Second derivative test

Second-derivative test for relative extremum. If f 0 (x0 ) = 0, then f (x0 ) will

be
• A relative maximum if f 00 (x0 ) < 0.
• A relative minimum if f 00 (x0 ) > 0.

Figure: Conditions for a Relative Extremum

N. A. Dao Derivatives October 28, 2024 14 / 30

Ex =
f(x) = n + 1 , (0 .
2; 4]
U

Max , mis f(r)

j'(x) 1 n- 1 in (0 ,2 : 4)
-
= + =
= 0

n2

=
1 =
1 = 0 n = 1
ra

Step 2 : We have f(1) = 2

, g(0 2) .
=
5 2 .

: (4) =
, 25
4+ 0 = 4 25
.

Thus and
, maxf() =
5 .
2
; minfln-
[0 .
2 , 4]
Chain rule

d
f (g (x )) = f 0 (g (x ))g 0 (x ).
dx

Example 16
d
Find dx sin x 3 .

N. A. Dao Derivatives October 28, 2024 15 / 30

Derivatives of Implicit functions
dy
Let y satisfy y 3 + xy (x ) + x 2 x = 2024. Find dx (0).

Proof.
Observe that y (0) = 20241/3 . Di↵erentiating both sides of the equation
yields
d 3
y + xy (x ) + x 2 x = 0
dx
3y 2 (x )y 0 (x ) + y (x ) + xy 0 (x ) + 2x 1 = 0.

At x = 0, we have

1 y (0) 1 20241/3
3y 2 (0)y 0 (0) + y (0) 1 = 0 () y 0 (0) = = .
3y 2 (0) 3 ⇥ 20242/3

N. A. Dao Derivatives October 28, 2024 16 / 30

Derivatives of Inverse functions

Figure: Inverse function

df 1 1
(y ) = 0 , y = f (x ).
dx f (x )
Apply this formula to find arcsin0 (x ), arccos0 (x ), arctan0 (x ), . . .

N. A. Dao Derivatives October 28, 2024 17 / 30

Derivatives of functions of multi-variables
Definition 17
Let D be an open set in R2 , and let f (x, y ) : D ! R. Then, its partial
derivatives of f (x, y ) are the functions fx , fy defined by

f (a + h, b ) f (a, b )
fx (a, b ) = lim ,
h !0 h
f (a, b + h ) f (a, b )
fy (a, b ) = lim ,
h !0 h
for (a, b ) 2 D .

We denote rf (or Df ) by the first derivative of f , and

rf (a, b ) = (fx (a, b ), fy (a, b )) .
Notation.
∂f ∂f
• ∂x = Dx f = f x ; ∂y = Dy f = f y .
N. A. Dao Derivatives October 28, 2024 18 / 30
f(x : y) x Luy
yz
f(k z) nyz
= + +
y = =
:

fu(x y) : = 2n +
2y DJ (vo ; yo : 20)
=
(8 ,
(Noiyo ; 20) ,
Sy (no : yo ; 20) fz(rigoo
,

by (x , y) =
2y + 20 8x(10 :%: 20) =
y0z0 Jy) ) = Kozo faL ) =

by

(Ju(a b) (a b) Pflvo : yo : zo) (Ho20 Koz0 Hoyo (

Dflaib) by
=
>
- , ,
=
, ,
,

(2a + 1b (b)

-is
= ,
fu +

Example 18
Let f (x, y ) = x 2 y 3 + sin xy . Find Df (0, 1), Df (1, 1).
On (wiy) = 2rys +
yes my ,

by (r, y) =
n 3y + n .

Sky

/Guillcost >
-
Df(01) = (1 = 0)

031)
It i +c
=

>
-
Df(1 : 1) = (C + 120 ,
3 +

N. A. Dao Derivatives October 28, 2024 19 / 30

Linear approximation

Let D be an open set in R2 , and let f : D ! R be di↵erentiable at

(x0 , y0 ) 2 D . Then, we have

f (x0 + h, y0 + k ) ⇠ f (x0 , y0 ) + (h, k ) · Df (x0 , y0 )

= f (x0 , y0 ) + hfx (x0 , y0 ) + kfy (x0 , y0 ),

where h, k are small enough.

Example 19
Let f (x, y ) = sin(x 2 + y 2 ). Compute Df (0, 0), and estimate
f (0.01, 0.01).

N. A. Dao Derivatives October 28, 2024 20 / 30

 f(x + h
, yo
+ k) =
f(viyd + h k .

Df( ; ya
(fulvoiyo) Jy(m ; yo)) ,

= fo ; yo) + h .

fr (vi yo) +
k .

Jy (1 ; yo)
One-variable :
f( + h) = ((r) + h .

f(w)

#x 19 :
fluiy) =
Sin (n + y)
In (x : y) =
ancos(x + y4)
by (xi y) =
Ly .
cos(x2 yz)
+

D8(0 : 0) =
(fu(0 : 03 ,
fy(0 : 01) = 10 + 0
Butuo =

yu =
0 ,
h = 0 , 01 ,
k = -

0 , 01

Then it
follows from the Lineal Appro
that 8(0 01 -0 , 01) =
110 , 0) + , 018n
0 10 : 0) 0 . 01
fy (0 : 0)
-

,
·

= O

& (1 ,
01, -

0. 99)
+
10 =
+
= yo
= 1 ,
h = 0 , 01 ,
k = + 0 01
, .

f (1 , 01,
-
0 .
99) =
f( =
-

1) + 0 01 .
·

f (li -
1) + 0 ,
01 .

Jy(1 -
1)
 One-variable Multi-variables

y =
f(a) z =
f(r y),

dy =
f'(x) .
dr dz =
Df(x y) (du dy)
, ,

=
(fr(niy) by (n : y))( dy
, ,

= fudue Jydy .
+

f =
f(no = yo)
=

dy(u ; yo) =
f(no + bu , yo + xy) -

f(wiy))
Thi nho Linear Approximatin iv ro
Di↵erential

We denote the di↵erential of function f by df . Then, for (x0 , y0 ) 2 D we

have

df (x0 , y0 ) = fx (x0 , y0 )dx + fy (x0 , y0 )dy

⇡ f (x0 + Dx, y0 + Dy ) f ( x0 , y 0 ) .

Example 20
Let f (x, y ) = exp{x 2 + y 2 }. Find df (0, 0).

N. A. Dao Derivatives October 28, 2024 21 / 30

Exco :
f (wiy) = en + (exponential
y
ex + y-
2(n+ yz) Grea +

/fa(k
: y) =
=

Ok

Ify (u y) , =

<y .
e
dg(0 : 0) =
(n( : 0) ex + fy(0 , 0) dy = 0

dg(11) =
In (1) de +
fy(1 :
1)
dy
- Ledn + Ledy .
Marginal and Elastic

1. Marginal
Let f (x, y ) be a function. Then, we denote the marginal function f w.r.t
x, y alternatively by

Mfx (x0 , y0 ) := fx (x0 , y0 ) ⇡ f (x0 + 1, y0 ) f ( x0 , y0 ) ,

and
Mfy (x0 , y0 ) := fy (x0 , y0 ) ⇡ f (x0 , y0 + 1) f ( x0 , y0 ) .
2. Elastic 8
fx (x0 ,y0 )
>
<Ef ,x (x0 , y0 ) = f (x0 ,y0 ) ,
x0
> fy (x0 ,y0 )
:Ef ,y (x0 , y0 ) = f (x0 ,y0 ) .
y0

N. A. Dao Derivatives October 28, 2024 22 / 30

Example 21 (The Cobb–Douglas Production Function)
Cobb and
Douglas used the equation P (L, K ) = 1.01L0.75 K 0.25 to
model the American economy from 1899 to 1922, where
L is the amount of labor and K is the amount of capital.
(a) Calculate PL and PK .
(b) Find the marginal productivity of labor
and the marginal productivity of capital in the year 1920,
when L = 194 and K = 407 (compared with the assigned
values L = K = 100 in 1899). Interpret the results.
(c) In the year 1920
which would have benefited production more, an increase
in capital investment or an increase in spending on labor?

N. A. Dao Derivatives October 28, 2024 23 / 30

 751-025
25
& $1 1 01 0 10 ,

=
, , ,
·

-
0 75

(075
,

k = 1 01.
. .
0 . 25 .
1

& Dy (ega 402)

,

2) Whin rao
Marginal Theo Captio
Example 22
Let demand function Q = 100 P + 0.1Y , where P is the price of a
commodity, and Y is the income.
Find the elasticity of Q w.r.t. Y at P = 10, Y = 1000. If Y increases 5%,
-

then how much will Q change?

w r t
. .
: with respect to

N. A. Dao Derivatives October 28, 2024 24 / 30

2 Q = 100 -

P + 0, Y 1 : Price Y income
:

# Qy (10 ,
1000) =
Qy (10 , 1000) = 0
,1

Q(0 ,
1000) 198

1000 1000

= 10 &

Q =
100 =

P + 01
. . Y
& (10 , 1050 .
) = 100 -
10 + 0 11 .
1050 =
Second order derivatives

Let f (x, y ) be a function. Then, we have the second partial derivatives of

f at (a, b ) 2 D as follows:

∂2 f ∂2 f ∂2 f ∂2 f
( a, b ) , ( a, b ) , ( a, b ) , (a, b ).
∂x 2 ∂x ∂y ∂y ∂x ∂y 2
And,
!
∂2 f ∂2 f
∂x 2
(a, b ) ∂x ∂y (a, b )
D 2 f (a, b ) = D (Df )(a, b ) = ∂2 f ∂2 f (Hessian matrix)
∂y ∂x (a, b ) ∂y 2
(a, b )

Notation. ⇣ ⌘ ⇣ ⌘
∂2 f ∂ ∂f ∂2 f ∂ ∂f
• ∂x 2
= ∂x = fxx ; ∂y
∂x 2 = ∂y ∂y = fyy .
⇣ ⌘ ⇣ ⌘
∂2 f ∂ ∂f ∂2 f ∂ ∂f
• ∂x ∂y = ∂x ∂y = f yx ; ∂y ∂x = ∂y ∂x = fxy .

N. A. Dao Derivatives October 28, 2024 25 / 30

Example 23
Let f (x, y ) = ln(x 2 + y 2 ). Find Df (1, 1), and D 2 f (1, 1).
Moreover, show that

fxx (x, y ) + fyy (x, y ) = 0, 8(x, y ) 2 R2 \ {(0, 0)}.

Theorem 24 (Clairaut’s Theorem)

Suppose fxy and fyx are continuous on D (a domain in R2 ). Then

fxy (a, b ) = fyx (a, b ), 8(a, b ) 2 D .

Example 25
Whether or not there is a function f whose partial derivatives are
fx (x, y ) = x + 4y and fy (x, y ) = 3x y ?

N. A. Dao Derivatives October 28, 2024 26 / 30

Maximum and minimum values
Theorem 26 (Second derivatives test)
Suppose the second partial derivatives of f (x, y ) are continuous on a
domain D ⇢ R2 , and suppose that fx (a, b ) = fy (a, b ) = 0, (a, b ) 2 D
(that is (a, b ) is a critical point of f ). Let

D (a, b ) = fxx (a, b ) ⇥ fyy (a, b ) |fxy (a, b )|2 .

(a) If D (a, b ) > 0 and fxx (a, b ) > 0, then f (a, b ) is a local minimum.
(b) If D (a, b ) > 0 and fxx (a, b ) < 0, then f (a, b ) is a local maximum.
(c) If D (a, b ) < 0, then f (a, b ) is not a local maximum or minimum. In
this case, (a, b ) is a saddle point of f .

Remark 2.1
If D (a, b ) = 0, the test gives no information: f could have a local
maximum or local minimum at (a, b ), or (a, b ) could be a saddle point of
f.
N. A. Dao Derivatives October 28, 2024 27 / 30
Example 27
Find the local maximum and minimum values and saddle point(s) of the
function.
a) f (x, y ) = xy + x1 + y1 .
b) f (x, y ) = (x y )(1 xy ).

N. A. Dao Derivatives October 28, 2024 28 / 30

Lagrange Multipliers

Problem. Find the maximum and minimum values of f (x, y ) subject to

the constraint g (x, y ) = c.
Strategy
(a) Assume that rg (x, y ) 6= ~0 on the constraint g (x, y ) = c. Find all
values of x, y , l such that

rf (x, y ) = lrg (x, y )

and use the constraint.

(b) Evaluate f at all the points (x, y ) that result from step (a). The
largest of these values is the maximum value of f ; the smallest is the
minimum value of f .
Remark 2.2
The method still works for functions of multi-variables.

N. A. Dao Derivatives October 28, 2024 29 / 30

Example 28
Find the extreme values of the function f (x, y ) = 2x 2 + y 2 on the circle
x 2 + y 2 = 1.

Example 29
Find the extreme values of the function f (x, y ) = 2x 2 + y 2 on the circle
x 2 + y 2  1.

Example 30
A rectangular box without a lid is to be made from 1m2 of cardboard.
Find the maximum volume of such a box.

N. A. Dao Derivatives October 28, 2024 30 / 30