Chapter 4.

Partial Di↵erentiation

Outline of the chapter:

Functions of several variables

Unconstrained/constrained optimization

Partial elasticity and marginal functions

International School of Business, UEH UEH, Vietnam November 17, 2022 114 / 200
 Chapter 4. Partial Di↵erentiation 4.1. Functions of several variables

Most relationships in economics involve more than one variable. For

instance, the demand for a good depends not only on its own price but
also on the price of substitute and complementary goods, incomes of
consumers, advertising expenditure and so on. Likewise, the output from a
production process depends on a variety of inputs, including land, capital
and labor.

Note that
Substitutes are products/services in di↵erent forms that have the same
functionality. For example, butter and margarine are substitute goods.

Complementary goods/services is an item used in conjunction with another

good/service. Usually the complementary good has little to no value when
consumed alone, but when combined with another goods/services, it adds to
the overall value of the o↵ering. For example, an iPhone and apps with it
are complementary products.
International School of Business, UEH UEH, Vietnam November 17, 2022 115 / 200
 Chapter 4. Partial Di↵erentiation 4.1. Functions of several variables

To analyze general economic behavior we must extend the concept of a function

of one variable, and particularly the di↵erential calculus, to functions of several
variables.
Definition
A function, f , of n variables is a rule that assigns to each incoming tuple of
numbers, (x1 , x2 , . . . , xn ), a uniquely defined outgoing number, z. In symbols we
write this as z = f (x1 , x2 , . . . , xn ).

Domain of a function of n variables is defined by

{(x1 , x2 , . . . , xn )|f (x1 , x2 , . . . , xn ) is well defined}
a subset of IR ⇥ IR ⇥ · · · ⇥ IR ⌘ IRn called the n dimensional space. In particular, if
n = 2 then it is a two dimensional space or a real plane.
International School of Business, UEH UEH, Vietnam November 17, 2022 116 / 200
 Chapter 4. Partial Di↵erentiation 4.1. Functions of several variables

Graph of a function involving two variables is a surface in three

dimensional space. It is necessary to bear in mind that the graph of
function of one variable is a curve in xy plane.

To help you have a feeling of what is the graph of a function of two

variables, let us analyze how to sketch the graph of the following functions.
Example
Graph of the function f (x, y ) = x 2 + y 2 (D is the whole xy plane)

International School of Business, UEH UEH, Vietnam November 17, 2022 117 / 200
 Chapter 4. Partial Di↵erentiation 4.1. Functions of several variables

First-order derivatives.
Definition
First-order partial derivative w.r.t. x:
@f f (x0 + x,y0 ) f (x0 ,y0 )
@x (x0 , y0 ) ⌘ fx (x0 , y0 ) = lim x!0 x

First-order partial derivative w.r.t. y :

@f f (x0 ,y0 + y ) f (x0 ,y0 )
@y (x0 , y0 ) ⌘ fy (x0 , y0 ) = lim y !0 y

@f
Accordingly to compute for example @x , we need only to di↵erentiate
mathematical expressions with respect to x and let all remaining letters in
the mathematical expressions constants in disguise. It is the same for
other variables.

Gradient vector of f at (x0 , y0 ) is denoted rf and determined by

rf (x0 , y0 ) = [fx (x0 , y0 ); fy (x0 , y0 )]

International School of Business, UEH UEH, Vietnam November 17, 2022 118 / 200
 Chapter 4. Partial Di↵erentiation 4.1. Functions of several variables

Second-order derivatives.
@2f
Second-order partial derivative w.r.t. x: @x 2
⌘ fxx = (fx )x
@2f
Second-order partial derivative w.r.t. y : @y 2
⌘ fyy = (fy )y
@2f
Second-order partial derivative w.r.t. x and y : @x@y ⌘ fxy = (fy )x
@2f
Second-order partial derivative w.r.t. y and x: @x@y ⌘ fyx = (fx )y

Note that in general (fx )y 6= (fy )x . In the scope of this course, we will
consider functions that satisfy (fx )y = (fy )x . [According to Weierstrass
theorem, (fx )y (x0 , y0 ) = (fy )x (x0 , y0 ) if the first order partial derivatives
are continuous at (x0 , y0 ).]

International School of Business, UEH UEH, Vietnam November 17, 2022 119 / 200
 f- ✗ = 246-1 1) ✗ 2kt 5y3 4kg2-

f ,c=- =
4k -13g -15+0 = 4kcal -11)t5y3 -

4kg2
on
46h2 -1 1) +4N 22C
4y2
Fxx
-

= ✗

fy=¥y=3kt8y -16 = 12h2 -442-14

2
f- y = 3 ( y3 -
3)

Fxx __
(f)c) a
= 4

Fyy=(Fy)y=8
*
f- Cn ; g) = KY
AY
-2

YAY
-1

fxy=(fy)x =3 f- a = ; f××=y(y -

1) ✗

KY
RYxlnkxtnk-WYxlrinfxy-ffyla-KY-1-iy.it
lmk
Fyn =
a) y =3 fy =
✗
;
fyy= 1- 1-
enx=xY Hyena)
- -

✗
(

Fyn =
a) y=yxxY¥lnr×xY×a1
3g xY- 1- ( ylnie -11 )
fa= 4kt →
choy -_ const =

3N
fy = → chose = const

Fxx =
Cfa )✗= 4

Fyy = (
Fy )y=0
=3
foxy =

g) ✗
( =3
Fyn =
freely

ga= 12k3g +
2y2 5g
-
-18

324 +
Any g- 5k
-

gy =

Gorge = 36k2g
4k
8yy
=

(gy )q= 12h3 -14g 5

Gay
-

( 12ns -14g
ga)y
-5
Gyu
= =
 Chapter 4. Partial Di↵erentiation 4.1. Functions of several variables

Example. Use the small increments formula to estimate the change in

z = x 2y 4 x 6 + 4y

when
(a) x increases from 1 to 1.1 and y remains fixed at 0;

(b) x remains fixed at 1 and y decreases from 0 to -0.5;

(c) x increases from 1 to 1.1 and y decreases from 0 to -0.5.

International School of Business, UEH UEH, Vietnam November 17, 2022 121 / 200
 Chapter 4. Partial Di↵erentiation 4.2. Optimization problems

Recall that a local/relative minimum point is a point whose value is

smallest compared to the value of points around. It is analogous to
definition of relative/local maximum point.

A global/absolute minimum point is a point whose value is smallest

compared to every points in the domain.

International School of Business, UEH UEH, Vietnam November 17, 2022 122 / 200
 Chapter 4. Partial Di↵erentiation 4.2. Optimization problems

Let’s look at the three particular surfaces:

Example
Graph of the function f (x, y ) = x 2 + y 2 has the shape of a paraboloid
opened upwards.

(x = 0, y = 0) is called a minimum of f . Note that the tangent plane to

the surface at the minimum point is the xy plane and the surface is
convex, which means it holds water in itself.

International School of Business, UEH UEH, Vietnam November 17, 2022 123 / 200
 Chapter 4. Partial Di↵erentiation 4.2. Optimization problems

Example
Graph of the function f (x, y ) = x2 y 2 has the shape of a paraboloid
opened downwards

(x = 0, y = 0) is called the maximum of f . Note that the tangent plane to

the surface at the maximum point is the xy plane and the surface is
concave, which means it pours water.

International School of Business, UEH UEH, Vietnam November 17, 2022 124 / 200
 Chapter 4. Partial Di↵erentiation 4.2. Optimization problems

Example
Graph of the function f (x, y ) = x 2 y 2 has the shape of a hyperbolic
paraboloid

(x = 0, y = 0) is called the saddle point or minimax point of f . This

surface is neither convex nor concave at (0, 0).

International School of Business, UEH UEH, Vietnam November 17, 2022 125 / 200
 Chapter 4. Partial Di↵erentiation 4.2. Optimization problems

Steps to solve the unconstrained problem Min/Max z = f (x, y )

S1. [Fermat’s necessary condition] Find the stationary point/candidates (x0 , y0 )
of the function f by solving the simultaneous system fx = fy = 0
S2. [Sufficient condition investigating the concavity of the function] Classify the
candidates as minimum or maximum or saddle points by determining
= fxx (x0 , y0 ) ⇥ fyy (x0 , y0 ) fxy2 (x0 , y0 ). If
> 0 then (x0 , y0 ) is an extreme point. This point is the local
minimum if fxx (x0 , y0 ) > 0 and is the local maximum if fxx (x0 , y0 ) < 0.
Note that if does not consist of (x0 , y0 ) then minimum/maximum
obtained will be global.
< 0 then (x0 , y0 ) is the saddle point or so called minimax, that is,
neither min nor max.
= 0 then no conclusion is stated yet. To get further information, we
have to consider = f (x, y ) f (x0 , y0 ). If
 0 for all x, y then (x0 , y0 ) is a maximum
0 for all x, y then (x0 , y0 ) is a minimum
Otherwise, (x0 , y0 ) is a saddle point.
International School of Business, UEH UEH, Vietnam November 17, 2022 126 / 200
 Chapter 4. Partial Di↵erentiation 4.2. Optimization problems

Example. [Price discrimination] Suppose a monopolist is practicing

price discrimination in the sale of a product by charging di↵erent prices in
two separate markets.

In market A, the demand function is pA = 100 qA and in market B it is

pB = 84 qB , where qA and qB are the quantities sold per week in A and
B, and pA and pB are the respective prices per unit.

If the monopolist’s cost function is C = 600 + 4(qA + qB ), how much

should be sold in each market to maximize the profit?

International School of Business, UEH UEH, Vietnam November 17, 2022 127 / 200
 Slide 127 :

Irofit I =
[ PA ✗
9A +
PB ✗
9ps ] [600-1
-
4 ( 9A -19ps )
]
I[
9A ; 9 B) = 400 -9*7×9*-1 4- 9 B) ×
9ps
-
600 -4 (
9*-1 9ps
)

step 1 100-29*-4=0 48
{ Pa { 9A
: = =
*

84 40
P9B 29ps ¢ ◦
9ps
-
-

= = =

Step 2 -2 's
PQBQB & °
-

pqaq # PaA9B=
: = =

D= C- 2) × C- 2) -
02 = 470

Hence maximized when the market A sells 48 units and the market
profit is

B. sells 40 units .
 step 2 Fxx : = 4>9 fyy =3 foxy=3 ;

step 1 : Find the critical points D= 4×3 -

32=370
( called the ( 1 ; 2) is local minimum
or so
stationary points a
;
)
mi mum value is f ( 1 ; 2) = 49
OR the candiates
by solving the
system
-4N -18
for 2×-6=0 " =3
fx= 0 "= 2
{ { {
=

{y
=

4
fy =
8y -32=0 = 4
fy =
-6g + 24=0
y =

( =3 4) is the critical (x 2 4) is the critical

x
; y = =
; y =

point point
the critical
step 2 :
classify point fax = -4
;
fyy =
- 6
;
fay = 0

/
as a min Max
of Miramax D= 2470

fxx & 8
fjox <0 ( 2; 4) is a local
Fyy
'

=
=
;
and the value
foxy =Cfy)x= 0 maxi mi mum Max is 63

D=
Fxx fyy fg{y= 2×8-0 ✗ -

= 1670

(3; 4) is a local minimum and

foe =-3 122=0 14
{
"
-

{y
=

f- ( 3; 4)
1
min
f- ( x; y ) fy 2g 0 42
-

= = = =

fxx fay
-12
;
fyy -2s 0
= = =

step
Fx 4K
-13g 10=0
{
-

-13k -9=0
fy =
3g

{ 3k-13g
4k 10 " = 1-

{Y
=

g
-13g = = 2

Ck 1 2) is the critical point

; y
= =
 Chapter 4. Partial Di↵erentiation 4.2. Optimization problems

Example. [Constraint problem] A firm produces two goods A and B.

The weekly cost of producing x items of A and y items of B is

TC (x, y ) = 0.2x 2 + 0.05y 2 0.1xy + 2x 5y + 1000

(a) State the minimum value of TC in the case when there are no
constraints.

(b) Find the minimum value of TC when the firm is committed to

producing 500 goods of either type in total.

International School of Business, UEH UEH, Vietnam November 17, 2022 128 / 200
 Chapter 4. Partial Di↵erentiation 4.2. Optimization problems

Lagrange multiplier method for constraint problems.

Steps to find the optimal value of Min/Max z = f (x, y ) s.t. g (x, y ) = C

S1. Construct the Lagrangian function L(x, y , ) = f (x, y ) ⇥ [g (x, y ) C ],
in which is called the Lagrange multiplier.
S2. Optimal conditions: rf = rg [, fx = gx ; fy = gy ] and g (x, y ) C = 0.
Therefore, to find the optimal value we need to solve simultaneous system
Lx = Ly = L = 0 for (x0 , y0 ) [and if necessary].
Example. Use Lagrange multiplier method to optimize the function z = 2x 2 xy
subject to x + y = 12.
International School of Business, UEH UEH, Vietnam November 17, 2022 129 / 200
 Chapter 4. Partial Di↵erentiation 4.3. Partial elasticity and marginal functions

Our aim in this section is to study several applications of partial

di↵erentiation in economics that have been considered in the case of
functions of one variable, such as

Substitute commodities and Complementary commodities

Elasticity of demand

Marginal utility and marginal production and so on.

International School of Business, UEH UEH, Vietnam November 17, 2022 130 / 200
 Chapter 4. Partial Di↵erentiation 4.3. Partial elasticity and marginal functions

Substitute/complementary commodities. It’s worth recalling that two

commodities are said to be substitute commodities if an increase in the demand
for either results in a decrease in demand for the other. Substitute commodities
are competitive, like butter and margarine.

On the other hand, two commodities are said to be complementary commodities

if a decrease in the demand of either results in a decrease in the demand of the
other. An example is provided by digital cameras and recording disks. If
consumers buy fewer digital cameras, they will likely buy fewer recording disks,
too.

Suppose D 1 (p1 , p2 ) units of the first commodity and D 2 (p1 , p2 ) of the second are
demanded when the unit prices of the commodities are p1 and p2 , respectively. It
is reasonable to expect demand to decrease with increasing price, so Dp11 < 0 and
Dp22 < 0.

International School of Business, UEH UEH, Vietnam November 17, 2022 131 / 200
 Chapter 4. Partial Di↵erentiation 4.3. Partial elasticity and marginal functions

For substitute commodities, the demand for each commodity increases with
respect to the price of the other, so Dp12 > 0 and Dp21 > 0.

However, for complementary commodities, the demand for each decreases with
respect to the price of the other, and Dp12 < 0 and Dp21 < 0.

Example
Suppose the demand function for flour in a certain community is given by
10
D 1 (p1 , p2 ) = 500 + 5p2 ,
p1 + 2
while the corresponding demand for bread is given by
7
D 2 (p1 , p2 ) = 400 + 2p1 ,
p2 + 3
where p1 is the dollar price of a pound of flour and p2 is the price of a loaf of
bread. Determine whether flour and bread are substitute or complementary
commodities or neither.
International School of Business, UEH UEH, Vietnam November 17, 2022 132 / 200
 Chapter 4. Partial Di↵erentiation 4.3. Partial elasticity and marginal functions

Elasticity. Suppose that demand function Q = f (P, PA , Y ) in which P is

the price of a certain good, PA is the price of an alternative good, and Y
is the income of consumers. So
P @Q P
The price elasticity of demand is EP = Q ⇥ @P ⌘ Q ⇥ QP .
PA @Q PA
The cross-price elasticity of demand is EPA = Q ⇥ @PA ⌘ Q ⇥ QP A .

If the alternative good is substitutable then QPA > 0 and thus EPA > 0.

If the alternative good is complementary then QPA < 0 and thus

EPA < 0.
Y @Q Y
The income elasticity of demand is EY = Q ⇥ @Y ⌘ Q ⇥ QY .

If the good is inferior then QY < 0 and thus EY < 0.

If the good is normal then QY > 0 and thus EY > 0. Furthermore, if

EY > 1 then the good is superior.

International School of Business, UEH UEH, Vietnam November 17, 2022 133 / 200
 Chapter 4. Partial Di↵erentiation 4.3. Partial elasticity and marginal functions

Example. Given the demand function Q = 100 2P + PA + 0.1Y where

the price of a certain good P = 10, the price of an alternative good
PA = 12, and the income of consumers Y = 1000.
(1) Find
price elasticity of demand

cross-price elasticity of demand

income elasticity of demand

(2) Is the alternative good is substitutable or complementary?

(3) If income rises by 5%, calculate the corresponding percentage change

in demand. Would this good be classified as inferior, normal or
superior?

International School of Business, UEH UEH, Vietnam November 17, 2022 134 / 200
 Chapter 4. Partial Di↵erentiation 4.3. Partial elasticity and marginal functions

Utility. Consumers are faced with a choice of how many items of various goods to
buy and has a preference between options available. To analyze the behavior of
consumers quantitatively, we associate with each set of options a number, U,
called utility, which indicates the level of satisfaction. For instance,

If a customer prefers apples to oranges no matter the amount consumed, the

utility function could be expressed as U(apples) > U(oranges).

If a customer prefers (1 avocado and 3 apples) to (2 avocados and 1 apple),

the utility function could be expressed as U(1, 3) > U(2, 1).

A little more general, suppose that there are two goods, G1 and G2 , and that the
consumer buys x1 items of G1 and x2 items of G2 . The variable U is then a utility
function of x1 and x2 , which we write as U = U(x1 , x2 )

International School of Business, UEH UEH, Vietnam November 17, 2022 135 / 200
 Chapter 4. Partial Di↵erentiation 4.3. Partial elasticity and marginal functions

Suppose that utility is a function of two variables x1 and x2 and if x1

changes by a small x1 and x2 changes by a small x2 , then the net
change in U can be found from the the small increments formula:

U ⇡ U x 1 x1 + U x 2 x2
1/4 3/4
Example. Given the utility function U(x1 , x2 ) = x1 x2 . Determine the
values of marginal utilities Ux1 and Ux2 when x1 = 100, x2 = 200. Hence
estimate the change in utility when x1 decreases from 100 to 99 and x2
increases from 200 to 201.

In this example, we will evidence that Ux1 x1 < 0 and Ux2 x2 < 0. This
illustrates the law of diminishing marginal utility, meaning the
consumption of good G1 or G2 increases, each additional item of G1 or G2
bought confers less utility than the previous item.

International School of Business, UEH UEH, Vietnam November 17, 2022 136 / 200
 Chapter 4. Partial Di↵erentiation 4.3. Partial elasticity and marginal functions

Marginal rate of commodity substitution (MRCS). If two commodities are

substitutable, then in order to maintain a constant value of utility when x1
decreases by x1 units, it is necessary that x2 increases by a corresponding
quantity x2 such that U ⇡ Ux1 x1 + Ux2 x2 = 0.
x2 Ux 1
It turns out that MRCS ⌘ x1 = Ux 2

If fewer purchases are made of G1 , then the consumer has to compensate for this
by buying more of type G2 to maintain the same level of satisfaction.

Indi↵erence/level curve is a curve that consists all points of the same utility:
C (U, U0 ) = {(x1 , x2 )|U(x1 , x2 ) = U0 }
International School of Business, UEH UEH, Vietnam November 17, 2022 137 / 200
 Chapter 4. Partial Di↵erentiation 4.3. Partial elasticity and marginal functions

p p
Example. Given the utility function U = x1 x2 find a general
expression for MRCS in terms of x1 and x2 .
(a) Calculate the particular value of MRCS for the indi↵erence curve that
passes through (300, 500).

(b) Hence estimate the increase in x2 required to maintain the current

level of utility when x1 decreases by 3 units.

International School of Business, UEH UEH, Vietnam November 17, 2022 138 / 200
 Chapter 4. Partial Di↵erentiation 4.3. Partial elasticity and marginal functions

Production. The output, Q, of any production process depends on a

variety of inputs, known as factors of production. These comprise land,
capital, labor and enterprise.

For simplicity we restrict our attention to capital and labor. Capital, K ,

denotes all man-made aids to production such as buildings, tools and plant
machinery. Labor, L, denotes all paid work in the production process. The
dependence of Q on K and L may be written

Q = f (K , L)

which is called the production function.

Example
If Q = 100K 1/3 L1/2 then the inputs K = 27 and L = 100 lead to an
output
Q = 100(27)1/3 (100)1/2 = 100(3)(10) = 3000
International School of Business, UEH UEH, Vietnam November 17, 2022 139 / 200
 Chapter 4. Partial Di↵erentiation 4.3. Partial elasticity and marginal functions

Production functions can be studied in a similar way to utility functions.

So we also have Q ⇡ @Q @Q
@K ⇥ K + @L ⇥ L ⌘ QK ⇥ K + QL ⇥ L

Marginal rate of technical substitution (MRTS) is the amount by

which capital needs to rise to maintain a constant level of output when
K QL
labour decreases by 1 unit: MRTS ⌘ L = QK .

Example
p
Given Q = 2LK + L. If the current level of K and L are 7 and 4,
respectively, estimate the increase in capital needed to maintain the
current level of output given a 1-unit decrease in labour.

International School of Business, UEH UEH, Vietnam November 17, 2022 140 / 200
 Chapter 4. Partial Di↵erentiation 4.3. Partial elasticity and marginal functions

Homogeneous function. A production function f (K , L) is said to be homogeneous

of degree n if, for any number ,
n
f ( K , L) = f (K , L)

A production function is then said to display decreasing returns to scale, constant

returns to scale or increasing returns to scale, depending on whether < 1, = 1
or > 1, respectively. One useful result concerning homogeneous functions is
known as Euler’s formula, which states that
@f @f
K +L = nf (K , L)
@K @L

The special case n = 1 is worthy of note because the right-hand side of the
formula is then simply f (K , L), which is equal to the output, Q.

Euler’s formula for homogeneous production functions of degree 1 states that if

each input factor is paid an amount equal to its marginal product then each term
on the left-hand side gives the total bill for that factor.
International School of Business, UEH UEH, Vietnam November 17, 2022 141 / 200
 Chapter 4. Partial Di↵erentiation 4.3. Partial elasticity and marginal functions

The Cobb—Douglas production function given by Q = AK ↵ ⇥ L , in

which A, ↵, are positive constants, is an homogeneous function of degree
↵+ .

The Cobb-Douglas production function is based on the empirical study of

the American manufacturing industry made by Cobb-Douglas. It is widely
used to represent the technological relationship between only two inputs,
labour and capital.

International School of Business, UEH UEH, Vietnam November 17, 2022 142 / 200
 Chapter 5. Integration

Outline of the chapter:

Indefinite integration

Definite integration

Applications in Business

Probability [for your further reference]

Consumers’ surplus - Producers’ surplus

Continuous money flow

Discounting

International School of Business, UEH UEH, Vietnam November 17, 2022 143 / 200
 Chapter 5. Integration 5.1. Indefinite integration

Definition
If F 0 (x) = f (x) for all x in an interval I then we say that F (x) is an
anti-derivative of f (x) on I .

Theorem
If F is an anti-derivative of f then all other anti-derivative have the form
F (x) + C where C is a real number.

International School of Business, UEH UEH, Vietnam November 17, 2022 144 / 200
 Chapter 5. Integration 5.1. Indefinite integration

Definition
If the function f has an anti-derivative then the integral of f denoted by
R R
f or f (x)dx is the collection of all anti-derivatives of f .

If F is an anti-derivative of f then the integral of f is often written as

Z
f (x)dx = F (x) + C

where C is any real number.

R ⇣ 3 ⌘0
x3 x
Example. x 2 dx = 3 + C because 3 = x 2.

International School of Business, UEH UEH, Vietnam November 17, 2022 145 / 200
 Chapter 5. Integration 5.2. Definite integration

To define the definite integral, let’s study the following area problem.

Problem. Find the area of the region S enclosed by the curve y = x 2 , the x-axis,
and the vertical lines x = 0 and x = 1.

One of the ways to find the area of this region is approximating each strip by a
small rectangle. As we have seen in the picture below, the more the number of
sub-intervals, the better the approximation.

International School of Business, UEH UEH, Vietnam November 17, 2022 146 / 200
 Chapter 5. Integration 5.2. Definite integration

Area between curves. Consider the region S enclosed by the curve

y = f (x) 0, the x-axis, and the vertical lines x = a and x = b, where
f (x) is continuous function over x 2 [a, b]

International School of Business, UEH UEH, Vietnam November 17, 2022 147 / 200
 Chapter 5. Integration 5.2. Definite integration

The area of the indicated region can be calculated by performing the

following steps (Rectangular Approximation Method, or RAM for short).
1. We first divide up [a, b] into n subintervals of equal length by n + 1
points
x 0 = a < x1 < x2 < · · · < xn = b

2. Set
b a
x ⌘ xi xi 1 = , 8i = 1, 2, . . . , n
n
3. We take any point xi⇤ 2 [xi 1 , xi ], 8i = 1, . . . , n
4. Area of a rectangle of width x and height f (xi⇤ ) is f (xi⇤ ) x
P
5. Total areas of rectangles = ni=1 f (xi⇤ ) x [is called Riemann sum].
6. Exact area
P P Rb
= limn!1 ni=1 f (xi⇤ ) x ⌘ n=1 ⇤
i=1 f (xi ) x ⌘ a f (x) dx

International School of Business, UEH UEH, Vietnam November 17, 2022 148 / 200
 Chapter 5. Integration 5.2. Definite integration

Theorem (Fundamental theorem of Calculus)

If f is continuous on [a, b] and F is an anti-derivative of f then
Z b
f (x)dx = F (x)|ba = F (b) F (a) [Newton-Leibnitz formula]
a

In summary, to compute a definite integral we can possibly use definition or

theorem.
Rb Pn=1
[by definition] a
f (x) dx = i=1 f (xi⇤ ) x

b a
where x0 = a, xi = a + i ⇥ x, xn = b; x = xi xi 1 = ; xi⇤ 2 [xi 1 , xi ]
n
Rb
[by theorem] a
f (x)dx = F (x)|ba = F (b) F (a)

R1
Example. Let compute the integral 0 x 2 dx by both the Riemann sum and the
Fundamental theorem, and then compare results obtained.
International School of Business, UEH UEH, Vietnam November 17, 2022 149 / 200
 Chapter 5. Integration 5.3. Applications in Business

Probability of a continuous random event

The probability density function or PDF of a continuous random variable gives
the relative likelihood of any outcome in a continuum occurring. Unlike the
case of discrete random variables, for a continuous random variable any single
outcome has probability zero of occurring.

Probability that a random variable X takes a value in the open or close interval
[a, b] is given by the integral of a function called probability density function fX
Z b
P(a  X  b) = fX (x)dx and P(X = c) = 0 for any c 2 [a, b]
a

International School of Business, UEH UEH, Vietnam November 17, 2022 150 / 200
 Chapter 5. Integration 5.3. Applications in Business

Consumers’ surplus.
Given a demand function P = f (Q). At Q = Q0 the price P = P0 .
P0 is the price that consumers are prepared to pay for the last unit that they buy,
which is the Q0 th good. For quantities up to Q0 they would actually be willing to
pay the higher price given by the demand curve. The consumer’s surplus is the
shaded area which represents the benefit to the consumer of paying the fixed
price of P0
Z Q0 Z Q0
CS = f (Q)dQ P0 ⇥ Q0 = PdQ P0 ⇥ Q0
0 0

International School of Business, UEH UEH, Vietnam November 17, 2022 151 / 200
 Chapter 5. Integration 5.3. Applications in Business

Producers’ surplus.
Given a supply function P = g (Q). Assuming that all goods are sold, the total
amount of money received is then Q0 P0 , where P0 is the price at which the
producer is prepared to supply the last unit, which is the Q0 th good. For
quantities up to Q0 they would actually be willing to accept the lower price given
by the supply curve.
The producer’s surplus is the shaded area representing the benefit to the
producer of selling at the fixed price of P0 :
Z Q0 Z Q0
PS = P0 ⇥ Q0 g (Q)dQ = P0 ⇥ Q0 PdQ
0 0

International School of Business, UEH UEH, Vietnam November 17, 2022 152 / 200
 Chapter 5. Integration 5.3. Applications in Business

Example 1. The demand equation for a product is q = 400 p 2 and the

supply equation is p = q/60 + 5. Find the producers’ surplus and
consumers’ surplus under the market equilibrium.

Example 2. The supply equation for a product is given by the following

table, where p is the price per unit (in dollars) at which q units are
supplied to the market.

Use the trapezoidal rule to estimate the producers’ surplus if the selling
price is 80.

International School of Business, UEH UEH, Vietnam November 17, 2022 153 / 200
 Chapter 5. Integration 5.3. Applications in Business

Continuous money flow

In an earlier chapter we looked at the concepts of present value and future value
when a lump sum of money is deposited in an account and allowed to accumulate
interest. In some situations, however, money flows into and out of an account
almost continuously over a period of time. Examples include income in a store,
bank receipts and payments, and highway tolls. Although the flow of money in
such cases is not exactly continuous, it can be treated as though it were
continuous, with useful results.

In this section, we introduce the following features:

Total money flow.

Present value of money flow.

Accumulated amount of money flow.

International School of Business, UEH UEH, Vietnam November 17, 2022 154 / 200
 Chapter 5. Integration 5.3. Applications in Business

Total money flow

If I (t) is the income at time t then the total income/money flow over [0, T ] is
n=1
X n=1
X Z T Z T
[I (ti ) I (ti 1 )] = I 0 (ti 1 )⇥ t = I 0 (t)dt ⌘ I (T ) I (0) = f (t)dt
i=1 i=1 0 0

where f (t) represent the rate of flow of money per unit of time.
Notice that this “total money flow” does not take into account the interest the
money could earn after it is received. It is simply the total income.

Example. The income from a soda machine is growing exponentially. When the
machine was first installed, it was producing income at a rate of 1000 per year.
By the end of the 1st year, it was producing income at a rate of 1020.2 per year.
Find the total income produced during its first 3 years of operation.
International School of Business, UEH UEH, Vietnam November 17, 2022 155 / 200
 Chapter 5. Integration 5.3. Applications in Business

Present value of money flow

If f (t) represent the rate of the continuous flow at a discount rate r compounded
continuously for T years, the present value of a continuous money flow is
n=1
X n=1
X Z T
0 rti rti rt
P0 = [I (ti ) t] e = [f (ti ) t] e = f (t) ⇥ e dt
i=1 i=1 0

Example. A company expects its rate of annual income during the next three
years to be given by f (t) = 75, 000t, 0  t  3. What is the present value of this
income over the 3-year period, assuming a discount rate of 8% compounded
continuously?
International School of Business, UEH UEH, Vietnam November 17, 2022 156 / 200
 Chapter 5. Integration 5.3. Applications in Business

Accumulated amount of money flow

If f (t) is the rate of money flow at an annual rate r at time t, the
accumulated amount of money flow at time T is
Z T
rT rt
A=e f (t)e dt.
0

This amount includes interest received on the money after it comes in.
Example
If money is flowing continuously at a constant rate of 2000 per year over
5 years at 6% interest compounded continuously. Find
a. The total money flow over the 5-year period.

b. The accumulated amount of money flow, compounded continuously,

at time T = 5.
International School of Business, UEH UEH, Vietnam November 17, 2022 157 / 200
 Chapter 5. Integration 5.3. Applications in Business

Discounting. To calculate the present value, P0 , when a single future

value, P, is discounted at r % interest continuously for t years, we can use
the following formula: P0 = Pe rt

If the fund is to provide a continuous revenue stream for n years at an

annual rate of P dollars per year, then the present value can be found by
Rn
evaluating the definite integral P0 = 0 Pe rt dt

Example. Calculate the present value of a continuous revenue stream of

1,000 a year for 5 years if the discount rate is 9%.

International School of Business, UEH UEH, Vietnam November 17, 2022 158 / 200
 Chapter 6. Matrices

Outline of the chapter:

What is a matrix?

Basic matrix operations

Matrix inverse

Cramer’s rule

International School of Business, UEH UEH, Vietnam November 17, 2022 159 / 200
 Chapter 6. Matrices 6.1. What is a matrix?

A rectangular table consisting of m rows and n columns

2 3
a11 a12 . . . a1n
6 7
6 a21 a22 . . . a2n 7
A=6 .6 7
. 7
4 . 5
am1 am2 . . . amn

is called a matrix of order m ⇥ n.

Rows of a matrix are horizontal, columns of a matrix are vertical.

Notation. Am⇥n = [aij ], i = 1, . . . , m, j = 1, . . . , n in which aij is called

the ij entry. The set of matrices of order m ⇥ n is denoted Mm⇥n .

International School of Business, UEH UEH, Vietnam November 17, 2022 160 / 200
 Chapter 6. Matrices 6.1. What is a matrix?

In particular,
If m = 1 then A is also called row vector.
If n = 1 then A is called column vector.
If m = n then the matrix is said to be square and in this case, we call
A is the square matrix of order n.

Identity matrix of order n: aii = 1 for i = 1, . . . , n and aij = 0 for

i 6= j, i, j = 1, . . . , n
2 3
1 0 ... 0
6 7
60 1 . . . 07
In = 6
6 ..
7⌘I
7
4. 5
0 0 ... 1

International School of Business, UEH UEH, Vietnam November 17, 2022 161 / 200
 Chapter 6. Matrices 6.2. Matrix Operations

Let A = [aij ] and B = [bij ] be two matrices of order m ⇥ n, i.e.,

A, B 2 Mm⇥n .

Equality.
A = B , aij = bij

Addition/Subtraction.
A ± B = [aij ± bij ]

Scalar Multiplication. Let k be a real number.

k A = [k aij ]

Transposition. Let A = (aij ) be a matrix of order m ⇥ n. The transposition of A

denoted by AT is defined by symbolically AT = (aji ). This means that the first
row of AT is the first column of A, the second row of AT is the second column of
A, and so on. Thus A 2 Mm⇥n , while AT 2 Mn⇥m

International School of Business, UEH UEH, Vietnam November 17, 2022 162 / 200
 Chapter 6. Matrices 6.2. Matrix Operations

Scalar/Dot/Inner product. Given a = (a1 a2 . . . an ) and

b = (b1 b2 . . . bn ). The scalar product of a and b is defined by

n
X
ha, bi = ab T = a1 b1 + a2 b2 + · · · + an bn = ai bi
i=1

Example
a = (1 3 5), b = ( 1 0 4) ) ab T = 1 ⇥ ( 1) + 3 ⇥ 0 + 5 ⇥ 4 = 19

Recall that the product of 2 vectors perpendicular to each other is 0. For

instance

a = (1 0), b = (0 1) ) ab T = 0

International School of Business, UEH UEH, Vietnam November 17, 2022 163 / 200
 Chapter 6. Matrices 6.2. Matrix Operations

Matrix Multiplication. Let A be a matrix of order m ⇥ n and B be a

matrix of n ⇥ p. The product of A and B is

n
X
AB = [cij ] where cij = aik bkj
k=1

Note that we can only multiply two matrices when the number of columns
in the first matrix is equal to the number of rows in the second matrix.

" # " #
1 ⇥ 7 + 2 ⇥ 9 + 3 ⇥ 11 1 ⇥ 8 + 2 ⇥ 10 + 3 ⇥ 12 58 64
= =
4 ⇥ 7 + 5 ⇥ 9 + 6 ⇥ 11 4 ⇥ 8 + 5 ⇥ 10 + 6 ⇥ 12 139 154
International School of Business, UEH UEH, Vietnam November 17, 2022 164 / 200
 Chapter 6. Matrices 6.2. Matrix Operations

Observe that
" #" # " #
1 0 a11 a12 a11 a12
=
0 1 a21 a22 a21 a22
and
" #" # " #
a11 a12 1 0 a11 a12
=
a21 a22 0 1 a21 a22

In general, we have that

AIn = A = In A

International School of Business, UEH UEH, Vietnam November 17, 2022 165 / 200
 Chapter 6. Matrices 6.2. Matrix Operations

Tabulations are handy while storing the raw data as we will see in the
following examples.
Example
A firm orders 12, 30 and 25 items of goods G1, G2 and G3. The cost of
each item of G1, G2 and G3 is 8, 30 and 15, respectively.
(a) Write down suitable price and quantity vectors, and use matrix
multiplication to work out the total cost of the order.

(b) Write down the new price vector when the cost of G1 rises by 20%,
the cost of G2 falls by 10% and the cost of G3 is unaltered. Use
matrix multiplication to work out the new cost of the order and hence
find the overall percentage change in total cost.

International School of Business, UEH UEH, Vietnam November 17, 2022 166 / 200
 Chapter 6. Matrices 6.2. Matrix Operations

Example
A chain of sports shops, A, B and C , sells T-shirts, trainers and tennis racquets.
The weekly sales and profit per item are shown in the tables below:

The 3 ⇥ 3 matrices formed from the sales and profit tables are denoted by S and
P, respectively.

(a) If A = SP T , find the element a11 and give a brief interpretation of this
number.

(b) If B = S T P, find the element b33 and give a brief interpretation of this
number.
International School of Business, UEH UEH, Vietnam November 17, 2022 167 / 200
 Chapter 6. Matrices 6.3. Inverse Matrix

By multiplication matrix operation, we can easily see that

2 32 3 2 3
a11 a12 . . . a1n x1 b1
6 76 7 6 7
6 a21 a22 . . . a2n 7 6x2 7 6 b2 7
6 . 76.7=6 . 7
6 . 76.7 6 . 7
4 . 54.5 4 . 5
am1 am2 . . . amn xn bm
8
>
> a11 x1 + a12 x2 + · · · + a1n xn = b1
>
>
>
>
<a21 x1 + a22 x2 + · · · + a2n xn = b2
, ..
>
> .
>
>
>
>
:a + am2 x2 + · · · + amn xn
m1 x1 = bm
This means that a system of linear equations can be written under a matrix
equation AX = B. We know that the ordinary equation ax = b has a solution
x =a 1
b if a 6= 0, where a 1
= 1a . Now, being seeking solution of the matrix
1
equation above, we will define A .
International School of Business, UEH UEH, Vietnam November 17, 2022 168 / 200
 Chapter 6. Matrices 6.3. Inverse Matrix

Let A be a matrix of order n ⇥ n. If there exists a matrix B such that

AB = In = BA

then A is said to be invertible and B is called the inverse matrix of A,

denoted by A 1 .
Example
# " " #
1 2 2 1
Let a matrix A = . Verify that is the inverse of A.
3 4 3/2 1/2

International School of Business, UEH UEH, Vietnam November 17, 2022 169 / 200
 Chapter 6. Matrices 6.3. Inverse Matrix

How to find an inverse matrix?

1. Write an augmented matrix consisting of A on the left side and the
identity matrix on the right side
2 3
a11 a12 . . . a1n | 1 0 . . . 0
6 7
6a21 a22 . . . a2n | 0 1 . . . 07
6 . 7
6 . 7
4 . 5
an1 an2 . . . ann | 0 0 . . . 1
2. Then we attempt to transform the augmented matrix by row
equivalent operations called Gaussian elimination method to one of
the form 2 3
1 0 . . . 0 | b11 b12 . . . b1n
6 7
60 1 . . . 0 | b21 b22 . . . b2n 7
6. 7
6. 7
4. 5
0 0 ... 1 | bn1 bn2 . . . bnn
International School of Business, UEH UEH, Vietnam November 17, 2022 170 / 200
 Chapter 6. Matrices 6.3. Inverse Matrix

Gaussian elimination method

1. Interchange any two rows

8 8
<3x + 4y = 5 <x + 2y = 1
,
:x + 2y = 1 :3x + 4y = 5

2. Multiply each entry in a row by the same nonzero constant

8 8
<x + 2y = 1 <2x + 4y = 2
,
:3x + 4y = 5 :3x + 4y = 5

3. 8
Add a nonzero multiple of8one row to another row.
<x + 2y = 1 <x + 2y = 1
,
:3x + 4y = 5 :0x 2y = 8

International School of Business, UEH UEH, Vietnam November 17, 2022 171 / 200
 Chapter 6. Matrices 6.3. Inverse Matrix

Example. Solve the following system

8
>
> x + 3y + 2z =9
>
<
x y + 3z = 16
>
>
>
:3x 4y + 2z = 28

(a) by the Gaussian elimination method,

(b) by the inverse matrix.

International School of Business, UEH UEH, Vietnam November 17, 2022 172 / 200
 Chapter 6. Matrices 6.3. Inverse Matrix

How to check a matrix invertible or not? one of the ways is finding its
determinant. If the determinant of a matrix is non-zero then the matrix is
invertible.
Matrix of order 2 ⇥ 2

a b
= ad cb
c d
Matrix of 3 ⇥ 3:
a11 a12 a13
a21 a22 a23 =
a31 a32 a33

a22 a23 a21 a23 a21 a22

a11 - a12 + a13
a32 a33 a31 a33 a31 a32

International School of Business, UEH UEH, Vietnam November 17, 2022 173 / 200
 Chapter 6. Matrices 6.3. Inverse Matrix

Now the determinant of matrix of order n ⇥ n is defined inductively as

follows

For a square matrix A = [aij ]. The cofactor Aij of an element aij is

Aij = ( 1)i+j Mij ,

where Mij is the determinant of the matrix formed by deleting the ith row
and the jth column of A.

The determinant of A of n ⇥ n is defined by

n
X
|A| = aij Aij ,
j=1

where i 2 1, n.

International School of Business, UEH UEH, Vietnam November 17, 2022 174 / 200
 Chapter 6. Matrices 6.3. Inverse Matrix

Steps to find determinant of a square matrix

For any square matrix A of order n ⇥ n (n > 1), we define the determinant
of A, denoted |A|, as follows:
1. Choose any row or column

2. Multiply each element in that row or column by its cofactor and add
the results
Note that

The determinant of a 1 ⇥ 1 matrix is simply the element of the matrix.

The value of a determinant will be the same no matter which row or column
is chosen.

International School of Business, UEH UEH, Vietnam November 17, 2022 175 / 200
 Chapter 6. Matrices 6.4. Cramer’s rule

Solving systems of linear equations using determinants

Cramer’s rule for 2 ⇥ 2 systems

The solution of the system

8
<a x + b y = c
1 1 1
:a x + b y = c
2 2 2

Dx Dy
is given by x = ;y=
D D

a1 b1 c1 b 1 a 1 c1
where D = ; Dx = Dy = ; and D 6= 0
a2 b2 c2 b 2 a 2 c2

International School of Business, UEH UEH, Vietnam November 17, 2022 176 / 200
 Chapter 6. Matrices 6.4. Cramer’s rule

Cramer’s rule for 3 ⇥ 3 systems

The solution of the system

8
>
> a x + b1 y + c1 z = d 1
>
< 1
a2 x + b2 y + c2 z = d 2
>
>
>
: a3 x + b3 y + c3 z = d 3

Dx Dy Dz
is given by x = ;y= ;z=
D D D
a1 b1 c1 d 1 b1 c1 a1 d 1 c1
where D = a2 b2 c2 ; Dx = d2 b2 c2 Dy = a2 d2 c2 ;
a3 b3 c3 d 3 b3 c3 a3 d 3 c3
a1 b 1 d1
Dz = a2 b2 d2 ; and D 6= 0
a3 b 3 d3
International School of Business, UEH UEH, Vietnam November 17, 2022 177 / 200
 Chapter 6. Matrices 6.4. Cramer’s rule

As for n ⇥ n systems, we do the same. Cramer’s rule in general case

gives

D1 D2 Dn
x1 = , x2 = , . . . , xn =
D D D

provided that D 6= 0.

International School of Business, UEH UEH, Vietnam November 17, 2022 178 / 200
 Chapter 6. Matrices 6.4. Cramer’s rule

Example. Solve the following system

8
>
> x + 3y + 2z =9
>
<
x y + 3z = 16
>
>
>
:3x 4y + 2z = 28
by the Cramer rule.

International School of Business, UEH UEH, Vietnam November 17, 2022 179 / 200
 Chapter 6. Matrices 6.5. Some solvable methods for a system of linear equations

Some solvable methods a system of linear equations

(1) We can solve a general system of m linear equations and n variables by

Gaussian elimination method,

Reducing matrix using elementary row operations to row echelon form.

For example,

(2) For a system of n linear equations and n variables, we can possibly solve by
Using inverse matrix,

Using Cramer rule.

International School of Business, UEH UEH, Vietnam November 17, 2022 180 / 200
 Chapter 6. Matrices 6.5. Some solvable methods for a system of linear equations

Example. A firm produces 3 products A, B, and C that require processing

of 3 machines M1, M2, and M3. The time in hours required for processing
one unit of each product by the three machines is given by the following
table:

M1 is available for 440 hours, M2 for 310 hours, and M3 for 560 hours.
Find how many units of each product should be produced to make use of
all available time on the machines.

International School of Business, UEH UEH, Vietnam November 17, 2022 181 / 200
 Chapter 6. Matrices 6.5. Some solvable methods for a system of linear equations

Example. An investment company three types of pooled funds: Standard (S),

Deluxe (D), and Gold Star (G). Each unit of S contains 12 shares of stock A, 16
stocks B, and 8 of stock C. Each unit of D contains 20 shares of stock A, 12
stocks B, and 28 of stock C. Each unit of G contains 32 shares of stock A, 28
stocks B, and 36 of stock C.

Suppose an investor wishes to purchase exactly 220 shares of stock A, 176 shares
of stock B, and 264 shares of stock C by buying units of the three funds.

(a) Determine those combinations of units of S, D, and G that will meet the
investor’s requirement exactly.

(b) Suppose the investor pays 300 for each unit of S, 400 for each unit of D,
and 600 for each unit of G. Which of the combination from part (a) will
minimize the total cost to the investor?

International School of Business, UEH UEH, Vietnam November 17, 2022 182 / 200
 Chapter 7. Linear Programming

Outline of the chapter: In this chapter, we will introduce basically linear

programmings and methods for solving them.
What is a linear programming

Standard form of a linear programming

Geometric method for solving a linear programming of 2 variables

Simplex method for solving a linear programming

International School of Business, UEH UEH, Vietnam November 17, 2022 183 / 200
 Chapter 7. Linear Programming 7.1. What is a linear programming?

Let’s begin this chapter with an example in order to better understand what a
linear programming is?

Example. A manufacturer produces two models of racing bike, B and C, each of

which must be processed through two machine shops.

Machine shop 1 is available for 120 hours per month and machine shop 2 for 180
hours per month.

The manufacture of each bike of type B takes six hours in shop 1 and three hours
in shop 2. The corresponding times for C are 4 and 10 hours, respectively.

If the profit is 180 and 220 per bike of type B and C, respectively, how should
the manufacturer arrange production to maximize total profit?

International School of Business, UEH UEH, Vietnam November 17, 2022 184 / 200
 Chapter 7. Linear Programming 7.2. Standard form of a linear programming

Standard form of a linear programming

A linear programming is a problem to maximize or minimize a linear

function Z = c1 x1 + c2 x2 + · · · + cn xn subject to the following conditions
8
>
> a11 x1 + a12 x2 + · · · + a1n xn  b1
>
>
>
>
<a21 x1 + a22 x2 + · · · + a2n xn  b2
> ..
>
> .
>
>
>
:a x + a x + · · · + a x  b
m1 1 m2 2 mn n m

where aij are constants, bi are non-negative constants, cj are constants for
all 1  i  m, 1  j  n, and x1 , x2 , . . . , xn 0.

International School of Business, UEH UEH, Vietnam November 17, 2022 185 / 200
 Chapter 7. Linear Programming 7.3. Geometric method

Geometric method for linear programming of 2 variables

One of the ways to conduct the optimization problem of 2 variables, for

instance, max Z = P(x, y ) = 180x + 220y subject to
8
>
>  120
>6x + 4y
<
3x + 10y  180
>
>
>
:x, y 0.

is to used the geometric method as follows.

S1. Sketch the constraint set on the xy plane.

S2. Compare the value of Z at corners of the constraint set: the biggest
value is the optimal value of the max problem and the smallest value
is the optimal value of the min problem.
International School of Business, UEH UEH, Vietnam November 17, 2022 186 / 200
 Chapter 7. Linear Programming 7.3. Geometric method

Sketching linear inequalities

Graphing linear inequalities is nearly similar to doing with linear equation.

The di↵erence is that the solution to the inequality is not the drawn line
but the area of the coordinate plane that satisfies the inequality. So you
need to do as follows:
draw a line of equation. This line divides the coordinate plane into
two halves and is called boundary line. One side of the boundary line
contains all solutions to the inequality.

choose a test point not lying on the boundary line. If coordinates of

the test point satisfy the inequality then the solution set of the
inequality is the part which consisting the point, otherwise choose
other part as the solution set. The boundary line is dashed for > and
< and solid for and .
International School of Business, UEH UEH, Vietnam November 17, 2022 187 / 200
 Chapter 7. Linear Programming 7.3. Geometric method

The area shaded darkly in the figure is the set of all points (x, y ) satisfying
the inequality y  x + 2

International School of Business, UEH UEH, Vietnam November 17, 2022 188 / 200
 Chapter 7. Linear Programming 7.3. Geometric method

Example. An Italian restaurant o↵ers a choice of pasta or pizza meals. It

costs 3 to make a pasta dish, which it sells for 13. The corresponding
figures for pizzas are 2 and 10, respectively.

The maximum number of meals that can be cooked in a week is 1200 and
the restaurant has a weekly cost budget of 3000. How many pasta and
pizza dishes should be cooked each week to maximize profit?

International School of Business, UEH UEH, Vietnam November 17, 2022 189 / 200
 Chapter 7. Linear Programming 7.3. Geometric method

Example. A small firm manufactures and sells litre cartons of non-alcoholic

cocktails, ‘The Caribbean’ and ‘Mr Fruity’, which sell for 1 and 1.25,
respectively. Each is made by mixing fresh orange, pineapple and apple juices in
di↵erent proportions.

The Caribbean consists of one part orange, six parts pineapple and one part apple.
Mr Fruity consists of two parts orange, three parts pineapple and one part apple.

The firm can buy up to 300 litres of orange juice, up to 1125 litres of pineapple
juice and up to 195 litres of apple juice each week at a cost of 0.72, 0.64 and
0.48 per litre, respectively.

Find the number of cartons of ‘The Caribbean’ and ‘Mr Fruity’ that the firm
should produce to maximise profits. You may assume that non-alcoholic cocktails
are so popular that the firm can sell all that it produces.

International School of Business, UEH UEH, Vietnam November 17, 2022 190 / 200
 Chapter 7. Linear Programming 7.4. Simplex method for max problem

However, while the geometric method is used only for a linear programming of 2
variables, it is just as important to study methods for linear programmings
involving more than 2 variables.

Example. A manufacturer produces three models of bicycles. The time (in

hours) required for assembling, painting, and packaging each model is as follows:

The total time available for assembling, painting, and packaging is 36 hours, 60
hours and 66 hours, respectively. The profit per unit for each model is 3 (Model
A), 4 (Model B), and 5 (Model C). How many of each type should be
produced to obtain a maximum profit?
International School of Business, UEH UEH, Vietnam November 17, 2022 191 / 200
 Chapter 7. Linear Programming 7.4. Simplex method for max problem

Simplex method for max problem

Consider, for instance, the max problem of the function
Z = c1 x1 + c2 x2 + c3 x3 subject to
8
>
> a11 x1 + a12 x2 + a13 x3  b1
>
>
>
>
>
> a x + a22 x2 + a23 x3  b2
>
< 21 1
a31 x1 + a32 x2 + a33 x3  b3
>
>
>
>
>a41 x1 + a42 x2 + a43 x3
>  b4
>
>
>
:x1 , x2 , x3 0.

Note that b1 , b2 , b3 and b4 are supposed to be non-negative constants.

In 1947, George B. Dantzig created a simplex algorithm to solve linear programmings for planning and decision-making in

large-scale enterprises. The algorithm’s success led to a vast array of specializations and generalizations that have dominated

practical operations research for half a century later.

International School of Business, UEH UEH, Vietnam November 17, 2022 192 / 200
 Chapter 7. Linear Programming 7.4. Simplex method for max problem

The latter system can be written as Z c 1 x1 c 2 x2 c3 x3 = 0 subject to

8
>
> a11 x1 + a12 x2 + a13 x3 +s1 = b1
>
>
>
>
<a x + a x + a x +s2 = b2
21 1 22 2 23 3
>
> a31 x1 + a32 x2 + a33 x3 +s3 = b3
>
>
>
>
:a41 x1 + a42 x2 + a43 x3 +s4 = b4 ,
where s1 , s2 , s3 and s4 are non-negative and called the slack variables. So
now we have a departing simplex table as follows.

International School of Business, UEH UEH, Vietnam November 17, 2022 193 / 200
 Chapter 7. Linear Programming 7.4. Simplex method for max problem

Steps to find the optimal solution of the max problem:

S1. If all indicators are non-negative then STOP and Z has a maximum
with the current list of basic variables.

S2. If there is any negative indicator then choose a column w.r.t. the
biggest negative indicator as a pivot column.

S3. Divide each entry in R by the corresponding positive entry (Z and

negative entry are not included) in the pivot column and choose a
pivot row which has the smallest positive quotient.

S4. Create a new simplex table in which the common entry of the pivot
column and the pivot row above is transformed into 1 and other
entries in the pivot column zeros by using the elementary row
operations [executing with pivot row only!]

S5. Turn back to S1.

International School of Business, UEH UEH, Vietnam November 17, 2022 194 / 200
 Chapter 8. Di↵erential Equations [for your further reference] 8.1. Dynamics

So far, all of our models have been static. In practice, economic variables take
time to vary.

The incorporation of time into economic models is known as dynamics, and it

enables us to decide whether equilibrium values are actually achieved and, if so,
exactly how individual variables approach these values.

When t, is a discrete variable taking whole number values. This describes

the situation in which variables change only after fixed periods. For example,
the prices of certain agricultural products change from one season to the
next but are fixed throughout each season.

When time is a continuous variable taking all possible values in a certain

interval. This describes the situation in which variables change from one
moment to the next. For example, the prices of certain commodities, such as
oil, e↵ectively change instantaneously and are not fixed on a seasonal basis.

International School of Business, UEH UEH, Vietnam November 17, 2022 195 / 200
 Chapter 8. Di↵erential Equations [for your further reference] 8.2. Di↵erential equations

When time is a continuous variable, we express this time dependence using the
usual function notation, P(t). It is appropriate to model the rate of change of P
with respect to t using the derivative, P 0 (t). Equations that involve the
derivatives of an unknown function are called di↵erential equations.

Example. A bank pays 3.6 % interest on its certificate of deposit accounts, but
charges a 10 annual fee. Write an equation for the rate of change of the balance.

If the balance P(t) has a unit of dollars, the rate of change of the balance is
P 0 (t) which has a unit of dollars per year. Since the amount of increase in the
balance is 0.036P(t) dollars/year and the fee is 10 dollars/year, we obtain that

P 0 (t) = 0.036P(t) 10.

International School of Business, UEH UEH, Vietnam November 17, 2022 196 / 200
 Chapter 8. Di↵erential Equations [for your further reference] 8.2. Di↵erential equations

Ideal growth models

As we have studied so far, if the interest is compounded continuously, the

compound amount after t years is P = P0 e rt . This is the solution of the
di↵erential equation:
P0 = r ⇥ P

under the initial condition P(t = 0) = P0 .

Example
A bank pays 3.6% interest each year, compounded continuously. If the
initial investment is 5, 000 dollars, the investment after t years can be
modeled by P 0 = 0.036P and therefore P = 5, 000e 0.036⇥t .

International School of Business, UEH UEH, Vietnam November 17, 2022 197 / 200
 Chapter 8. Di↵erential Equations [for your further reference] 8.2. Di↵erential equations

Logistic models

However, capital cannot increase at an equal rate endlessly, the more so if the
system is completely or partially closed. When growing in such a system, capital
exhausts the limited resources in its environment. In other words, it enters into
self – competition which diminishes its growth – the system gets ‘satiated’.

Therefore, the growth of capital will be described by the logistic function of

growth proposed by Pierre-François Verhulst in the 1840s:
✓ ◆
P
P0 = P 1 ⇥ lnr
Pm

where Pm is the maximum (marginal) capital value (potential capital).

International School of Business, UEH UEH, Vietnam November 17, 2022 198 / 200
 Chapter 8. Di↵erential Equations [for your further reference] 8.3. Solving di↵erential equations

Separable equations

A separable equation is a first-order di↵erential equation in which the expression

for dP
dt can be factored as a function of t times a function of P. That is,

dP g (t)
= .
dt h(P)

We can use the Chain rule and the Fundamental theorem of Calculus to derive
R R
that h(P)dP = g (t)dt.

Example
Solve the following equation
dP t2
a. dt = P2 .

b. Find the solution of this equation that satisfies the initial condition
P(0) = 2.

International School of Business, UEH UEH, Vietnam November 17, 2022 199 / 200
 Chapter 8. Di↵erential Equations [for your further reference] 8.3. Solving di↵erential equations

To be continued ...

International School of Business, UEH UEH, Vietnam November 17, 2022 200 / 200