University of Toronto

Scarborough

MATA33H3
Calculus for Management II
Winter 2018
Final Exam

Prof: Moore, E.

Exam Guide
Table of Contents:
2.8 Introduction to Functions of Several Variables
17.1 − 17.2 Partial Derivatives
17.2 - 17.3 Implicit Partial Differentiation
17.4 Higher-Order Derivatives
17.5 Chain Rule
17.6 Extrema for Functions of Two Variables
17.7 Langrange Multipliers
17.9 Multiple Integration
Review

2.8 Introduction to Functions of Several Variables

Functions → A function 𝑓𝑓 = a rule that assigns to each element 𝑥𝑥 in a certain set 𝐴𝐴, exactly one
element, called 𝑓𝑓(𝑥𝑥), in some other set 𝐵𝐵. The element 𝑓𝑓(𝑥𝑥) is called the value of 𝑓𝑓 at 𝑥𝑥

Domain and Range → Set 𝐴𝐴 and Set 𝐵𝐵 are sets of real numbers

𝐴𝐴: Domain of the function

𝐵𝐵: range of 𝑓𝑓 = set of all possible values of 𝑓𝑓(𝑥𝑥) as 𝑥𝑥 varies in domain.

Independent Variable: → Arbitrary number in domain of a function 𝑓𝑓

Dependent Variable → A number in range of 𝑓𝑓

Machine diagram for a function 𝒇𝒇

• Function set as machine 𝑥𝑥 is the domain of function𝑓𝑓; then when 𝑥𝑥 enters the machine,
accepted as input and machine, it produces an output 𝑓𝑓(𝑥𝑥) according to the rule of
function.

Example 1:

𝑓𝑓(𝑥𝑥1 , 𝑥𝑥2 , 𝑥𝑥3 , 𝑥𝑥4 )

There are four input variables in the above function

Input: the ordered pairs

𝑥𝑥1 𝑥𝑥2
= det �𝑥𝑥 𝑥𝑥4 � = 𝑥𝑥1 𝑥𝑥4 − 𝑥𝑥3 𝑥𝑥2 ⇒ ℝ
3
 1 0
𝑓𝑓(1,0,0,1) = det � � = (1 ⋅ 1) − (0 ⋅ 0) = 1
0 1
Example 2:

Let

𝐺𝐺(𝑥𝑥, 𝑦𝑦) = −4𝑥𝑥 + 7𝑦𝑦

𝐺𝐺(5,4) = −4(5) + 7(4) = −20 + 28 = 8

𝐺𝐺 has a real output plus ordered pairs of input

Example 3:

𝑍𝑍 = ℎ(𝑥𝑥, 𝑦𝑦) = �𝑥𝑥 + 𝑦𝑦

We see that 𝐷𝐷 is the subset of ℝ2 = ℝ + ℝ(𝑥𝑥, 𝑦𝑦 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝)

Domain of ℎ(𝑥𝑥) = {(𝑥𝑥, 𝑦𝑦)|𝑥𝑥 ∈ ℝ 𝑎𝑎𝑎𝑎𝑎𝑎 𝑑𝑑𝑑𝑑𝑓𝑓𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖}

𝑥𝑥 + 𝑦𝑦 ≥ 0, 𝑦𝑦 ≥ −𝑥𝑥

Domain

(−1,0) (1,0)

y = −x

Point (−1,0) and Domain

√−1=undefined

𝐷𝐷 = {(𝑥𝑥, 𝑦𝑦)|𝑥𝑥, 𝑦𝑦 ∈ ℝ, 𝑦𝑦 2 − 𝑥𝑥}

Therefore
 Point (1,0) → ℎ(1,0) = √1 + 0 = √1 = 1

Example 4:

Find the domain of the function 𝑓𝑓(𝑥𝑥, 𝑦𝑦) = �9 − 𝑥𝑥 2 − 𝑦𝑦 2

𝑓𝑓 is defined as (𝑥𝑥, 𝑦𝑦) ↔ 9 − 𝑥𝑥 2 − 𝑦𝑦 2 ≥ 0 ↔ 𝑥𝑥 2 + 𝑦𝑦 2 ≤ 9

𝑓𝑓(4,0) = undefined, 𝑓𝑓(1,1) = √7

𝐷𝐷 = {(𝑥𝑥, 𝑦𝑦)|𝑥𝑥, 𝑦𝑦 ∈ ℝ, 𝑥𝑥 2 + 𝑦𝑦 2 ≤ 9}

Machine diagram concept

• Inputs are same ordered pairs

• Output is a real number

Rough sketch of 𝑫𝑫

In the circle or on borders, the domain is defined

(0,3)

Domain
(3,0)

(−3,0)

(0, −3)
 17.1 – 17.2 Partial Derivatives

𝑓𝑓(𝑎𝑎+ℎ)−𝑓𝑓(𝑎𝑎)
1) Slope of tangent 𝑚𝑚 = lim ℎ
ℎ→0
𝑓𝑓(𝑎𝑎+ℎ)−𝑓𝑓(𝑎𝑎)
2) Velocity 𝑣𝑣(𝑎𝑎) = lim ℎ
𝑛𝑛→0
′ (𝑎𝑎) 𝑓𝑓(𝑎𝑎+ℎ)−𝑓𝑓(𝑎𝑎)
3) Rate of ∆ 𝑓𝑓 = lim ℎ
if this limit exists
ℎ→0

* Definition (Partial derivatives of 𝑓𝑓 𝑎𝑎𝑎𝑎 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 (𝑎𝑎, 𝑏𝑏))

𝑓𝑓(𝑥𝑥 + ℎ, 𝑦𝑦) − 𝑓𝑓(𝑥𝑥, 𝑦𝑦)

𝑓𝑓𝑥𝑥 (𝑥𝑥, 𝑦𝑦) = lim
ℎ→0 ℎ
𝑓𝑓(𝑥𝑥, 𝑦𝑦+ℎ)−𝑓𝑓(𝑥𝑥,𝑦𝑦)
𝑓𝑓𝑦𝑦 (𝑥𝑥, 𝑦𝑦) = lim ℎ
ℎ−0

𝐸𝐸𝐸𝐸. 1: 𝑧𝑧 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) we write :

𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
𝑓𝑓𝑥𝑥 (𝑥𝑥, 𝑦𝑦) = 𝑓𝑓𝑥𝑥 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) = = 𝑓𝑓1 = 𝐷𝐷1 𝑓𝑓 = 𝐷𝐷𝐷𝐷𝐷𝐷
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑 𝑑𝑑 𝑑𝑑𝑑𝑑
𝑓𝑓𝑦𝑦 (𝑥𝑥, 𝑦𝑦) = 𝑓𝑓𝑦𝑦 = = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) = = 𝑓𝑓2 = 𝐷𝐷2 𝑓𝑓 = 𝐷𝐷𝐷𝐷𝐷𝐷
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑

𝐸𝐸𝐸𝐸. 2 : 𝑧𝑧 = 𝑥𝑥 + 𝑦𝑦 2 (𝑎𝑎, 𝑏𝑏) = (−1, 1)

𝑑𝑑𝑑𝑑
• (−1, 1) = Derivative with regards to 𝑥𝑥 of [replace 𝑦𝑦 = 𝑏𝑏 = 1] at 𝑥𝑥 = 𝑎𝑎 = −1
𝑑𝑑𝑑𝑑

= 𝑥𝑥 + 1

=1
𝑑𝑑𝑑𝑑
• (−1, 1) = derivative with regards to 𝑦𝑦 of [replace 𝑥𝑥 = 𝑎𝑎 = −1] at 𝑦𝑦 = 𝑏𝑏 = 1
𝑑𝑑𝑑𝑑

= 𝑦𝑦 = −1

= 2𝑦𝑦 at 𝑦𝑦 = 1 = 2
 𝐸𝐸𝐸𝐸 3: 𝑧𝑧 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) = 2𝑥𝑥 2 + 3𝑥𝑥𝑥𝑥 + 4𝑦𝑦 2 + 5𝑥𝑥 + 6𝑦𝑦 − 7 (𝑎𝑎, 𝑏𝑏) = (−1, 1)
𝑑𝑑𝑑𝑑
• (−1, 1) = 𝑓𝑓𝑥𝑥 (−1, 1) = derivative with regards to 𝑥𝑥 of 𝑓𝑓(𝑥𝑥, 1) at 𝑥𝑥 = −1
𝑑𝑑𝑑𝑑

𝑓𝑓(𝑥𝑥, 1) = 2𝑥𝑥 2 + 3𝑥𝑥 + 4 + 5𝑥𝑥 + 6 − 7 = 4𝑥𝑥 + 81 = 4

𝑥𝑥 = −1 𝑥𝑥 = −1

𝑓𝑓𝑥𝑥 (𝑥𝑥, 1) = 4𝑥𝑥 + 3 + 5 = 4𝑥𝑥 + 8

𝐸𝐸𝐸𝐸 4 ∶ 𝑔𝑔(𝑥𝑥, 𝑦𝑦) = (𝑥𝑥 2 + 1)2 + (𝑦𝑦 3 − 3)3 + 5𝑥𝑥𝑦𝑦 3 − 2𝑥𝑥 2 𝑦𝑦 2 Chain rule

𝑔𝑔𝑔𝑔 (𝑎𝑎, 𝑏𝑏) −derivative with regards to 𝑦𝑦 of 𝑔𝑔(𝑎𝑎, 𝑦𝑦) at 𝑦𝑦 = 𝑏𝑏. (𝑢𝑢3 )′ = 3𝑢𝑢2 + 𝑢𝑢′

Arbitrary 𝑝𝑝 + 𝑠𝑠 → = [a2 +1)1]’y + 3(y2-3)2 * 3y2 +5a *3y2 -2a2 *2y

=0

Plug in 𝑎𝑎 = −1 and 𝑦𝑦 = 𝑏𝑏 = 1

 𝑧𝑧 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) = (𝑥𝑥 3 + 𝑦𝑦 2 )𝑒𝑒 𝑥𝑥𝑥𝑥+3𝑥𝑥−𝑦𝑦 (𝑒𝑒 4 )′ = 𝑒𝑒 4 ∗ 4′

𝑑𝑑𝑑𝑑
• (𝑥𝑥, 𝑦𝑦) = 𝑓𝑓𝑓𝑓(𝑥𝑥, 𝑦𝑦) = 3𝑥𝑥 2 ∗ 𝑒𝑒 𝑥𝑥𝑥𝑥+3𝑥𝑥−𝑦𝑦 + (𝑥𝑥 3 + 𝑦𝑦 2 )𝑒𝑒 𝑥𝑥𝑥𝑥+𝑥𝑥−𝑦𝑦 ∗ (𝑦𝑦 + 1)
𝑑𝑑𝑑𝑑

𝑥𝑥 = 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣
Last time (𝑎𝑎, 𝑏𝑏) 𝑓𝑓𝑓𝑓𝑓𝑓 = 3𝑥𝑥 2 𝑒𝑒 𝑥𝑥𝑥𝑥+𝑥𝑥−𝑦𝑦 (𝑥𝑥 − 1)0 + 2𝑦𝑦𝑒𝑒 𝑥𝑥𝑦𝑦 + 𝑥𝑥−𝑦𝑦
∗ (𝑦𝑦 + 1)
𝑦𝑦 = 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐

+(𝑥𝑥 3 + 𝑦𝑦 2 )𝑒𝑒 𝑥𝑥𝑥𝑥+𝑥𝑥−𝑦𝑦 (𝑥𝑥 − 1)(𝑦𝑦 + 1)

Arbitrary 𝑥𝑥 = 𝑎𝑎 +(𝑥𝑥 3 + 𝑦𝑦 2 )𝑒𝑒 𝑥𝑥𝑥𝑥+𝑥𝑥−𝑦𝑦 − 1

why not
use 𝑥𝑥, 𝑦𝑦 𝑦𝑦 = 𝑏𝑏

𝑑𝑑𝑑𝑑
• 𝑑𝑑𝑑𝑑
= 𝑓𝑓𝑦𝑦 (𝑥𝑥, 𝑦𝑦) = 2𝑦𝑦𝑒𝑒 𝑥𝑥𝑥𝑥+𝑥𝑥−𝑦𝑦 + (𝑥𝑥 3 + 𝑦𝑦 2 )𝑒𝑒 𝑥𝑥𝑥𝑥+𝑥𝑥−𝑦𝑦 (𝑥𝑥 − 1)

𝑦𝑦 = 𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉
𝑓𝑓𝑓𝑓𝑓𝑓 = 2𝑦𝑦𝑒𝑒 𝑥𝑥𝑥𝑥+𝑥𝑥−𝑦𝑦 (𝑦𝑦 + 1) + 3𝑥𝑥 2 𝑒𝑒 𝑥𝑥𝑥𝑥+𝑥𝑥−𝑦𝑦 (𝑥𝑥 − 1)
𝑥𝑥 = 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
+(𝑥𝑥 3 + 𝑦𝑦 2 )𝑒𝑒 𝑥𝑥𝑥𝑥+𝑥𝑥−𝑦𝑦 (𝑦𝑦 + 1)(𝑥𝑥 − 1) + (𝑥𝑥 3 + 𝑦𝑦 2 )𝑒𝑒 𝑥𝑥𝑥𝑥+𝑥𝑥𝑥𝑥 |
 𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑
= 𝑓𝑓𝑓𝑓

Are functions of (𝑥𝑥, 𝑦𝑦)

𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑
= 𝑓𝑓𝑓𝑓 *can evaluate for any 𝑥𝑥 and 𝑦𝑦.

2 Variable case: Let 𝑧𝑧 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) be a function of variables 𝑦𝑦, 𝑥𝑥 (2 partial derivatives)
𝑑𝑑𝑑𝑑
(𝑎𝑎) (𝑥𝑥, 𝑦𝑦) = partial derivative of 𝑓𝑓 𝑤𝑤𝑤𝑤𝑤𝑤ℎ respect to 𝑥𝑥.
𝑑𝑑𝑑𝑑

𝑑𝑑𝑑𝑑
(𝑏𝑏) (𝑥𝑥, 𝑦𝑦) = partial derivative of 𝑓𝑓 𝑤𝑤𝑤𝑤𝑤𝑤ℎ respect to 𝑦𝑦
𝑑𝑑𝑑𝑑

 Definition :
𝑔𝑔(𝑥𝑥+ℎ)−𝑔𝑔(𝑥𝑥)
• In MATA 32 : 𝑦𝑦 = 𝑔𝑔(𝑥𝑥) 𝑔𝑔′ (𝑥𝑥) = lim � �
ℎ→0 ℎ

MATA 33: 𝑧𝑧 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦)

𝑑𝑑𝑑𝑑 𝑓𝑓(𝑥𝑥+ℎ,𝑦𝑦)−𝑓𝑓(𝑥𝑥,𝑦𝑦)
(𝑥𝑥, 𝑦𝑦) = lim � �
𝑑𝑑𝑑𝑑 ℎ→0 ℎ

𝑑𝑑𝑑𝑑 𝑓𝑓(𝑥𝑥,𝑦𝑦+ℎ)−𝑓𝑓(𝑥𝑥,𝑦𝑦)
(𝑥𝑥, 𝑦𝑦) = lim � �
𝑑𝑑𝑑𝑑 ℎ→0 ℎ

Notation:
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑
(𝑥𝑥, 𝑦𝑦) or 𝑓𝑓𝑓𝑓(𝑥𝑥, 𝑦𝑦) or 𝑑𝑑𝑑𝑑
or 𝑍𝑍𝑍𝑍

𝐴𝐴 lot easier to differentiate.

 Second - order partial derivatives:

 Second - order mixed derivatives

:
 Higher - order partial and mixed derivatives:

Examples:

1. Z = (4x+9)(8x+5y)
2.

3. Introduction to Second derivatives

17.2 - 17.3 Implicit Partial Differentiation

Ex: Let 𝑥𝑥, 𝑦𝑦, 𝑧𝑧 = variables

4𝑥𝑥 2 𝑧𝑧 − 5𝑥𝑥𝑥𝑥 + 𝑦𝑦 𝑧𝑧 = 1 − 𝑧𝑧 + 3𝑥𝑥𝑦𝑦 2 Make it simple enough to isolate "𝑧𝑧"

4𝑥𝑥 2 2 + 𝑦𝑦𝑦𝑦 + 𝑧𝑧 = 1 + 3𝑥𝑥𝑦𝑦 2 + 5𝑥𝑥𝑥𝑥

𝑧𝑧(4𝑥𝑥 2 + 𝑦𝑦 + 1) = 1 + 3𝑥𝑥𝑦𝑦 2 + 5𝑥𝑥𝑥𝑥

1+3𝑥𝑥𝑦𝑦 2 +5𝑥𝑥𝑥𝑥
∴ 𝑧𝑧 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) = 4𝑥𝑥 2 +𝑦𝑦+1
→ Writing "𝑧𝑧" explicitly as function of 𝑥𝑥, 𝑦𝑦.

→ Now you can find partial derivatives

Implicit partial differentiation:

Allow us to find partial derivative without finding explicit function

Assume each variable acts implicitly as a function of the other

Assume𝑧𝑧 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) whether or not we can write function

Find 𝑧𝑧𝑧𝑧 and 𝑧𝑧𝑧𝑧 implicitly

𝑑𝑑 𝑑𝑑
a) 𝑧𝑧𝑥𝑥 : 𝑑𝑑𝑥𝑥 (4𝑥𝑥 2 𝑧𝑧 − 5𝑥𝑥𝑥𝑥 + 𝑦𝑦𝑦𝑦) = 𝑑𝑑𝑥𝑥 (1 − 𝑧𝑧 + 3𝑦𝑦 2 )

8𝑥𝑥𝑥𝑥 + 4𝑥𝑥 2 𝑧𝑧𝑥𝑥 − 5𝑦𝑦 + 𝑦𝑦 𝑧𝑧𝑥𝑥 = − 𝑧𝑧𝑥𝑥 + 3𝑦𝑦 2

Now solve for 𝑧𝑧𝑥𝑥

𝑧𝑧𝑥𝑥 (4𝑥𝑥 2 + 𝑦𝑦 + 1) = 3𝑦𝑦 2 + 5𝑦𝑦 − 8𝑥𝑥𝑥𝑥

3𝑦𝑦 2 +5𝑦𝑦−8𝑥𝑥𝑥𝑥
∴ 𝑧𝑧𝑥𝑥 = 4𝑥𝑥 2 +𝑦𝑦+1
→ 𝑧𝑧𝑥𝑥 = function of 𝑥𝑥, 𝑦𝑦 + 𝑧𝑧

𝑑𝑑 𝑑𝑑
b) 𝑧𝑧𝑧𝑧: 𝑑𝑑𝑦𝑦 (4𝑥𝑥 2 𝑧𝑧 − 5𝑥𝑥𝑥𝑥) = 𝑑𝑑𝑦𝑦 (1 − 𝑧𝑧 + 3𝑦𝑦 2 )

Same thing (vice versa)

Example: 𝑥𝑥 + 2𝑦𝑦 + 1 = 0
𝑥𝑥+1 1
𝑦𝑦 = 𝑦𝑦(𝑥𝑥) = − 2
𝑦𝑦 ′ = 2

Example: 𝐹𝐹(𝑥𝑥, 𝑦𝑦) = 𝑥𝑥 3 + 𝑦𝑦 3 − 2𝑥𝑥𝑥𝑥 𝐹𝐹(𝑥𝑥, 𝑦𝑦) = 0

a) (1, 1) is on curve 𝐹𝐹 = 0
𝐹𝐹(1,1) = 1 + 1 − 2 = 0
b) 𝑦𝑦 ′ (1) =?
𝑑𝑑
𝐹𝐹�𝑥𝑥, 𝑦𝑦(𝑥𝑥)� = 0 𝑑𝑑𝑑𝑑

Insert 𝑥𝑥 = 1, 𝑦𝑦 = 1 → 3𝑥𝑥 2 + 3𝑦𝑦 2 ∙ 𝑦𝑦 2 − 2𝑦𝑦 − 2𝑥𝑥𝑥𝑥 = 0.

�𝑦𝑦−𝑥𝑥 2 +4
Ex: 𝑧𝑧 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) =
�5−𝑦𝑦

𝑦𝑦 − 𝑥𝑥 2 + 4 ≥ 0
𝐷𝐷 = {(𝑥𝑥, 𝑦𝑦)𝐸𝐸𝑅𝑅 2 �
−𝑦𝑦 + 5 > 0

Intersection 𝑦𝑦 = 5 𝑦𝑦 = 5 𝑥𝑥

𝑦𝑦 − 𝑥𝑥 2 + 4 = 0 𝑥𝑥 2 = 9 𝑦𝑦 = 𝑥𝑥 2 − 4
𝑥𝑥+4
Ex: 𝑓𝑓(𝑥𝑥, 𝑦𝑦) = 𝑥𝑥−𝑦𝑦 𝑓𝑓𝑥𝑥 (2,4) =? 𝑌𝑌

(𝑥𝑥,𝑦𝑦)=1(𝑥𝑥−𝑦𝑦)−1(𝑥𝑥+𝑦𝑦) −2𝑦𝑦 −2(4) −8

𝑓𝑓𝑥𝑥 (𝑥𝑥−𝑦𝑦)2
= (𝑥𝑥−𝑦𝑦)2 = (2−4)2 = 4
= −2

Ex: 𝑧𝑧 = 𝑥𝑥 2 𝑒𝑒 𝑥𝑥𝑥𝑥
𝑑𝑑𝑑𝑑
→ Product Rules 𝑑𝑑𝑑𝑑 = 2𝑥𝑥𝑒𝑒 𝑥𝑥𝑥𝑥 + 𝑥𝑥 2 𝑒𝑒 𝑥𝑥𝑥𝑥 ∙ 𝑦𝑦 = (2𝑥𝑥 + 𝑥𝑥 2 𝑦𝑦)𝑒𝑒 𝑥𝑥𝑥𝑥

𝑥𝑥𝑧𝑧 2 −𝑦𝑦2 𝑧𝑧−1 𝑑𝑑2 𝑧𝑧

Ex: 𝐹𝐹
=0 𝑑𝑑2 𝑥𝑥
=?

𝑧𝑧 2 𝑧𝑧 2 𝑑𝑑 𝑧𝑧 2
𝐹𝐹𝑥𝑥 = 𝑧𝑧 2 → 𝑧𝑧𝑥𝑥 = 2𝑥𝑥𝑥𝑥−𝑦𝑦 2 = 𝑦𝑦 2 −2𝑥𝑥𝑥𝑥 𝑧𝑧𝑥𝑥 = (𝑧𝑧𝑧𝑧) = 𝑑𝑑𝑥𝑥 [𝑦𝑦 2 −2𝑥𝑥𝑥𝑥]

2𝑧𝑧∙𝑧𝑧𝑥𝑥 �𝑦𝑦 2 −2𝑥𝑥𝑥𝑥�−𝑧𝑧 2 (−2𝑧𝑧−2𝑥𝑥𝑧𝑧𝑥𝑥 )

𝐹𝐹𝑧𝑧 = 2𝑥𝑥𝑥𝑥 − 𝑦𝑦 2 = (𝑦𝑦 2 −2𝑥𝑥𝑥𝑥)2

𝑧𝑧 2 2𝑧𝑧𝑧𝑧𝑥𝑥 �𝑦𝑦 2 −2𝑥𝑥𝑥𝑥�+2𝑧𝑧 2 (𝑧𝑧+𝑥𝑥𝑧𝑧𝑥𝑥 )

Plug in 𝑧𝑧𝑥𝑥 = 𝑦𝑦 2 −2𝑥𝑥𝑥𝑥 →= ( )2
Applications
hours
Marginal Productivity: 𝑃𝑃 = √𝑅𝑅 Let 𝑡𝑡 = # of week
k = capital

e.g. 𝑡𝑡 = 400, 𝑘𝑘 = 16; marginal productivity is

𝑑𝑑𝑑𝑑 √𝑘𝑘 4
𝑑𝑑𝑑𝑑
= 2 𝑡𝑡� 𝑘𝑘=16 = 2.20 = 0.1 ← ∆ of production increate 𝑡𝑡 by one and keeping 𝑘𝑘 constant = 0.1
√
𝑡𝑡=400

𝑑𝑑𝑑𝑑 √𝑡𝑡 20
𝑑𝑑𝑑𝑑
= 2√𝑘𝑘� 𝑘𝑘=16 = 2.4 = 2.5 ← ∆ of production increase 𝑘𝑘 by one and keeping 𝑡𝑡 instant = 2.5
𝑡𝑡=400

Competitive complimentary Products: 2 related that ∆ of price affects one another

Definition

𝑑𝑑𝑑𝑑𝑑𝑑 𝑓𝑓𝑓𝑓 = 𝑓𝑓(𝑝𝑝𝐴𝐴 , 𝑃𝑃𝐵𝐵 ) − demand 𝑓𝑓 for 𝐴𝐴

>0
𝑑𝑑𝑑𝑑𝑑𝑑 𝑓𝑓𝑓𝑓 = 𝑔𝑔(𝑃𝑃𝐴𝐴 , 𝑃𝑃𝐵𝐵 ) − demand 𝑓𝑓 for 𝐵𝐵

Exercise: Marginal cost at given production level

a) 𝑐𝑐 = 0.03(𝑥𝑥 + 𝑦𝑦)3 − 0.6(𝑥𝑥 + 𝑦𝑦)2 + 9.5(𝑥𝑥 + 𝑦𝑦) + 7700

𝑑𝑑𝑑𝑑
= 0.03 ∙ 3(𝑥𝑥 + 𝑦𝑦)2 − 0.6 ∗ 2(𝑥𝑥 + 𝑦𝑦)3 + 9.5
𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑
(50,80) = 0.09(130)2 − 1.2(130) + 9.5 = 9.169 − 12.13 + 9.5
𝑑𝑑𝑑𝑑
= 1521 − 156 + 9.5 = 1374.5

Competitive and Complement Products

Definitions:
𝑑𝑑𝑑𝑑𝐴𝐴 𝑑𝑑𝑑𝑑𝐵𝐵
1) Competitive 𝑑𝑑𝑑𝑑𝑑𝑑 > 0 𝑑𝑑𝑑𝑑𝑑𝑑
>0 (+)
𝑑𝑑𝑑𝑑𝐴𝐴 𝑑𝑑𝑑𝑑𝐵𝐵
2) Complementary 𝑑𝑑𝑑𝑑𝑑𝑑 < 0 𝑑𝑑𝑑𝑑𝑑𝑑
<0 (−)
 17.4 Higher-Order Derivatives

Main idea−2 variables: 𝑧𝑧 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) be MATA 33 function

The Partial Derivatives

𝑓𝑓𝑓𝑓(𝑥𝑥, 𝑦𝑦) and 𝑓𝑓𝑓𝑓(𝑥𝑥, 𝑦𝑦) are MATA 33 functions of 𝑥𝑥 + 𝑦𝑦

1𝑠𝑠𝑠𝑠 Order

 Second (2𝑛𝑛𝑛𝑛 ) order:

𝑓𝑓𝑥𝑥𝑥𝑥 (𝑥𝑥, 𝑦𝑦) and 𝑓𝑓𝑦𝑦𝑦𝑦 (𝑥𝑥, 𝑦𝑦)

𝑓𝑓𝑓𝑓𝑓𝑓(𝑥𝑥, 𝑦𝑦) 𝑓𝑓𝑓𝑓𝑓𝑓(𝑥𝑥, 𝑦𝑦)

𝑑𝑑2 𝑧𝑧 𝑑𝑑 𝑑𝑑𝑑𝑑
• 𝑓𝑓𝑥𝑥𝑥𝑥 (𝑥𝑥, 𝑦𝑦) = �𝑓𝑓𝑥𝑥(𝑥𝑥, 𝑦𝑦)�𝑥𝑥 or 𝑑𝑑𝑥𝑥 2
= 𝑑𝑑𝑑𝑑 �𝑑𝑑𝑑𝑑�
𝑑𝑑2 𝑧𝑧 𝑑𝑑 𝑑𝑑𝑑𝑑
𝑓𝑓𝑓𝑓𝑓𝑓(𝑥𝑥, 𝑦𝑦) = �𝑓𝑓𝑓𝑓(𝑥𝑥, 𝑦𝑦)�𝑦𝑦 or 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
= 𝑑𝑑𝑑𝑑 �𝑑𝑑𝑑𝑑� → Mix 2𝑛𝑛𝑛𝑛 order partial using differently
variables
𝑑𝑑2 𝑧𝑧 𝑑𝑑 𝑑𝑑𝑑𝑑
• 𝑓𝑓𝑓𝑓𝑓𝑓(𝑥𝑥, 𝑦𝑦) = �𝑓𝑓𝑓𝑓(𝑥𝑥, 𝑦𝑦)�𝑥𝑥 or 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
= 𝑑𝑑𝑑𝑑 �𝑑𝑑𝑑𝑑�
𝑑𝑑2 𝑧𝑧 𝑑𝑑 𝑑𝑑𝑑𝑑
𝑓𝑓𝑓𝑓𝑓𝑓(𝑥𝑥, 𝑦𝑦) = �𝑓𝑓𝑓𝑓(𝑥𝑥, 𝑦𝑦)�𝑦𝑦 or 𝑑𝑑𝑦𝑦 2
= 𝑑𝑑𝑑𝑑 �𝑑𝑑𝑑𝑑�
  There are 3𝑟𝑟𝑟𝑟 order etc :

With every higher order, the 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 of partial derivatives doubles (e. g. 3𝑟𝑟𝑟𝑟 order has 8)

𝑓𝑓" = (𝑓𝑓′)′ 𝑓𝑓 ′′′ = (𝑓𝑓 ′′ )′

𝑑𝑑 𝑑𝑑2 𝑧𝑧
𝑑𝑑𝑑𝑑 𝑑𝑑𝑥𝑥 2
= 𝑓𝑓𝑓𝑓𝑓𝑓

𝑑𝑑 𝑑𝑑 𝑑𝑑2 𝑧𝑧
𝑑𝑑𝑑𝑑
𝑓𝑓𝑓𝑓(𝑥𝑥, 𝑦𝑦) 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
= 𝑓𝑓𝑓𝑓𝑓𝑓

𝑧𝑧 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) They are equal

𝑑𝑑 𝑑𝑑 𝑑𝑑2 𝑧𝑧
𝑑𝑑𝑑𝑑
𝑓𝑓𝑓𝑓(𝑥𝑥, 𝑦𝑦) 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
= 𝑓𝑓𝑓𝑓𝑓𝑓

𝑑𝑑 𝑑𝑑2 𝑧𝑧
𝑑𝑑𝑑𝑑 𝑑𝑑𝑦𝑦 2
= 𝑓𝑓𝑓𝑓𝑓𝑓

There are 3 second derivatives

𝑓𝑓𝑓𝑓𝑓𝑓, 𝑓𝑓𝑓𝑓𝑓𝑓 = 𝑓𝑓𝑓𝑓𝑓𝑓, 𝑓𝑓𝑓𝑓𝑓𝑓

 𝑢𝑢 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦2) = 𝑔𝑔𝑒𝑒 2𝑥𝑥𝑥𝑥𝑥𝑥 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 =?

2𝑥𝑥𝑥𝑥𝑥𝑥
• Step 1 → 𝑓𝑓𝑓𝑓 = 𝑔𝑔𝑒𝑒 ∗ 2𝑥𝑥𝑥𝑥 = 18𝑒𝑒 2𝑥𝑥𝑥𝑥 ∗ 𝑥𝑥𝑥𝑥
• Step 2 → 𝑓𝑓𝑓𝑓𝑓𝑓 = 18𝑒𝑒 2𝑥𝑥𝑥𝑥𝑥𝑥 ∗ 2𝑦𝑦𝑦𝑦 ∗ 𝑥𝑥𝑥𝑥 + 18𝑒𝑒 2𝑥𝑥𝑥𝑥𝑥𝑥 ∗ 𝑧𝑧
• Step 3 → 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓
Examples

1.

2.
3.
17.5 Chain Rule
Chain rules → Purpose is to differentiate functions that are compositions

→ MATA 32: 𝑦𝑦 = (𝑓𝑓𝑓𝑓)(𝑥𝑥) = 𝑓𝑓�𝑔𝑔(𝑥𝑥)� * imposition at function 𝑓𝑓 and 𝑔𝑔

𝑑𝑑𝑑𝑑
𝑦𝑦 ′ = 𝑑𝑑𝑑𝑑 = 𝑓𝑓 ′ �𝑔𝑔(𝑥𝑥)�, 𝑔𝑔′(𝑥𝑥)

OR
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑
= 𝑑𝑑𝑑𝑑 . 𝑑𝑑𝑑𝑑 → 𝑦𝑦 = 𝑓𝑓(𝑤𝑤) where 𝑤𝑤 = 𝑔𝑔(𝑥𝑥)

Variable function

→ MATA 33 chain Rules

Generalize: let 𝑧𝑧 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) 𝑥𝑥 = 𝑔𝑔(𝑟𝑟, 𝑠𝑠) 𝑦𝑦 = ℎ(𝑟𝑟, 𝑠𝑠)

𝑟𝑟, 𝑠𝑠 = variables

𝑥𝑥, 𝑦𝑦, 𝑧𝑧 =functions * 𝑟𝑟, 𝑠𝑠 are ultimate variables of 𝑧𝑧

∴ 𝑧𝑧 = 𝑓𝑓�𝑔𝑔(𝑟𝑟, 𝑠𝑠), ℎ(𝑟𝑟, 𝑠𝑠)�

 Partial differentiation:
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
𝑟𝑟 & 𝑠𝑠 𝑑𝑑𝑑𝑑
= 𝑑𝑑𝑑𝑑 . 𝑑𝑑𝑑𝑑 + 𝑑𝑑𝑑𝑑 . 𝑑𝑑𝑑𝑑 * differentiate each intermediate 𝑈𝑈𝑈𝑈𝑈𝑈 to each ultimate

𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑

Ultimate 𝑑𝑑𝑑𝑑
= 𝑑𝑑𝑑𝑑 . 𝑑𝑑𝑑𝑑 + 𝑑𝑑𝑑𝑑 . 𝑑𝑑𝑑𝑑

𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑠𝑠
= �𝑑𝑑𝑑𝑑 . 𝑑𝑑𝑑𝑑� � 𝑑𝑑𝑑𝑑 �
𝑑𝑑𝑑𝑑
𝑠𝑠

(2)

Can be seen as tow matrices

Ex. 1 𝑧𝑧 = 5𝑥𝑥𝑦𝑦 2 + 𝑥𝑥 3 𝑥𝑥 = 𝑟𝑟 2 𝑠𝑠 𝑦𝑦 = 𝑟𝑟 + 𝑦𝑦 𝑠𝑠 2′

𝑓𝑓(𝑥𝑥, 𝑦𝑦) 𝑔𝑔(𝑟𝑟, 𝑠𝑠) ℎ (𝑣𝑣, 𝑠𝑠)

 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
Find 𝑑𝑑𝑑𝑑
� 𝑑𝑑𝑠𝑠1

𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑

𝑑𝑑𝑑𝑑
= 𝑑𝑑𝑑𝑑 . 𝑑𝑑𝑑𝑑 + 𝑑𝑑𝑑𝑑 . 𝑑𝑑𝑑𝑑

= �5𝑦𝑦 2 + 3𝑥𝑥 2 �(2𝑟𝑟𝑟𝑟) + (10𝑥𝑥𝑥𝑥)(1) simpler = 𝑆𝑆𝑆𝑆

= (5𝑦𝑦 2 + 3𝑥𝑥 2 )(2𝑟𝑟𝑟𝑟) + (10𝑥𝑥𝑥𝑥)(1)

Sub 𝑥𝑥 and 𝑦𝑦 → = (5(𝑟𝑟 + 4𝑠𝑠 2 )2 + 3(𝑟𝑟 2 𝑠𝑠)2 ) ∗ 2𝑟𝑟𝑟𝑟 + �10(𝑟𝑟 2 𝑠𝑠)(𝑟𝑟 + 4𝑠𝑠 2 )�
𝑑𝑑𝑑𝑑
Evaluate �
𝑑𝑑𝑑𝑑 𝑟𝑟=2
𝑠𝑠=1

*Method 1 = sub in 𝑟𝑟 and 𝑠𝑠

𝑑𝑑𝑑𝑑 2
�𝑟𝑟=2 = �5(2 + 4(1)2 )2 + 3�22 (1)� � ∗ 2(2) + (10�22 (1)�(2 + 4(1)2 )
𝑑𝑑𝑑𝑑
𝑠𝑠=1

= (180 + 48)4 + (10 ∗ 24)

= 4(228) + 240 = 912 + 240

= 1152

*Method 2 = Use inter mediate variables

𝑥𝑥 = 𝑟𝑟 2 𝑠𝑠 = 4(1) = 4

𝑦𝑦 = 𝑟𝑟 + 4𝑠𝑠 2 = 6

Now 𝑆𝑆𝑆𝑆𝑆𝑆 in SE:

𝑑𝑑𝑑𝑑
= (5𝑦𝑦 2 + 3𝑥𝑥 2 )(2𝑟𝑟𝑟𝑟) + (10𝑥𝑥𝑥𝑥)
𝑑𝑑𝑑𝑑
= (180 + 48)(4) + (10.24)

= 4(228) + 240 = 912 + 240

= 1152
 Chain Rule (𝑒𝑒")′ = 𝑒𝑒 4 ∗ 𝑢𝑢′ 𝑓𝑓(𝑥𝑥) = 𝑒𝑒 4 (𝑥𝑥)

𝑓𝑓(𝑥𝑥) = 𝑒𝑒 2 = exp(𝑥𝑥) 𝑔𝑔(𝑥𝑥) = 𝑒𝑒

𝑔𝑔(𝑥𝑥) = 4(𝑥𝑥) (𝑓𝑓𝑓𝑓𝑓𝑓)𝑥𝑥 = 𝑓𝑓�𝑔𝑔(𝑥𝑥)� = exp�𝑢𝑢(𝑥𝑥)�

= 𝑒𝑒 4𝑥𝑥

𝑓𝑓(𝑥𝑥) = ln(𝑥𝑥 3 + 𝑥𝑥 2 + 1) 𝑓𝑓(𝑥𝑥) = ln 𝑥𝑥 𝑔𝑔(𝑥𝑥) = 𝑥𝑥 3 + 𝑥𝑥 2 + 1

1
𝑓𝑓 ′ (𝑥𝑥) = ln′�𝑔𝑔(𝑥𝑥)� ∗ 𝑔𝑔′ (𝑥𝑥) = 𝑥𝑥 3 +𝑥𝑥 2 +1 ∗ (3𝑥𝑥 2 + 2𝑥𝑥)

𝑑𝑑𝑑𝑑
• 𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
𝑦𝑦 = 𝑓𝑓(𝑥𝑥) = 𝑓𝑓�𝑔𝑔(𝑥𝑥)� 𝑦𝑦 ′ = 𝑑𝑑𝑑𝑑 . 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
�• 𝑢𝑢(𝑥𝑥) = 4
𝑑𝑑𝑑𝑑
• 𝑑𝑑𝑑𝑑

𝑧𝑧 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) = 𝑓𝑓�𝑥𝑥(𝑠𝑠, 𝑡𝑡), 𝑦𝑦(𝑠𝑠, 𝑡𝑡)�

𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑
= 𝑑𝑑𝑑𝑑 . 𝑑𝑑𝑑𝑑 + 𝑑𝑑𝑑𝑑 . 𝑑𝑑𝑑𝑑

𝑥𝑥 𝑦𝑦 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑

= . + .
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑

𝑠𝑠 𝑠𝑠 𝑡𝑡
𝑡𝑡
 17.6 Extrema for Functions of Two Variables

Definition: 𝑧𝑧 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) have a relative maximum at point (𝑎𝑎, 𝑏𝑏) if all points (𝑥𝑥, 𝑦𝑦) in the plane
that are sufficiently close to (𝑎𝑎, 𝑏𝑏) we have

𝑓𝑓(𝑎𝑎, 𝑏𝑏) ≥ 𝑓𝑓(𝑥𝑥, 𝑦𝑦)

*For 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 Min we replace ≥ by ≤

Definition: 𝑧𝑧 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) have a relative min at point (𝑎𝑎, 𝑏𝑏) if all points (𝑥𝑥, 𝑦𝑦) in the plane that
are sufficiently close to (𝑎𝑎, 𝑏𝑏) we have

𝑓𝑓(𝑎𝑎, 𝑏𝑏) ≤ 𝑓𝑓(𝑥𝑥, 𝑦𝑦)

The 𝟏𝟏𝒔𝒔𝒔𝒔 derivative test:

Fermat’s theorem −𝑓𝑓 has local max/min 𝑎𝑎𝑎𝑎 𝑐𝑐

And if 𝑓𝑓′(𝑐𝑐) exists, then 𝑓𝑓 ′ (𝑐𝑐) = 0

• Let 𝑓𝑓 = function of 2 variables. If it has a local max/min 𝑎𝑎𝑎𝑎 (𝑎𝑎, 𝑏𝑏) and the first−order partial
derivatives of 𝑓𝑓 exist there, then 𝑓𝑓𝑥𝑥 (𝑎𝑎, 𝑏𝑏) = 0 𝑓𝑓𝑦𝑦 (𝑎𝑎, 𝑏𝑏) = 0
• Tangent plane to the graph of 𝑓𝑓 𝑎𝑎𝑎𝑎 local max/min = horizontal

* A point (𝑎𝑎, 𝑏𝑏) such that 𝑓𝑓𝑓𝑓(𝑎𝑎, 𝑏𝑏) = 𝑓𝑓𝑓𝑓(𝑎𝑎, 𝑏𝑏) = 0 is called a critical point of 𝑓𝑓

Locate local extreme of a function, we must determinists critical points. (Not necessarily local
extrema)

Ex: 𝑓𝑓(𝑥𝑥, 𝑦𝑦) = 𝑦𝑦 2 − 𝑥𝑥 2

𝑓𝑓𝑥𝑥 (𝑥𝑥, 𝑦𝑦) = −2𝑥𝑥 ∴ (0, 0) only a critical point

𝑓𝑓𝑦𝑦 (𝑥𝑥, 𝑦𝑦) = 2𝑦𝑦 (Not local enticement)

Ex.: 𝑓𝑓(𝑥𝑥, 𝑦𝑦) = 𝑥𝑥 2 + 𝑦𝑦 2 − 2𝑥𝑥 − 6𝑦𝑦 + 14

𝑓𝑓𝑥𝑥 (𝑥𝑥, 𝑦𝑦) = 2𝑥𝑥 − 2 → 𝑥𝑥 = 1

𝑓𝑓𝑦𝑦 (𝑥𝑥, 𝑦𝑦) = 2𝑦𝑦 − 6 → 𝑦𝑦 = 3

∴ (1, 3) = 𝑎𝑎bsolute min of f

 (𝑥𝑥 − 1)2 ≥ 0 → 𝑓𝑓(𝑥𝑥, 𝑦𝑦) ≥ 4

(𝑦𝑦 − 3)2 ≥ 0

2 −𝑦𝑦 2 √2
𝐸𝐸𝐸𝐸.: 𝑓𝑓(𝑥𝑥, 𝑦𝑦) = 𝑥𝑥𝑒𝑒 −𝑥𝑥 ∴ � 2 , 0� = 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 max

2 −𝑦𝑦 2 2 −𝑦𝑦 2 √2 −√2

𝑓𝑓𝑥𝑥 = 𝑒𝑒 −𝑥𝑥 + 𝑥𝑥(−2𝑥𝑥)𝑒𝑒 −𝑥𝑥 → 𝑥𝑥 = ± � , 0� = 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 min
2 2

2 −𝑦𝑦 2
𝑓𝑓𝑦𝑦 = 𝑥𝑥(−2𝑦𝑦)𝑒𝑒 −𝑥𝑥 → 𝑦𝑦 = 0

Critical Points

• 𝐴𝐴 Critical Point which = neither local max/min = is called saddle point

• 𝑎𝑎𝑎𝑎 critical point → No max, has absolute min value
• 𝑎𝑎𝑎𝑎 critical points has absolute mix and max value.

𝟐𝟐𝒏𝒏𝒏𝒏 Derivative test 𝒘𝒘𝒘𝒘𝒘𝒘𝒘𝒘 𝟐𝟐 variables

• Test points (a,b) for the max/min/saddle points (critical points)

• Suppose 2𝑛𝑛𝑛𝑛 partial derivatives of 𝑓𝑓 are continuous about (𝑎𝑎, 𝑏𝑏) and point (𝑎𝑎, 𝑏𝑏) =critical
for 𝑓𝑓.
𝑓𝑓𝑥𝑥 (𝑎𝑎, 𝑏𝑏) = 0 𝑓𝑓𝑦𝑦 (𝑎𝑎, 𝑏𝑏) = 0
• We denote:
𝑓𝑓𝑥𝑥𝑥𝑥 𝑓𝑓𝑥𝑥𝑥𝑥 2
𝐷𝐷 = 𝐷𝐷(𝑎𝑎, 𝑏𝑏) = det � � = 𝑓𝑓𝑥𝑥𝑥𝑥 (𝑎𝑎, 𝑏𝑏)𝑓𝑓𝑦𝑦𝑦𝑦 (𝑎𝑎, 𝑏𝑏) − �𝑓𝑓𝑥𝑥𝑥𝑥 (𝑎𝑎, 𝑏𝑏)�
𝑓𝑓𝑥𝑥𝑥𝑥 𝑓𝑓𝑦𝑦𝑦𝑦

• Holds: (1) 𝐷𝐷 > 0 𝑓𝑓(𝑎𝑎, 𝑏𝑏) = local min

𝑓𝑓𝑥𝑥𝑥𝑥 (𝑎𝑎, 𝑏𝑏) > 0

(2) 𝐷𝐷 > 0 𝑓𝑓(𝑎𝑎, 𝑏𝑏) =local max

𝑓𝑓𝑥𝑥𝑥𝑥 (𝑎𝑎, 𝑏𝑏) < 0

(3) 𝐷𝐷 < 0 → 𝑓𝑓(𝑎𝑎, 𝑏𝑏) 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝑛𝑛𝑛𝑛𝑛𝑛 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 local max/min. In this case, point (𝑎𝑎, 𝑏𝑏) = called saddle
point of 𝑓𝑓 and graph of 𝑓𝑓 crosses its tangent plane 𝑎𝑎𝑎𝑎 (𝑎𝑎, 𝑏𝑏)

(4) 𝐷𝐷 = 0 → Test gives no information

APPLICATION EXAMPLES:

Let f(x,y) = x^4 + y^4 – 4xy + 1

fx = 4x^3 – 4y = 0  x^3 = y
fy = 4y^3 – 4x = 0  y^3 = x
Second Derivative Test:
fxx = 12x^2

fyy = 12y^2 D = (12x^2)(12y^2) – 16 = 144x^2y^2 - 16

fxy = -4
Compute D at each critical point:

(0,0)  D = -16 < 0 not local max/min

(1,1)  D = 128 > 0 local min

(-1,-1)  D = 128 local min

Second Derivative Test with multiple variables:

Suppose a = (a1, a2 … an) = critical point of function z = f (x1, x2 … xn)

a. Hessian of f at a is defined by:

b. Test distinguishes:
i. Det (A) not = 0  if det (A1) > 0, det (A2) > 0 , det (An) > 0 
a = local min if -det (A1) < 0, +det (A2) > 0 , det (An) > 0  a =
local max if any other combo, it yields a saddle point (neither
min/max)
ii. Det (A) = 0  no conclusion drawn
17.7 – Lagrange Multipliers

Definition:

o is called the Lagrange Multiplier.

o A technique/method to find and optimize local extrema of a function of > 2 variables

where the function is subject to > 1 constraint equations
o Max and Min of f (x,y,z) are subjects to one constraint g(x,y,z) = K

1. Method of lagrange multipliers for one constraint:

a. To find max and min of f (x,y,z) subject to one constraint g(x,y,z) = K

(assume extreme values exist and g is not 0 on the surface g (x,y,z) = K)

b. Find all values of x, y, z / fx = gx

fy = gy

fz = gz

c. Evaluate f at all points ( x, y, z ) that result from the step above then:
i. Largest of these values = max of f
ii. Smallest of these values = min of f

APPLICATION EXAMPLE:

Compute the max and min of function f (x, y) = x^2 + y^2

 Subject to constraint x^4 + y^4 = 1

STEP 1:

Find all values of x, y, z, such that:

fx = gx
1.
fy = gy

2.
fz = gz

x=0y=+1
y=0x=+ 1 contradiction

fx = gx  2x = 4x^3 x=0y=+1
x=0y=+1
x = 0  1 = 2x^2 y = 0  1 = 2y^2 f(x,y) = 1

x = 0  1 = 2x^2
y=0x=+ 1 f(x,y) = 1

fy = gy  2y = 4y^3 y=0x=+1
x = 0  1 = 2x^2
y=01= 2y^2 y = 0  1 = 2y^2 x^2 = y^2

2x^4 = 1  x +
2y^4 = 1  y +

** x ^ 4 + y ^ 4 = 1 **

STEP 2:

Evaluate f at all points ( x , y , z ) that result from step 1:

o Largest of these values = max of f

o Smallest of these values = min of f

Overview Example:
Let z = f ( x , y ) = 5x + 2y which is subject to constraint of 5x^2 + 2y^2 = 14

a) E = { ( x , y ) | 5x^2 + 2y^2 = 14 }

**we only consider points that are on the ellipse**

b) Find all constrained critical points.

Constrained function : g(x,y) = 5x^2 + 2y^2 – 14
**we have g(x,y) = 0 5x^2 + 2y^2 = 14**

2. Method of lagrange multipliers for two constraints

a. To find max and min of f ( x , y , z ) with the following constraints:

i. g(x,y,z)= K

ii. h(x,y,z)= l
b. Find all values of x, y, z , , such that:
+ hx

+ hy
i.
+ hz

ii. g(x,y,z)=K

iii. H ( x, y , z ) = l

c. Evaluate f at all points ( x, y, z ) that result from the step above:

i. Largest of these values = max of f

ii. Smallest of these values = min of f

APPLICATION EXAMPLE:

Compute the max and min of the function u = f (x,y,z)

g(x,y,z)=K
Subject to constraints
h(x,y,z)=l
Must solve: g(x,y,z)=K After: Compute f (x,y,z) at the solutions

h(x,y,z)=l See which is max, and which is min.

fx = gx + hx

fy = gy+ hy

fz = gz+ hz

g=x+y+z=1
f ( x , y , z ) = x^2 + 2y^2 + 3yz + 4z^2
h = 3x + 2y + z = 2

x+y+z=1

3x + 2y + z = 2

2x = +3 = + 9y + 24z - 3 = 9y + 24z – 2 ( 2y + 13z ) = 5y – 2z

4y + 3z = +2 = + 6y + 16z - 2 = - + 6y + 16z  = 2y + 13z

3y + 8z = +  = 3y + 8z -

First Derivative:

x+y+z=1 Solve 3x3 system x=?

3x + 2y + z = 2 y=?

2x – 5y + 2z = 0 z=?

Second Derivative:

Compute f(x,y,z) – It is a minimum

17.9 – Multiple Integrals

Basic Integration:

Let z = f (x , y) ; we can imagine

EXAMPLE: z = f (x , y) = 3x + 5y

Definition:

 Integrated Integral ( first with respect to ‘y’, then with respect to ‘x’)
  Similarly, define iterated integral ( first with respect to ‘x’, then with respect to ‘y’) by
formula

Fubini’s Theorem:

If f is continuous on the rectangle R = {( x , y ) | a < x < b , c < y < d }

Double Integrals:

These are the basic properties of double integrals.

APPLICATION EXAMPLES:

1.

2.

3.
Review

Review - Multiple Integers

Definition:

 Integrated Integral ( first with respect to ‘y’, then with respect to ‘x’)

 Similarly, define iterated integral ( first with respect to ‘x’, then with respect to ‘y’) by
formula

Practice

1. Determine and classify all critical points of f = x^3 + xz^2 – 3x^2 + y^2 + 2z^2

a. First Derivative Test:

x = -2
fx = 0 3x^2 + z^2 – 6x = 0   12 + z^2 + 12 = 0  z^2 = -24
y=0

x = -2
x=0
 3x^2 – 6x = 0  3x ( x-2 ) 
y=0
x=2
fy = 0 2y = 0  y = 0
 fz = 0 2xz + 4z = 0  2z ( x + 2 )  x = -2 OR z = 0

3x^2 + z^2 – 6x = 2xz + 4z

Therefore critical points are ( 0 , 0 , 0 ) and ( 2 , 0 , 0 )

b. Second Derivative Test:

fxx fxy fxz 6x – 6 0 2z

H = fxy fyy fyz  0 2 0
fxz fyz fzz
2z 0 2x + 4

therefore if ( 0 , 0 , 0 ) = -6 0 0 det1 = -6 , det2 = -12 , det3 = -48

0 2 0 therefore, neither min/max = saddle point

0 0 4

c. For 2 variables: D = fxx (x,y) fyy (x,y) - [ fxy (x,y ) ]^2

d. For 3 variables: det(A1) > 0 , det(A2) > 0 … det(An) > 0  local min.

-det(A1) < 0 , +det(A2) > 0 , -det(A3) < 0  local max.

Any other sign combos = saddle point (neither min/max)

therefore if ( 2 , 0 , 0 ) = -6 0 0 det1 = 6 , det2 = 12 , det3 = 96

0 2 0 therefore, local minimum ( + , + , + )

0 0 8
 2. Determinant Review:

det 6 1 2 = 6det 4 -1 - 1det 2 0 + 2det 0 0

3 4 -1 2 8 0 8 0 8

0 2 8

= 6 ( 32 – 2 ) – 24 + 2 ( 6 )
= 180 – 24 + 12
= 168

3. Integral Practice:

4. Application Word Problems:

Revenue of a company is R = 100x^2/3 y^1/3 x = unit of labour y = unit of steel

Cost/labour = $20 Total Budget = $20 000
Cost/steel = $2000

Optimize the company’s production.

Constraints: g = 20x + 2000y = 20 000

f = R = 100x^2/3 y^1/3

x + 100y = 1 000
g = 20 000 x + 100y = 1 000
x^-1/3 y^1/3 = 3/10
fx = gx 100* 2/3x^-1/3 y^1/3 = 20 =
x^2/3 y^-2/3 = 60
fy = gy 100* 1/3y^-2/3 x^2/3 = 2 000

Eliminate the by dividing  X^2/3-(-1/3) y^-2/3-1/3 = 60 * 1/3 x + 100y = 1000

x/y = 200  x + x/2 = 1000

 x = 2000/3
x = 200y  y = 10/3

5. Demand function of 2 products:

qA = e^-pA-pB qB = 20 / pApB

qA/ qB = - e^-pA-pB qB/ qA = -20 / pA^2pB

qA/ qB < 0 qB/ qA < 0

Complimentary products
6. Find critical points:

a. g = x^2 + y^2 + xz + yz

gx = 2x + z = 0 x–y=0

gy = 2y + z = 0 x=y=z=0 ( 0, 0, 0 )

gz = x + y = 0  x+y=0

2 0 1 Det1 = 1
H=
0 2 1 Det2 = 4 Therefore a saddle point

1 1 0 Det3 = 2 (-1) + 1 (-2) = - 4

b. f = 4x^2 e^y - 2x^4 – e^4y

fx = 8xe^y – 8x^3 = 0 x ( e^y – x^2) = 0

e^y (x^2 – e^3y) = 0

fy = 4x^2e^y – 4e^4y = 0

x^2 = e^3y x = 0  e^3y = 0 DNE impossible

x^2 = e^y  e^3y = e^y  e^2y = 1 y=0, x^2=1

So the points would be ( 1 , 0 ) and ( -1 , 0 )

8-24 8 -16 -8
H(1,0) = =
fxx = 8e^y – 24x^2
8 4-16 8 -12
fyy = 4x^2 e^y – 16e^4y

fxy = 8xe^y
8-24 -8 -16 -8
H(-1,0) = =
-8 4-16 -8 -12
det (1,0) = ( 8 – 24 )( 4 – 16 ) – 8^2

= (-16)(-12) – 64 = 128 > 0

Local Max
det(-1,0) = ( 8 – 24 )( 4 – 16 ) – (-8)^2

= (-16)(-12) – 64 = 128 > 0

fxx (x,y ) < 0

Good Luck!!