑥) = 𝑒 −5𝑥 𝑓 𝑥 , and that
there is a point c such that 𝑓 ′ 𝑐 = 5𝑓 𝑐 .

30
Derivatives
1. Derivatives of real functions of a real variable.
2. Derivatives of real functions of a vector variable.
3. Derivatives of vector-valued functions.
4. Derivatives of higher order.
5. Differentiable functions: some properties.
6. The Chain Rule

31
Partial derivatives
▪ In the case of functions of more than one variable (vector fields
and scalar fields), the situation is not so simple:
▪ There is no point that can be called the derivative of 𝑓 at a point 𝒙𝟎 in ℝ𝑛
▪ There are infinitely many directional derivatives of 𝑓 at 𝒙𝟎 .
▪ The existence of directional derivatives at 𝒙𝟎 does not imply that 𝑓 is
differentiable at 𝒙𝟎 .
▪ The existence of directional derivatives at 𝒙𝟎 does not imply that 𝑓 is
continuous at 𝒙𝟎 .

32
Partial derivatives: definition
▪ The partial derivative with respect to variable 𝑥𝑖 of a function 𝑓
from ℝ𝑛 to ℝ𝑚 at point 𝒂 is the directional derivative along
standard vector 𝑒𝑖 of 𝑓 at 𝒂.

33
Partial derivatives: definition

https://www.geogebra.org/m/tZgrSxQ4#material/EWMQ8qnr

34
Partial derivatives: notation

35
Partial derivatives
▪ Let 𝑓 𝑥, 𝑦 = 𝑥 2 𝑦 ▪ This swirly-d symbol, ∂, called "del", is
used to distinguish partial derivatives
from ordinary single-variable
derivatives.
▪ We want to see how the function
changes as we let just one of those
variables change while holding all the
others constant.

36
Partial derivatives: example
▪ Let 𝑓: ℝ𝑛 → ℝ𝑚 , 𝑓 𝑥, 𝑦 = 𝑥 2 + 𝑦 2 .
Find the partial derivatives of this function,
at the input point 2,1 .

37
Gradient
▪The gradient of a function 𝑓, denoted as ∇𝑓, is the collection of all its
partial derivatives into a vector.

𝜕𝑓 𝜕𝑓 𝜕𝑓
∇𝑓 𝒂 = 𝒂 𝒂 … 𝒂
𝜕𝑥1 𝜕𝑥2 𝜕𝑥𝑛
▪ ∇𝑓 is a vector-valued function, specifically one with an n-dimensional
input and an 𝑛-dimensional output.
▪ The gradient of 𝑓, if evaluated at an input 𝒂 points in the direction of
steepest ascent.

38
Directional derivatives
▪ The directional derivative along vector 𝑣 of a function 𝑓 from ℝ𝑛 to ℝ𝑚
at a point 𝑎 is the limit, if it is finite, of the average rate of change of 𝑓 at
𝑎 in the direction of 𝑣, when the number of times 𝑣 is use is becoming
small.
′
𝜕𝑓 𝑓 𝒂 + 𝑡𝒗 − 𝑓 𝒂
𝑓𝑣 𝒂 = 𝒂 = lim
𝜕𝑣 𝑡→0 𝑡
▪ The directional derivative along a vector 𝒗 at a point 𝒂 is the rate of
growth of the function when the arguments move away from 𝒂 in the
direction of 𝒗.

39
Directional derivatives
Let 𝑓: ℝ3 → ℝ1 be a differentiable function.
Then the directional derivative along the vector v is computed as:

40
Directional derivatives: example
▪ Let 𝑓: ℝ𝑛 → ℝ𝑚 , 𝑓 𝑥, 𝑦 = 𝑥 2 + 𝑦 2 .
Find the directional derivative of 𝑓 at the
point 2,1 along the vector 1,1 .

41
Normalized Directional Derivative
▪ The normalized directional derivative along non – null vector 𝑣 of a
function 𝑓 from ℝ𝑛 → ℝ𝑚 at a point 𝑎 is the directional derivative along
𝑣
of 𝑓 at 𝑎.
𝑣
𝑓𝑣ො′ 𝑎 = 𝑓 ′𝑣 𝑎
𝑣

▪ The directional derivative depends not only on the direction of the

vector it is associated to, but also on its norm. To compare directional
derivatives, it is useful that they are associated to vectors with the same
norm.

42
 Tangent plane to the surface
▪ Tangent plane to the surface 𝑧 = 𝑓 𝑥, 𝑦
at the point 𝑥0 , 𝑦0 , 𝑓 𝑥0 , 𝑦0

https://www.geogebra.org/m/tZgrSxQ4#material/EWMQ8qnr

𝜕𝑓 𝜕𝑓
𝑧 = 𝑓 𝑎, 𝑏 + 𝑎, 𝑏 𝑥 − 𝑎 + 𝑎, 𝑏 𝑦 − 𝑎
𝜕𝑥 𝜕𝑦

43
 2
Differentiability in ℝ
▪ A function 𝑓 is differentiable at 𝑎, 𝑏 if in a neighbourhood of
point 𝑎, 𝑏, 𝑓 𝑎, 𝑏 we can say that:

𝜕𝑓 𝜕𝑓
𝑓 𝑥, 𝑦 = 𝑓 𝑎, 𝑏 + 𝑎, 𝑏 𝑥 − 𝑎 + 𝑎, 𝑏 𝑦 − 𝑏 + E 𝑥 − 𝑎, 𝑦 − 𝑏
𝜕𝑥 𝜕𝑦
such that:
𝐸 𝑥 − 𝑎, 𝑦 − 𝑏
lim =0
𝑥,𝑦 → 𝑎,𝑏 𝑥 − 𝑎, 𝑦 − 𝑏

44
Differentiability
2
in ℝ
THE PARTIAL DERIVATIVES MAY
BE DIFFERENT, AND STILL THE
FUNCTION IS DIFFERENTIABLE!

45
Example
▪ Let 𝑓 𝑥, 𝑦 = 𝑥 2 + 𝑦 2

Show that 𝑓 is differentiable at (0,0).

46
Example
𝑦 3 −2𝑥 2 𝑦 2
, if 𝑥, 𝑦 ≠ 0,0
𝑥 2 +𝑦 2
▪ Let 𝑔 𝑥, 𝑦 =
0 , if 𝑥, 𝑦 = 0,0

Show that 𝑔 is not differentiable at (0,0).

47
Results
▪ If a function has no partial derivatives at 𝒂, then the function is not
differentiable at 𝒂.
▪ If a function is differentiable at 𝒂, then the partial derivatives at
𝒂 exist.
▪ If the partial derivatives at 𝒂 exist, the function may be
differentiable or not.
▪ If f is differentiable at 𝒂, then f is continuous at 𝒂.

48
Derivatives
1. Derivatives of real functions of a real variable.
2. Derivatives of real functions of a vector variable.
3. Derivatives of vector-valued functions.
4. Derivatives of higher order.
5. Differentiable functions: some properties.
6. The Chain Rule

49
Derivative of vector-valued function
▪ Let 𝑟 𝑡 = 𝑥 𝑡 , 𝑦 𝑡 describe the movement of a particle,
according to the parametric functions 𝑥 𝑡 = t , y 𝑡 = 𝑡 2 .
▪ Then, 𝑟 ∶ ℝ → ℝ2 with expression 𝑟 𝑡 = 𝑡 , 𝑡 2 .
▪ To take the derivative of a vector-valued function, take the
derivative of each component.

50
Derivative of vector-valued function
▪ If you interpret the initial
function as giving the position
of a particle as a function of
time, the derivative gives the
velocity vector of that particle
as a function of time.

51
Jacobian Matrix: definition
▪ The jacobian matrix of a function 𝑓 from ℝ𝑛 to ℝ𝑚 at 𝑎 is the
matrix in which each row contains the partial derivatives, at 𝑎, of a
component function of 𝑓, and each column contains the partial
derivatives with respect to a variable, at 𝑎, of 𝑓.

52
Jacobian Matrix
▪ Find the Jacobian matrix of the function

𝑓 ∶ ℝ2 → ℝ3
𝑥 𝑦
𝑓 𝑥, 𝑦 = 2 2
, 2 2
, 2𝑥𝑦
𝑦 +𝑥 𝑦 +𝑥

evaluated at the point 1,1 .

53
Derivatives
1. Derivatives of real functions of a real variable.
2. Derivatives of real functions of a vector variable.
3. Derivatives of vector-valued functions.
4. Derivatives of higher order.
5. Differentiable functions: some properties.
6. The Chain Rule

54
First Order Derivative
▪ Allows us to study the function’s monotonicity and extreme values.

▪ If the first derivative is positive, then the function is strictly increasing

▪ If the first derivative is negative, then the function is strictly decreasing
▪ If the first derivative is zero (𝑓 is stationary at 𝑥0 , a critical point) and
changes sign at a certain point, there is an extreme value:
Minimum if it goes from negative to positive
Maximum if it goes from positive to negative

55
Minimum and Maximum values

http://tutorial.math.lamar.edu/Classes/CalcI/MinMaxValues.aspx
56
Second Order Derivative
▪Allows us to study the shape and the points of inflection of a function.
▪ If the second order derivative is positive, then the function is convex
(sometimes said concave up).
▪ If the second order derivative is negative, then the function is concave
(sometimes said concave down).
▪ If the second order derivative is zero and changes sign at a certain point,
there is an inflection point. At this point, the function changes shape
from concave to convex or the other way around.

57
Higher Order Partial Derivatives
▪ By capturing all the second-derivative
information of a multivariable function, the
Hessian matrix often plays a role analogous
to the ordinary second derivative in single
variable calculus.
▪ The second partial derivative test, which
helps you find the maximum/minimum of a
multivariable function.

58
The Hessian Matrix

59
Exercise
▪ Compute the Hessian of

𝑓 𝑥, 𝑦 = 𝑥 3 − 2𝑥𝑦 − 𝑦 6

at the point 1 2 :

60
Derivatives
1. Derivatives of real functions of a real variable.
2. Derivatives of real functions of a vector variable.
3. Derivatives of vector-valued functions.
4. Derivatives of higher order.
5. Differentiable functions: some properties.
6. The Chain Rule

61
Differentiable functions : Properties
▪ A vector-valued function 𝒇 = 𝑓1 , 𝑓2 , … , 𝑓𝑚 is differentiable at
𝒂 = 𝑎10 , 𝑎20 , … , 𝑎𝑛0
if its component functions 𝑓1 , 𝑓2 , … , 𝑓𝑚 are differentiable at 𝒂.

𝜕𝑓1 𝜕𝑓1 𝜕𝑓1

𝒂 𝒂 … 𝒂
𝑓1 𝒂 + 𝒖 𝑓1 𝒂 𝜕𝑥1 𝜕𝑥2 𝜕𝑥𝑛 𝑢1 𝑜1 𝜌
𝜕𝑓2 𝜕𝑓2 𝜕𝑓2
𝑓2 𝒂 + 𝒖 𝑓 𝒂 𝒂 𝒂 … 𝒂 𝑢2 𝑜2 𝜌
= 2 + 𝜕𝑥1 𝜕𝑥2 𝜕𝑥𝑛
⋮ +
⋮ ⋮ ⋮ ⋮ ⋮ ⋮
𝑓𝑚 𝒂 + 𝒖 𝑓𝑚 𝒂 𝜕𝑓𝑚 𝜕𝑓𝑚 𝜕𝑓𝑚 𝑢𝑛 𝑜𝑚 𝜌
𝒂 𝒂 … 𝒂
𝜕𝑥1 𝜕𝑥2 𝜕𝑥𝑛

62
Exercise
▪ Study, regarding differentiability, the
following function at 𝑡 = 0:

1
−𝑡 , 𝑡 2 sin 𝑖𝑓 𝑡 < 0
𝑡
𝑓 𝑡 =
𝑡2 − 𝑡 𝑒𝑡, 0 𝑖𝑓 𝑡 ≥ 0

63
Differentiable functions : Properties
▪ If 𝑓𝑥′1 , 𝑓𝑥′2 , … , 𝑓𝑥′𝑛 exist on a neighbourhood of 𝒂 and are
continuous at 𝒂, then 𝑓 is differentiable at 𝒂. It is actually enough
that 𝑛 − 1 partial derivatives exist and are continuous at 𝒂.
▪ Notice that this is a sufficient condition for differentiability, but not
a necessary one!

64
Exercise
▪ It is known of a function 𝑓 that

and that

Is 𝑓 differentiable at 0,0 ?

65
Continuously differentiable function
▪ We say that 𝑓 is continuously differentiable on a subset 𝑆 of ℝ𝑛 if
𝑆 is contained in an open set on which 𝑓𝑥′1 , 𝑓𝑥′2 , … , 𝑓𝑥′𝑛 are
continuous. We call this a 𝐶 1 function.
▪ The previous result implies that such a function is differentiable at
each 𝒂 in 𝑆.
▪ Result: every 𝐶 1 function at a point is differentiable at that point.

66
𝐶 𝑘 Function: definition
▪ A function from ℝ𝑛 to ℝ𝑚 is 𝐶 𝑘 at a point 𝑎 if all its partial
derivatives of order lower or equal to 𝑘 are continuous at 𝑎.

67
Second Order Derivatives
Schwarz – Young Theorem
▪ Let 𝑓: 𝐷 ⊆ ℝ𝑛 → ℝ. The partial derivatives 𝑓𝑥′′𝑖 𝑥𝑗 = 𝑓𝑥′′𝑗 𝑥𝑖 coincide
at every point in D where 𝑓𝑥′𝑖 and 𝑓𝑥′𝑗 are differentiable.

◦ Corollary 1:
If 𝑓: 𝐷 ⊆ ℝ𝑛 → ℝ is a 𝐶 2 function, then the partial derivatives 𝑓𝑥′′𝑖 𝑥𝑗 and 𝑓𝑥′′𝑗 𝑥𝑖
coincide.
◦ Corollary 2:
If 𝑓: 𝐷 ⊆ ℝ𝑛 → ℝ is a 𝐶 𝑘 function, then the order of derivation up to order k
is indifferent.

68
Example
▪ Check the validity of the Schwarz – Young
Theorem when applied to the function
with expression:

𝑓 𝑥, 𝑦 = 3𝑥 2 𝑦 3 + 𝑥𝑦

69
Differentiable functions again
▪ Let 𝑓: 𝐷 ⊆ ℝ𝑛 → ℝ be a differentiable function at 𝒂.
Then the directional derivative of 𝑓 at 𝒂 in the direction of 𝒖, 𝑓𝑢′ 𝑎
is equal to:

Remark: the above equality does not ensure that 𝑓 is differentiable

at point a (the converse of the proposition is not necessarily true).

70
Exercises
▪ Let 𝑓: 𝐷 ⊆ ℝ2 → ℝ be a function such that
∇𝑓 0,0 = 0 0 and 𝑓 ′1,2 0,0 = 1.
Can this function be differentiable at (0,0)?

71
Relative Minimums And Maximums
▪ The point (𝑎, 𝑏) is a critical point (or a stationary point) of 𝑓(𝑥, 𝑦)
provided one of the following is true:
1. ∇𝑓 0,0 = 0 0 (this is equivalent to saying that 𝑓𝑥′ 𝑎, 𝑏 = 0
and 𝑓𝑦′ 𝑎, 𝑏 = 0.
2. 𝑓𝑥′ 𝑎, 𝑏 and/or 𝑓𝑦′ 𝑎, 𝑏 do not exist.

72
Fact on Relative Minima And Maxima
▪ If the point (𝑎, 𝑏) is a relative extrema of the function 𝑓(𝑥, 𝑦) and
the first order derivatives of 𝑓(𝑥, 𝑦) exist at (𝑎, 𝑏) then (𝑎, 𝑏) is also
a critical point of 𝑓(𝑥, 𝑦) and in fact we’ll have ∇𝑓 0,0 = 0 0 .

73
Exercise
▪ Find all the critical points of

𝑓(𝑥, 𝑦) = 4 + 𝑥 3 + 𝑦 3 − 3𝑥𝑦

▪ What can you conclude about the critical

points you found?

74
Derivatives
1. Derivatives of real functions of a real variable.
2. Derivatives of real functions of a vector variable.
3. Derivatives of vector-valued functions.
4. Derivatives of higher order.
5. Differentiable functions: some properties.
6. The Chain Rule

75
The chain rule in real functions
▪ Suppose that 𝑔 is differentiable at 𝑥0 and 𝑓 is differentiable at
𝑔 𝑥0 . Then the composite function ℎ = 𝑓 ∘ 𝑔, defined by
ℎ 𝑥 =𝑓 𝑔 𝑥
is differentiable at 𝑥0 , with
ℎ′ 𝑥0 = 𝑓′ 𝑔 𝑥0 𝑔′ 𝑥0

76
Exercise
2𝑥+5
▪ If 𝑝 𝑥 = 𝑥 and 𝑞 𝑥 = , calculate:
𝑥−1

a) 𝑝 ∘ 𝑞 ′ 2
b) 𝑞 ∘ 𝑝 ′ 𝑥

77
The Chain Rule – general case

▪ For the functions 𝑓 and 𝑔, with expressions

defined below, compute the derivatives of
𝑓 ∘ 𝑔 and 𝑔 ∘ 𝑓, in case they are compatible:

𝑓 𝑥, 𝑦 = 2𝑥 − 𝑦 2 𝑔 𝑡 = (𝑡, 2𝑡)

78
The Chain Rule – general case

▪ For the functions 𝑓 and 𝑔, with expressions

defined below, compute the derivatives of
𝑓 ∘ 𝑔 and 𝑔 ∘ 𝑓, in case they are compatible:
𝑓 𝑥, 𝑦 = 𝑥 2 + 3𝑦
𝑔 𝑠, 𝑡 = (𝑠 + 2𝑡, 𝑠 2 − sin 𝑡)

79
The Chain Rule – general case

▪ For the functions 𝑓 and 𝑔, with expressions

defined below, compute the derivatives of
𝑓 ∘ 𝑔 and 𝑔 ∘ 𝑓, in case they are compatible:
𝑓 𝑥, 𝑦 = 2𝑥 + 𝑦 2
𝑔 𝑡 = (𝑡 , 𝑡 2 , 𝑡 3 )

80
The Chain Rule – general case

3 𝑧 𝑦
▪ Let 𝑤 𝑥, 𝑦, 𝑧 = 𝑥 𝑓 , . Calculate:
𝑥 𝑥

𝜕𝑤 𝜕2 𝑤
and
𝜕𝑥 𝜕𝑥𝜕𝑦

81