MATRICES AND VECTORS

1.0 Introduction
A matrix is a rectangular array of numbers considered as an entity with m rows and n columns

in the array. Hence we have an m -by- n matrix (written as m×n ). The dimension/order of a

matrix is stated as the number of rows and column. A matrix is denoted by bold capital letters such as
A,B,.. e.t.c. an example of an m×n matrix is given by:

a11 a12 ........ a1 n

A=
( a21 a22 ........ a2 n
..... ...... ........ ......
am1 am2 ......... a mn
) (1)

 The matrix A in (1) is said to have order m×n .

 The elements/entries are the mn numbers that constitute A, e.g. aij denotes an element
th th
in the i row and j column.
 A matrix with only one row is called a row vector and a matrix with only one column is
called a column vector.

−1 2

Example1: Given A=
(35 −28 ) , B= (−1 2 √ 3 16 ) and C=
[ ]8
7
1
5
6
1 . State the order of the

matrices,
a21 ,c32 ,b23 .

Solution to be provided in class

Example 2: Construct a 4×3 matrix A=

(aij )4×3 with
aij =2i+ j

Solution to be provided in class

 if m=n , the matrix has the number of columns as rows, and it is called a square matrix of
order n

 If
A=(aij )m×n , then the elements
a11 ,a 22 ,...,amn constitute the main diagonal that runs

from the top left (

a11 ) to the bottom right (amn ) .
 Square matrices with the property that they are symmetric about the main diagonal are called

2 −1 5 a b c

symmetric matrices. For example:

(−3 2
2 0 ) ,
−1 −3 2
5 2 8 ,
( )
b d e
c e f
Symmetric matrices are characterized by the fact that they are equal to their transposes:

The matrix A is symmetric ⇔ A=A T

T
 For any m×n matrix A, its transpose, denoted by A , is defined as the n×m matrix

whose first column is the first row of A, whose second column is the second row of A and so on.
Thus:

a b c a d g

A=
( )
d e f
g h i ( ) T
⇒A = b e h
c f i

−1 0

Example: Let A =
( ) (
3 2
5 −1 ,B=
1 −1 0 4
2 1 1 1 ) . Find the transpose of A and B.
Solution to be provided in class

1.1 Matrix Operations

a) Equal Matrices: If
A=(aij )m×n and
B=( bij )m×n are two m×n matrices, then A and

B are said to be equal and it implies that A=B, provided that

aij =bij for all

i=1,2,3 ...m and j=1,2,3...n . Hence two matrices are equal if they have the same
dimensions and if all their corresponding entries are equal. If A and B are not equal, we write
A ¿ B.
 Example: When is
(23t t−1 = t
u )( 2v
u+1 t+w )
Solution to be provided in class

b) Matrix addition and subtraction: If

A=(aij )m×n and
B=( bij ) m×n , we define the sum of

A and B as the m×n matrix (aij +bij ) m×n . Thus:

(aij )m×n +(bij )m+n =(a ij+bij )m×n … (3)
Hence to add matrices, add the corresponding elements. For example:

(ac bd )+( eg fh )=(ca+e+g b+f

d+h )
To subtract matrices, subtract the corresponding elements, for example:

(ac bd )−( eg hf )=( a−e

c−g
b−f
d −h )
1 2 0 2 3
Example: Given the following matrices A=
(
−2 4 ) , B=
( 1 0 5 ) and C=

3 −2
( 5 0 )
(i) Calculate A+C and A-C
(ii) Why is it not possible to calculate A+B or A-B? Hence state the restrictions on
matrix addition and subtraction.
Solution to be provided in class

c) Multiplication of a matrix by a scalar: A scalar is an ordinary number, 2,5 -8 e.t.c. If α is a

real number, we define α A by: α A= α (aij )m×n =(αa ij )m×n ….. (4). Hence when a
matrix is multiplied by a real number, each element in the matrix is multiplied by the real
number. For example:

k a b = ka kb
[ ][
c d kc kd ]
 Example: Given the matrix A=
(−21 24 ) calculate 5A,
1
− A
2 .
Solution to be provided in class

 The matrix (−1) A is usually denoted by –A. the difference between the two
matrices A and B of the same dimension, A-B, means the same as A+ (-1)B.

d) Matrix multiplication: suppose that

A=(aij )m×n and
B=( bij ) n× p , then the product

C=AB is the m× p matrix C=(c ij )m× p , whose element in the ith row and the
th
j column is the product sum (or dot product):
n
c ij= ∑ air brj =ai 1 b1 j + ai 2 b 2 j +.. . ..+a ik bkj +. . .. ..+ ain bnj
r=1 of the ith row of A and the
th
j column of B. For example:

a b e f = ae+bg af +bh
[ ][ ] [
c d g h ce +dg cf +dh ]
Dimension: (2×2 )×(2×2)=(2×2 )

= 1 2 (01 2 2
Example: Given the matrices A
(
−2 4 ) , B= 0 5 ) and C=
(35 −20 ) ,
Evaluate AC, CA, AB and BA. Compare answers for AC and CA.
Solution will be provided in class.

 The product AB is defined only if the number of columns in A is equal to the number
of rows in B.
 If A and B are two matrices, then AB might be defined even if BA is not. For
example, if A is 6 by 3 and B is 3 by 5, then AB is defined (as its 6 by 5) whereas BA
is not defined.

Class Exercise
 1. Given the matrices A=
(10 −4−9 ) , B=
(−74 30 ) and C=
(125 −1 −1
0 2 ) ,
calculate if possible:
(i) A+B, A-B, A+4B and A+C
(ii) BC, CB, AC+B

2. Find the transpose of A=

(3 5 8 3
−1 2 6 2 ) and B =
()
1
−1
2 ,C= (1 4 2 6 )

1.3 Applications of matrix operations

A distributor records the weekly sales of personal computers (PCs) in three retail outlets in
different parts of the country.
Table 1: Number of Computers sold in each shop

Basic Extra Latest

Shop A 150 320 180
Shop B 170 420 190
Shop C 201 63 58

The cost price of each model is: Basic $480, Extra $600, Latest $1020. The retail price of
each model in each of the three shops is given in the Table 2 below:
Table 2

Basic Extra Latest

Shop A 560 750 1580
Shop B 520 690 1390
Shop C 590 720 1780

Use matrix multiplication to calculate:

(i) The total weekly cost of computers to each shop
(ii) The total weekly revenue for each model for each shop
(iii) The total weekly profit for each shop. Which shop makes the greatest overall profit?

Solution: (i) the numbers sold from Table 1 may be written as a matrix, Q:
 PC→ Basic Extra Latest

Q=
shopA
shopB
shopC
( 150
170
201
320
420
63
180
190
58
)
Write the cost of each type of computers as a column matrix:

480
C= 600
( )
1020

If this cost matrix C is pre-multiplied by the numbers sold matrix, Q, the product will be a 3 ¿ 1matrix,
in which the elements in each row give the total cost of computers to each shop:

Q*C = total cost

Dimension: (3×3)(3×1 )=(3×1)

150 320 180 480 150(480 )+320(600)+180(1020 )

( )( ) (
170 420 190 600 = 170(480 )+420(600 )+190(1020)
201 63 58 1020 201(480 )+63(600)+58(1020 ) )
447600

=
( )
527400
193440

Cost to: shop A=$447 600, shop B=$527 400, shop C=$193 440

(ii) Total revenue = price * quantity. The quantities are given in matrix Q. the prices are obtained in table
2. Matrix multiplication, however, is carried out by multiplying rows by columns; therefore in order to
multiply quantity and price for each PC, the rows in table 2 must be written as columns in the prices
matrix. That is, prices matrix must be transposed before pre-multiplying by the quantities matrix, Q.

Multiplying Q by the transpose of P:

 150 320 180 560 520 590 604, 400

Q*P =
( 170 420 190
201 63 58 ) ( 750 690 720
1580 1390 1780 ) ( )
=
642,300
267,190

(iv) Profit = TR – TC

604,400 447600 156,800

( ) ( ) ( )
= 642,300
267,190 -
527400
193440 =
114 ,900
73,750
Therefore shop A makes the highest profit.

Class Exercise

a) The percentage of voters who will vote for party candidates A,B and C is given in the following
table:

A B C No. of voters
Area1 60% 20% 20% 25 000
Area 2 45% 30% 25% 60 000
Area 3 38% 30% 32% 98 000
Use matrix multiplication to calculate the total number of votes for each candidate
1.4 Determinants
A determinant is a square array of numbers or symbols, for example:

1 0 −2
a b 2 5 |2 2 3 |
| | | |
A= c d B= 3 −4 D= 1 3 2
How to evaluate a 2×2 determinant

A 2×2 determinant is evaluated as follows:

a b
| |=(a)(d)−(c )( d)
c d
Hence the value of the determinant B is:

2 5
| |=(2)(−4 )−(3)(5)=−8−(15)=−23
3 −4
 Cramer’s rule: Expansion by Cofactors

a b c

Given A =
( )d e f
g h i , its determinant has a value of:

a b c
e f d f d e
|d e f |=a| |−b| |+c| |
h i g i g h
g h i
= a(ei−hf )−b(di−gf )+c (dh−ge )

= aei−ahf −bdi+bgf +cdh−cge

= scalar
Example: Evaluate the following determinants:

1 2 −1 0 1 0 1 −1 0
|3 1 0 | |3 1 1 | |1 3 2 |
0 0 1 , 1 2 −1 and 1 2 1
Solution to be provided in class

Sarrus’ Rule

Write down the determinant twice, except that the last column in the second determinant
should be omitted:

a b c a b
|d e f d e |
g h i g h
First multiply along the three lines falling to the right, giving all these products a plus sign:
aei+bfg+cdh
Then multiply along the three lines rising to the right, giving these products a minus sign:

−gec−hfa−idb

The sum of all the six terms is exactly equal to |A|

Example: Use Sarrus’s Rule for the previous example
1.5 The Inverse of a Matrix
Given a matrix A, we say X is an inverse of A if there exists a matrix X such that : AX = XA
= I. A is said to be invertible. Because XA = AX = I, the matrix A is also an inverse of X, that
is, A and X are inverses of each other.
NB: only square matrices can have matrices, but not all square matrices have inverses.

A square matrix A has an inverse ⇔ |A|≠0

A square matrix A is said to be singular if |A|=0 and non-singular if |A|≠0 .
−1
The inverse is written as A .
Procedure

(i) Find the |A| and proceed iff |A|≠0

(ii) Find the cofactors of all the elements of A and arrange them as a cofactor matrix

C=[|C ij|] , that is:

Assume A is a non-singular matrix given by:

a b c
A n= d e f
g h i ( ) , since each element of A has a cofactor
[|C ij|] , it is possible

to form a matrix of cofactors by replacing each element

aij in A by its cofactor

[|C ij|] . Such a cofactor matrix denoted by,

C=[|C ij|] , must also be of order n.
(iii) Take the transpose of C to get the adjoint of A abbreviated adj A.

(iv) Divide the adjoint of A by the |A| . The result is the inverse of A. That is:
1
A−1 = adjA
|A|
4 1 −2

Example: Find the inverse of B =

[ ]
0 3 2
3 0 7
Solution to be provided in class

Class exercise
 a) Evaluate the following determinants:

1 0 −2 15 −4 14
|2 2 3 | |−4 6 34 |
D= 1 3 2 E= −3 −2 1

b) Find the inverses of matrices D and E in a) if they are non-singular.

1.6 Using Cramer’s rule to solve simultaneous equations

Consider the system:
a11 x 1+a 12 x 2+....+a1 n x n=b1
a21 x1+a22 x 2+.....+a2 n x n=b2
..................................................
an1 x n +a n2 x n+......+ann xn =b n
The system of equations is equivalent to the matrix Ax=b where A is the matrix of
coefficients, x is the matrix of the unknown variables and b is the matrix of the right hand
sides. The system can be written as:
A x = b

a11 a12 .. .. . a1 n x 1 b1

( a21 a22
. .. .. . .. . .
a n1 a n2
.. .. . .
.. .. . .
.. .. . .
)( ) ( )
a2 n x 2 = b 2
.. .. .. .. .. ..
ann x n bn

The general linear system of equations with n equations and n unknowns has a unique solution iff

A is non-singular ( |A|≠0 ). The solution is:

D1 D2 Dn
x 1= , x2 = ,. . .. . , x n =
|A| |A| |A|

Where
D1 , D2 ,...,Dn are defined as the determinant obtained from |A| by replacing

the jth column vector with the column vector whose components are
b1 ,b 2 ,...,b n

Example: find the solutions of the following system for all values of p:
 px+ y=1
x− y+z=0
2 y−z=3

Solution to be provided in class

Class exercise
1. Given the supply and demand functions for two related goods, A and B,

¿¿¿ ¿¿¿
Good A: ¿ Good B: ¿
(a) Write down the equilibrium for each good. Hence deduce two equations in
Pa and

Pb
(b) Use Cramer’s rule to find the equilibrium prices and quantities for goods A and B

2. Use Cramer’s rule to find Y and C when:

Y =C +I 0 +Go and C=a+bY

Where Y is the national product and C is private consumption. The symbols

I 0 (private

investment),
G0 (public consumption and investment), a and b all represent

constants with b≺1 .

The Leontief Model

The Leontief model describes an economy with n interlinked industries, each of which produces a single
commodity using only one process of production. To produce its good, each industry must use inputs
from at least some other industries. For example, the steel industry needs goods from the coal industry as
well as from many other industries. In addition to supplying its own good to other industries that need it,
each industry also faces external demand for its product from consumers, governments, foreigners, and so
on. The amount needed to meet this external demand is called the final demand.

Let
x i denote the total number of units of good i that industry i is going to produce in a certain
year.
Let
aij x j= {the number of units of good i needed to produce
xj units of good i ……
(1)

In order that
x 1 units of good 1, x 2 units of good 2, ….., x n units of good n can all be produced,

industry i needs to supply a total of:

ai 1 x 1 +ai2 x 2 +. . ..+a in x n units of good i

If we require industry i also to supply bi units to meet final demand, then equilibrium between
supply and demand requires that:

x i=a i1 x 1 +ai 2 x2 +.. ..+ain x n +b i …… (2) for all i=1,2,....,n

So we arrive at the following system of equations:

x 1=a11 x1 +a12 x 2 +. . ..+a1n x n + b1

x 2=a21 x 1 +a 22 x 2 +. .. .+a2 n x n +b2

…………………………………... …….. (3)

x n=an 1 x1 +an 2 x 2 +. .. .+ann x n +b n

Moving all the terms involving

x 1 ,x 2 ,....,x n to the left hand side and rearranging gives the system of
equations:

(1−a 11 ) x1 −a 12 x 2−. . ..−a1 n x n=b1

−a21 x1 +(1−a22 )x 2 −. .. .−a 2n x n =b 2

………………………………………. (4)

−an 1 x 1−an 2 x 2 −. .. .+(1−ann ) x n =bn

This is the Leontief system. The numbers
a11 ,a12 ,.. .. ,ann are called input (technical) coefficients.

Given any collection of final demand quantities (

b1 ,b 2 ,...,b n ), a solution ( x 1 ,x 2 ,..., x n ) of (4) will
give outputs for each industry such that the combined inter-industry and final demands just can be met.

In matrix notation, (4) can be written as:

(1−a 11 ) −a12 . .. . −a1 n x1

[ −a21 (1−a22 )
. . ..
−an 1
. .. ..
−an 2
. .. . . −a2 n
. .. . . . .. . ..
. .. . .. (1−a mn )
] [] x2
...
xn
…. (5)

If the ones (1s) in the principal diagonal of the matrix on the left are ignored, the matrix is simply

−A=[−aij ] . On the other hand, the matrix is the sum of the identity matrix,
I n and the matrix
−A . Thus (5) can also be written as:

( I−A ) x=d

Where x and d are, respectively, the variable vector and the final demand vector.

Matrix ( I−A ) is called the technology matrix. If ( I−A ) is non-singular, then the inverse

( I−A )−1 can be found, and (5) will give the unique solution:

x=( I− A )−1 d

NB: in input/output analysis, total input = total output for each sector.

Example:

Given the input/output table for the three-sector economy:

Input to

Agric Industry Services Other demands Total Output

Output Agric 150 225 125 100 600

From Industry 210 250 140 300 900
 Services 170 0 30 100 300
If the final demands from each sector are changed to 500 from Agriculture, 550 from industry, 300 from
financial services, calculate the total output from each sector.
Solution:
Step 1: Use the underlying assumption total input = total output to complete the input/output table:
Input to

Agric Industry Services Other demands Total Output

Output Agric 150 225 125 100 600

From Industry 210 250 140 300 900
Services 170 0 30 100 300

Other inputs 70 425 5

Total input 600 900 300

Step 2: Calculate the matrix of technical coefficients, A, by dividing each column of inputs by total input.

Input to

Agric Industry Services Other demands Total Output

Output Agric 150/600 225/900 125/300 100 600

From Industry 210/600 250/900 140 /300 300 900
Services 170/600 0/900 30/300 100 300
Total input 600/600 900/900 300/300

−1
Step 3: get the inverse of the matrix ( I−A ) , since it is required in the equation x=( I− A ) d :

150 225 125

( )
600 900 300
1 0 0 0 .75 −0 .25 −0 . 42
( 0
0
1
0 1)
0 −
210
600
170
250
900
0
140
300
30
(
= −0 . 35 0 .72
−0 . 28 0 .00
−0 . 47
0 . 90 )
( I−A ) = 600 900 300

To calculate the inverse of ( I−A ) , determine the matrix of cofactors:

 +|0 .72 −0 .47| −|−0. 35 −0 .47 | +|−0 .35 0 .72 |

( )
0 .00 0 . 90 −0. 28 0. 90 −0 . 28 0 .90
0 . 648 0 .4466 0 .2016
C= −|

+|
−0 . 25 0 . 42
0 .00 0 . 90
−0 .25 −0 . 42
| +|

| −|
0.75
−0. 28
0 .75
−0. 42
0.90
−0.42
| −|

| +|
0 .75
−0 . 28
0 .75
−0 .25
0 .00
−0 .25
|
(
| = 0 . 225 0 .5574 0 .070
0 .4199 0 .4995 0 .4525
)
0 .72 −0 . 47 −0. 35 −0 .47 −0 . 35 0 .72

0. 648 0 .225 0.4199

T
(
C =adjA= 0. 4466 0. 5574 0.4995
0. 2016 0 .070 0.4525 )
0.648 0.225 0. 4199

The inverse of ( I−A ) =

adjA
|I− A| =
1
0. 289678 (
0.4466 0.5574 0. 4995
0.2016 0.070 0. 4525 )
Step 4: Finally, state the column of new external demands, d , and solve for x using the formula:

x=( I− A )−1 d

T agric 0 .648 0 . 225 0 . 4199 500 1980 .5

( )
T ind . =
T serv
1
0 . 289678 (
0. 4466 0 .5574
0 .2016 0 . 070 )( ) (
0 . 4995 550 = 2346 . 50
0 . 4525 300 949 .50 )

Class exercise

Given the inter-industry transaction table for two industries,

Input to X Input to Y Final demand

Output from X 600 400 200

Output from Y 600 200 0
(a) Write down the matrix of technical coefficients
(b) Calculate the total output required from each industry if the final demands from X and Y change
to 500 and 1000 units, respectively.
 2. COMPARATIVE STATIC ANALYSIS OF GENERAL FUNCTION MODELS

dy
In this topic, dx is reinterpreted as a ratio of two quantities, dy and dx .

2.1 Differentials and derivatives

Given a function y=f ( x ) , we can use the difference quotient Δy / Δx to represent the rate of

change of y with respect to x . Since it is true that:

 Δy≡ ( ΔyΔx ) Δx (1)

The magnitude of Δy can be found, once Δy / Δx and the variation in x are known. If Δx is

dy
infinitesimal, Δy will be infinitesimal and Δy / Δx approaches the derivative dx . Therefore, (1)
becomes:

dy
dy=( )dx '
dx or dy≡f ( x )dx (2)

The symbols dy and dx are called the differentials of y and x respectively. Dividing (2)

(dy ) dy (dy ) '

throughout by dx
⇒ ≡
(dx ) dx ( ) or (dx )
≡f (x )
(3)

(dy ) '
≡f (x )
This implies that (dx ) may be interpreted as the quotient of two separate differentials dy

and dx . Once given the derivative of a function y=f ( x ) , the differential dy can immediately be
written.

Example1: Given, y=x +3 x−2find

2 dy

Solution: f ' ( x )=2 x+3 , therefore the desired differential is: dy =( 2 x +3 ) dx

???? Assuming that x is to change from 2 to 2.001, calculate dy .

Set x=2and and dx=0.001substitute into dy =( 2 x +3 ) dx

dy =( 2 [ 2 ] +3 )∗0.01=0.07

???? How does this figure compare with the actual change in y ?

When x=2 ; y=( 22+ ( 3∗2 )−2 ) =8 , when X =2.001; y =2.0012+ ( 3∗2.001 ) −2=8.007001
Therefore the true change in y is ∆ y =8.007001−8=0.007001 ⇒ our answer dy =0.007
constitutes as approximation with an error of0.007001−0.007=0.000001 .

The process of finding the differential dy is called differentiation.

Example 2: Find the differential dy 2

for a function defined by x + 8 x+ 4. Find dy when x changes

from 4 to 4.004. What is the error involved in using dy as an approximation to Δy ?

Solution to be provided in class

2.2 Differentials and Point Elasticity

( ΔQ/Q )
For a demand function Q=f ( P) , the elasticity is defined as ( ΔP/P ) . If the change in P is

infinitesimal, the expressions ΔP and ΔQ will reduce to the differentials dP and dQ , and the

elasticity measure will then assume the sense of the point elasticity of demand, denoted by
εd .

dQ/Q dQ/dP
ε d≡ =
dP /P Q/ P

dQ/dP can be construed as the derivative, or the marginal function of the demand function

Q=f ( P) . Q/ P is the average function of the demand function. Therefore ε d is the ratio of the
marginal function to the average function of the demand function.

This implies that for any given total function y=f ( x ) , we can write the formula for the point
elasticity with respect to x as:

dy /dx marginalfunction
ε yx = =
y /x averagefunction

Example: Find
ε d if the demand function is Q=30−0.42 P

dQ
Solution: Marginal function: =−0.42
dP
 Q 30−0.42 P
Average function: =
P p

−0.42 P
So their ratio will give ε P=
30−0.42 P

Example 2: If the Savings function of an economy is defined as S=100+ 0.2Y −0.31 iwhere Sis

Savings and Y is national income and i is interest rate. Find the partial income elasticity of savings
and partial interest elasticity of savings.

Solution to be provided in class

Class Exercise

1. Given y=2 x 3 + x 2−3 x+8

(i) Find the differential dy

(ii) When changes from 5 to 5.02, what is the value of dy
(iii) Find the true value of Δy and determine the error involved when dy is used as an

approximation to Δy .
2. Given the import function, M =250+ 0.4 Y , where M is imports and Y is national income,

Find the income elasticity of imports

ε MY

2.3 Total Differentials

Consider a Savings function: S=f (Y ,i ) where S -savings, Y -national income, i -

interest rates
Assume the function to be continuous and to posses continuous (partial) derivatives, that is, it is

smooth and differentiable everywhere. The partial derivative ∂S/∂Y or

S Y measures the rate

of change of S with respect to an infinitesimal change in Y or MPS. Therefore the MPS may

be represented by the expression (∂ S/∂ Y )dy comparable to

'
f ( x )dx . The change in S

resulting from an infinitesimal change in i can be denoted by (∂ S/∂ i)di . The total change in
S will then be equal to:
dS=(∂ S /∂Y )dy+ ( dS/di ) di
 OR
dS=SY dy +Si di

The expression dS , being the sum of the changes from both sources, is called the total
differential of the savings function. The process of finding such a total differential is called total

differentiation. If Y change while i remains constant, di=0 , the total differential will reduce

to a partial differential dS=(∂ S/∂Y )dy . Dividing both sides by dY we get:

dS ∂ S
=
dY ∂Y ( ) icons tant

It is clear that the partial derivative (∂ S/∂ Y ) can also be interpreted as the ratio of two

differentials dS and dY with the proviso that i , the other independent variable in the
function, is held constant.
Example 1: find the total differential dz given z=4 x 2−8 y 2+ 6

Solution: dz= ( ∂∂ zx )dx +( ∂∂ zy )dy =( 8 x ) dx− (16 y ) dy

Example2: Find the total differential given z=x + x 2 y−2 y
Solution will be provided in class

2.4 Total Derivatives

Total derivatives help answer the question, “how do we find the rate of change of the function

C(Y ,T 0) with respect to

T 0 , when Y and T 0 are related?” unlike a partial derivative, a

total derivative does not require the argument Y to remain constant as

T 0 varies, and can
thus allow for the postulated relationship between the two arguments.

Consider any function y=f ( x , w ) where x=g(w ) with the variables y , x and w

related to one another, that is;

f g
y x w

f
 w , the ultimate source of change, can affect y through two channels:

(1) Indirectly, via the function g and then f (the straight arrows)

(2) Directly, via the function f (the curved arrow)

Whereas the partial derivative

f w is adequate for expressing the direct effect alone, a total derivative is

needed to express both effects jointly. To obtain the total derivative, first differentiate y totally, to get

the total differential

dy=f x dx+f w dw . When both sides of this equation are divided by the

differential dw , the result is:

dy dx dw
=f x +f w
dw dw dw

∂ y dx ∂ y
+
= ∂ x dw dw

Indirect effect direct effect

Example 1: Find the total derivative dy/dw , given the function y=f ( x , w )=3 x−w2 where

x=g(w )=2 w2 +w+4

Solution to be provided in class

2 2
Example 2: Suppose we have a production function Q=Q( K , L)=25 KL−K −2 L where the two

factors K and L ate functions of time, t: K=g(t )=0 .3 t , L=h(t )=0 . 2 t . Find the rate of change of
output with respect to time.

Solution to be provided in class

Suppose we have z=f ( x, y , w) where x=g(w ), y=h( w)

 dz dx dy dw
=f x +f y +f w
dw dw dw dw

Class Exercise

1. Find the total differential given U=2 x +5 xy+ y

2. Evaluate the total derivative dz/dt given the function z=f ( x , y )=x3 +4 xy − y 2 where
x=2 t , y=2−t
dP
3. Use the total derivatives to find
dY 0 for a market model whose equilibrium identity is

D( P ,Y 0 )−S( P)≡0 , where

P=P(Y 0 )

3.5 Rules of Differentials

Let u and v be two functions of the variables x 1 and x 2 , then the following rules are
valid:

(i) d (cu n )=cnun−1 du (power rule)

(ii) d (u+ v )=du±dv (sum-difference rule)

(iii) d (uv )=vdu+udv (product rule)

(iv)
d ( uv )= v1 ( vdu−udv )
2
(quotient rule)

dy y=5 x 2 + 3 x 2
Example 1: Find the total differential of the function 1

u=5 x v=3 x 2
Let 12 and
dy=d (5 x 2 )+d (3 x 2 )
1

Therefore: ¿ 10 x 1 dx 1 +3 dx 2 (by rule 2 and 1 respectively)

2 2
Example 2: Find the total differential of the function z=3 x +xy
Solution to be provided in class
Class Exercise
Use the rules of differentials to find the total differential of the following functions:
y=3 x 3 + x1 x
a) 1 22

x 1+ x 2
y=
x
b) 22

3.6 Economic Application

Consider a single commodity market, where the quantity demanded,

Qd is a function not only of

price P but also of an exogenously determined income Y 0 . The quantity supplied QS , on the
other hand, is a function of price alone. Hence the model can be written as:

Qd =Q s
Qd =D( P , Y 0 ) ( dD /dP≺0 ;dD/dY 0 )
QS =S (P ) ( dS /dP≻0 ) (1)

Both D and S functions are assumed to possess constinuous derivatives, or, in other words, have
smooth graphs. When income changes, it will upset a given equilibrium by causing a shift in the

demand curve. Similarly in (1),

Y 0 can cause a disequilibrating change through the demand

function. Hence,
Y 0 is the only exogenous variable, thus the comparative static analysis of this

model will be concerned exclusively with how a change in

Y 0 will affect the equilibrium position
of the model.

The equilibrium position of the market is defined by the equilibrium condition

Qd =Q s

D( P ,Y 0 )−S ( P)=0 (through substitution and rearrangement) (2)

Even though this equation cannot be solved explicitly for the equilibrium price, P , we shall assume
that there does exist a static equilibrium. From experience with specific function models, we expect

P to be a function of the exogenous variable

Y0 .
 P=P(Y 0 ) …. (3)

Rigorous foundation fro this expression can be provided by appealing to the implicit function

theorem. In as much as (2) is in the form of F( P , Y 0 )=0 , the satisfaction of the conditions of the

implicit function theorem will guarantee that every value of

Y 0 will yield a unique value of P in
the neighborhood of a point satisfying (2), that is, in the neighborhood of an (initial or old)

equilibrium solution. In that case, we can indeed write the implicit function
P=P(Y 0 ) , and

discuss its derivative,

d P /dY 0 , the very comparative-static derivative we desire, which is known
to exist. By analyzing/checking the conditions of applicability of the theorem, we can see that the
function satisfies all of them and hence the implicit function theorem applies and (3) is indeed
legitimate. According to the same theorem, the equilibrium condition (2) can now be taken to be an
identity in some neighbourhood of the equilibrium solution. Consequently we may write the
equilibrium identity:

D( P ,Y 0 )−S( P)≡0 …(4)

It then requires only a straight application of the implicit function rule to produce the comparative

static derivative,
d P /dY 0 . The result is:

∂ F /∂Y 0 ∂ D/∂Y 0
d P /dY 0 =− =− ≻0
∂ F /∂ P ∂ D/∂ P−∂ S/∂ P (5)

∂D
∂D/∂ P refers to the derivative ∂ P evaluated at the initial equilibrium, that is, at P=P ,a

similar interpretation attaches to ∂S/∂ P . The qualitative conclusion is that a small

increase/decrease in the income level will always result in an increase/decrease in the equilibrium
price. If the values which the derivatives of the demand and supply function take the initial
equilibrium are known, then (5) will yield a quantitative conclusion also.

2.6 Class Exercise

 In the very short run, the supply of a commodity is fixed, thus we may write
Qs =S0 . Assume that

the demand function of this commodity takes the form: Qd =D( P , t 0 ) where

∂ D/∂ P≺0;∂ D/∂ t0 and t 0 represents given consumer taste for the commodity, and where both
partial derivates are continuous.

(i) Is the implicit function theorem applicable? Why?

(ii) How would you write the equilibrium solution for P and Q?

(iii) How would P,Q , that is, equilibrium price and quantity respectively, vary with a change
in consumer taste?

3. DIFFERENTIAL EQUATIONS

3.1 First-Order Linear Differential Equations

Introduction

The first derivative dy/dt is the only one that appears in a first order differential equation, but
2 3
it may appear in various powers, dy/dt , (dy/dt ) or (dy/dt ) .

The word order refers to the highest order of the derivatives (or differentials) appearing in the
differential equation.

The solution to a differential equation is often referred to as the intergral of that equation.

The highest power attained by the derivative in the equation is referred to as the degree of the
differential equation.
In cases where the derivative dy/dt appears only in its first degree, and likewise the
dependent variable y , and furthermore, no product of the form y (dy/dt ) occurs, then the
equation is said to be linear. Thus a first order linear differential equation will generally take the
form:

dy
+u(t ) y=w(t )
dt ………….(1)

Where u and w are two functions of time as is y. in contrast to dy/dt and y , however, no
restriction is placed on the independent variable t, thus the functions u and w may represent
2 t
expressions as t ,e or some more complicated functions of t, on the other hand, u and w may
also be constants.

First-order Linear Differential Equations with a constant coefficient and a constant term

When the function u (the coefficient of the dependent variable y) is a constant, and when the
function w is a constant additive term, then (1) reduces to the special case of a first-order linear
differential equation with a constant coefficient and a constant term.

The homogeneous case

If u and w are constant functions and if w is identically (equal to) zero then (1) becomes:

dy
+ay =0
dt …………(2)

Where ‘a’ is some constant. This differential equation is said to be homogeneous, on account of
the zero constant term. More accurately, because every term therein is uniformly in the first
degree in terms of y and dy/dt , in particular, the constant 0 can be regarded as in the first
degree in terms of y because 0y=0. Equation (2) can be written alternatively as:

1 dy
=−a
y dt ……….(2’)

Taking integrals from both sides of 2’ yields:

1 dy
∫ y dt =∫ −a
1
∫ y dy=∫ −adt
 ln y=−at−c
y=e−at −c
y= Ae−at where A=e
−c

Therefore y (t )= Ae−at (General Solution)…..(3)

−at
OR y(t )= y(0)e (Definite Solution)……(3’)

In (3) there appears an arbitrary constant A, therefore, it is a general solution. When a particular
value is substituted for A, the solution becomes a particular solution of (2); there is an infinite
number of particular solutions, one for each possible value of A, including the value of y(0). This
latter value however, has a special significance: y(0) is the only value that can make the solution
satisfy the initial condition. Since this represents the result of definitising the arbitrary constant,
we shall refer to 3’ as the definite solution of the differential equation (2) or (2’).

NB: 1. The solution is not a number or some value, but rather a function y(t).

2. The solution y(t) is free of any derivative or differential expressions, so that as soon as a
specific value t is substituted into it, a corresponding value of y can directly be calculated.

The non-homogeneous case

When a non-zero constant takes the place of a zero in (2), we have a non-homogeneous
differential equation:

dy
+ay =b
dt ……(4)

The solution of this equation will consist of the sum of two terms, one of which is termed the
complementary function (denoted by
y c ), and the other known as the particular integral

(denoted by
y p ).

NB: A homogeneous equation is a reduced equation of (4), equation (4) is the complete equation.
Therefore, the complementary function is nothing but the general solution of the reduced
equation (homogeneous equation), whereas the particular integral is simply any particular
solution of the complete equation.
−at
Therefore y c =Ae (by 3)

Particular integral

First try the simplest possible type of solution:

 dy
y=k ⇒ =0
dt , then (4) becomes:

ay=b
b
y=
a
b
y p = (a≠0)
a
The sum of the complementary function and the particular integral then constitutes the general
solution of the complete equation (4).

b
⇒ y (t )= y c + y p = Ae−at +
a (General Solution, case of a≠0) …..(5)

Let y take the value y(o) when t=0; then by setting t=0 in (5), we get:

b
y (0 )= A+
a
b
⇒ A= y (0 )−
a
b b
y (t )=[ y (0 )− ]e−at +
Thus we can rewrite (5) into: a a (Definite Solution, case of a≠0)

dy
+4 y=0
Example 1: Solve the equation dt with the initial condition y(0)=1 .

Solution to be provided in class

dy
+2 y=6
Example 2: Solve the equation dt with the initial condition y(0 )=10

Solution to be provided in class

What if a=0 , then the solution in 5’ is undefined and the differential equation is of the
extremely simple form:

dy
=b
dt ….(6)

By straight integration, the solution can be found as:

∫ dy=∫ bdt
 y (t )=bt +c …..(7)

Where c is an arbitrary constant. The two terms in 7 can, in fact, again be identified as the
complementary function and the particular integral of the given differential equation,
respectively. Since a=0 , the complementary function can simply be expressed as:
−at 0
y c =Ae = Ae = A

OR ∫ dy=∫ 0dt
y=c
y c =A

yp?

dy
y=kt , =k
Let dt

⇒ k=b
∴ y p =bt
∴ y(t )= y c + y p = A+bt
(General Solution)

Therefore the definite solution if y(0) is the initial value of y is :

y (t )= y (0 )+bt (Definite solution, a=0 )

dy
=2
Example: Solve the equation dt , with the initial condition y (0 )=5

Solution to be provided in class

Practice Questions

Find the general solution and the definite solution of the following differential equations

dy
+4 y=8 , y (0 )=2
a) dt

dy
+10 y=15 , y (0 )=0
b) dt
 dy
=15 , y (0 )=1
c) dt

3.2 Separable Differential Equations

The differential equation

f ( y , t )dy +g ( y ,t )dt=0 …(1)

May possess a convenient property that the function f is in the variable y alone, while the
function g involves only the variable t so that the equation reduces to:

f ( y )dy+g (t )dt=0 ……(2)

(2) is said to be separable because the terms involving y consolidated into f(y) can
mathematically be separated from the terms involving t, which are collected under g(t). We use
simple integration techniques to solve such equations.
2
Example 1: Solve the equation 3 y dy−tdt =0

Solution to be provided in class

Example 2: Solve the equation 2tdy− ydt=0

Solution to be provided in class

Practice Questions

Solve the following differential equations

dy
=−2 y 2 t
a) dt
3
dy t
= 6
b) dt y +1

3.3 Economic Application of First-Order Differential Equations with a constant

coefficient and a constant term

Dynamics of Market Prices

The Framework

Suppose that, for a particular commodity, the demand and supply functions are as follows:
 Qd =a−bP ,( a ,b≻0 )
Qs =−c +dP ,( c , d≻0 ) ………..(1)

The equilibrium price is given as:

a+c
P= =
b+ d some positive constant (1’)

P is attainable (if ever) after a due process of adjustment, during which not only will price
change over time but
Qd ,Q s being functions of price, must change over time as well,
implying that P and Q can all be taken as functions of time.

Dynamic question: Does the time path P(t) tend to converge to P , as t→∞
The time path

Assume, for the sake of simplicity, that the rate of price change (w.r.t. time)at any moment is
always directly proportional to the excess demand ( Q d −Q s ) prevailing at that moment.

dP
=α (Q d −Qs ),(α≻0 )
i.e. dt ……(2)
α represents a (constant) adjustment coefficient.

dP
=0 Qd =Q s
dt iff

By virtue of the demand and supply functions in (1), we can express (2) in the form

dP
=α (a−bP+ c−dP )=α(a+ c )−α (b+d )P
dt
dP
or + α(b+ d )P=α (a+c )
dt
 4. DIFFERENCE EQUATIONS
1. Introduction

Difference equations are used to model systems where changes occur at discrete points in time. The
terminology used to describe and solve difference equations is similar to that used for differential
equations. At the end of this topic a student should be able to:

 Write down a difference equation to model simple systems that change at discrete points in time
 Solve first-order and second-order difference equations
 Apply concept to economic problems

2. Simple First-order Difference Equations

Let t=0,1,2...... denote different discrete time periods or points of time.

Call t=0 the initial period.

Let f (t , y) be a function defined for all positive integers t and all real numbers x

y t+1 =ay t +b ,
Consider the difference equation t=0,1 ,. . .. where a is a constant. Given y 0 , it is possible to

calculate
y t algebraically for small values of t . This can be done either by using the Iterative
method or the General method.

2.1 Iterative Method

Given any difference equation, and an initial value

y 0 , we can find y 1 from the equation. Once y 1

is found
y 2 will be immediately obtainable, and so forth by repeated application (iteration) of the
pattern of change specific in the difference equation. Results of iteration will then permit the inference of
a time path.

Example1 : find the solution of the difference equation Δy=2 assuming an initial value of
y 0 =15

Solution: to be provided during the lecture

2.2 General Method

Suppose a difference equation of the form

y t+1 +ay t =c ………………(1)

where a and c are two constants. The general solution will consist of the sum of the particular

integral,
yp and a complementary function,
yc

t
Complementary function: Let y t = Ab (with Ab t ≠0 ), find y t+1 and substitute into the
homogeneous equation. (Illustration to be provided in class)

t t
⇒ y c (= Ab )=A (−a ) …………………………..(2)
Particular integral: let
y t =k (a constant), evaluate y t+1 and substitute into the complete equation.
(Illustration to be provided in class)

c
⇒ y p =k= (a≠−1)
1+a …………………………..(3)

NB: Since
yp is a constant, a stationary equilibrium is indicated.

If a=−1 , (3) is not defined, hence let y t =kt , evaluate y t+1 and substitute into the complete
equation.

(Illustration to be provided in class)

⇒ y p (=kt )=ct ………………………………..(4)

Adding
y c and y p together gives the following general solution:

c
y t = A(−a )t +
1+a (General solution, a≠−1 )………………(5)

t
y t = A(−a ) +ct= A+ ct (General solution, a=−1 )…………..(6)

To find the definite solution, let

y t = y 0 and t=0 in (5) and (6). (Illustration to be provided in
class)

c c
y t =( y 0− )(−a)t +
The definite version of (5) is 1+1 1+ a ……………………………….(5’)

The definite version of (6) is

y t = y 0 +ct …………….(6’)

Examples: (a) y t+1 −5 y t =1( y 0 =7/4 (b) y t+2 +3 y t =4( y 0=4 ) (Illustration to be provided in
class)

3. Second-order Linear Difference Equations with Constant Coefficient and Term

A simple variety of a second-order difference equation takes the form:

 y t+2 +a1 y t+1 +a 2 y t =c ……………………(7)

(7) is linear, non homogeneous and with constant coefficients (

a1 , a 2 ) and a constant term, c.

Particular integral: the solution of (7) is expected to have two components: a particular integral and a

complementary function. To get

yp , let
y t =k , evaluate y t+1 and y t+2 and substitute into (7)
and solve for k.

c
⇒ y p (=k )=
1+a1 +a2 (case of a1 +a2 ≠−1 )…………………….(8)

Example: find the particular integral of

y t+2 −3 y t+1 +4 y t =6 (Illustration to be provided in class)

⇒ y p (=k )=3

If
a1 +a2 =−1 , the trial solution y t =k breaks down, then we try y t =kt instead. (Illustration
to be provided in class).

c
⇒ y p (=kt )= t
a 1 +2 (a case where
a1 +a2 =−1 )………………………..(9)

Example: find the particular integral of

y t+2 + y t +1 −2 y t =12

(Solution to be provided in class)

⇒ y p =4 t

NB: this particular integral represents a moving equilibrium.

t
Complementary Function: try a solution of the form y t = Ab , evaluate y t+1 , y t+2 and substitute
trial solution into the homogeneous equation. (Illustration to be provided in class)

Case 1(Distinct Real Roots):

When
a1 ≻4 a2 , the roots of the characteristic equation are real and distinct. In that event, bt1
t
and b2 are linearly independent and the complementary function can simply be written as a linear
combination of these expressions:

t t
y c =A 1 b1 + A 2 b2 ……………………..(10)

Example: find the solution of

y t+2 + y t +1 −2 y t =12

(Solution to be provided in class)

Case 2 (Repeated Real Roots)

When a21 =4 a2 , the roots of the characteristic equation would be repeated, that is,

−a1
b=(b1 =b 2 )=
2 . If we express the complementary function in the form of (10), the two

components will collapse into a single term. Both

b1 and b2 should appear in the general solution of
the homogeneous difference equation because, just as in the case of differential equations, this general
solution must consist of two linearly independent parts, each with its own multiplicative arbitrary
t
constant. To supply the missing component-which should be linearly independent of the term A3 b -

use the old trich of multiplying bt by the variable t . The complementary function for the repeated
root case is:

t t
y c =A 3 b + A 4 tb

Example: Find the complementary function and solution of

y t+2 +6 y t +1 +9 y t =4

(Solution to be provided in class)

4. Economic Application
1. A model by B.J Ball and E. Smolelesky is based on the following system:
Ct =cY t−1 , K=σY t−1 , Y t =C t +K t −K t−1
 Where
Ct denotes consumption, K t is capital stock, Y is net national product and c and
σ are positive constants. Give an economic interpretation of the equations. Derive and solve

the second order difference equation

2. A model by J.R Hicks uses the following differential equation:
t
Y t+2 −(b+k )Y t+1 + KY t =a(1+g) , t=0,1,2. .. . where a, b, g and k are constants

Find the special solution for

Y ¿t of the equation
3. The Samuelson Multiplier Accelerator Iterative Model gives rise to the following set of
equations:
Y t =C t +I t +G0
Ct =γY t−1 (0≺γ≺1)
I t =α(C t −C t−1 )(α≻0 )
What second order difference equation is implied by this version of the model.
Find the solution to this equation.
4. Consider a two-sector model:
Y t =C t +I t
Ct =0 .8 Y t−1 +100
I t =200
Find an expression for
Y t when Y 0 1700
5. Consider the supply and demand equations
Qst =P t−1 −8
Qdt =−2 Pt +22
Assuming equilibrium, find expressions for Pt and Qt when
Po =11

References

1) Chiang A.C (1995) , Fundamental Methods of Mathematical Economics. 4 th edition, McGraw-

Hill.

2) Archibald G.C and Lipsey R.G ( 1976). An introduction to Mathematical Economics

 LINEAR PROGRAMMING
1.0 General Formulation of Linear Programs
Linear programming is the name used for problems in which we maximize or minimize
linear functions subject to linear inequalities.
In principle, any linear programming problem (often called an LP problem) can be solved if
the solution exists.
In general any linear programming problem with only two decision variables involves
maximizing or minimizing a linear function:
z=c 1 x 1 +c 2 x 2 (criterion function)

Subject to m constraints
a11 x 1 +a 12 x 2≤b1
a21 x 1 +a22 x 2≤b2
−−−−−−−−−−
am 1 x 1 +a m 2 x2 ≤b m (inequality constraints)

Usually, we also impose explicit non-negativity constraints on x 1 and x 2 :

x 1≥0 , x 2 ≥0 (non-negativity constraints)

1.1 Graphical Approach to Simple LP Problems

LP problems with only two decision variables can be solved by a simple geometric method.

Procedure:

a) Graph the feasible region

b) Compute the coordinates of the corner points
c) Substitute the coordinates of the corner points into the objective function to see
which gives the optimal value
 d) If the feasible region is not bounded, this method can be misleading. Optimal
solutions always exist when the feasible region is bounded, but may or may not
exist when the feasible region is unbounded

Let us consider the following Problem

Example: you need to buy some filing cabinets. You know that cabinet X costs $10 per unit,
requires six square feet of floor space, and holds eight cubic feet of files. Cabinet Y costs $20 per
unit, requires eight square feet of floor space, and holds twelve cubic feet of files. You have been
given $140 for this purchase, though you don’t have to spend that much. The office has room for
no more than 72 square feet of cabinets. How many of which model should you buy, in order to
maximize storage volume?

Solution: the problem is to:

3 x1 +6 x 2≤150
x 1 +0 .5 x 2 ≤22
x 1 +x 2≤27 . 5
max z=20 x 1 +30 x2 s.t x 1≥0 , x 2 ≥0

The output pair ( x1 ,x 2 ) is called admissible (or feasible) for the problem if all the five

constraints are satisfied.

For example: for the flour constraint, 3 x1 +6 x 2≤150 , if we use all the flour then
3 x1 +6 x 2=150 , and we call this the flour border. We can find similar borders for other

inputs.

The diagram below shows the straight lines that represent the flour border, sugar border and the
butter border:

(Diagram to be provided in class)

In order for ( x1 , x 2 ) to be admissible, it has to be on or below (to the ‘southwest’ of) each of

the three borders simultaneously.

Because constraints x 1≥0 , x 2 ≥0 restrict ( x1 , x 2 ) to the nonnegative quadrant, the set of

admissible pairs for the problem, called the admissible (or feasible) region, is the set S shown
in the diagram above.

The five corner points O , A , B ,C and D are called extreme points of the set S

Point B (5,22 .5) maximizes profits for the baker and the total profit is 775.

Practice Question: a firm produces small and medium television sets. The profit is 800 for each
small and 1000 for each medium television set. Each television has to be processes in three
different divisions. Each small TV requires respectively, 4, 2 and 2 hrs in divisions 1, 2 and 3.
The corresponding numbers for the medium TV sets are 2, 8 and 4. Suppose divisions 1 and 2
both have a capacity of at most 16hrs per day and division 3 has a capacity of at most 11 hours
per day. Formulate and solve the LP problem for the firm

2. Introduction to Duality Theory

Confronted with an optimisation problem involving scarce resources, an economist will often
ask: what happens to the optimal solution if the availability of resources changes?

For linear programs, the answer to this question is intimately related to the so-called Duality
Theory of LP.

Consider the baker’s problem in the previous example.

Suppose the baker were to receive (free of charge) one extra kilo of flour. How much would this
extra kilo add to his maximum profit? How much would an extra kilo of sugar contribute to
profit? Or an extra kilo of butter?

If the baker receives one extra kilo of flour, his flour border would become 3 x1 + 6 x 2=151 .
The admissible set S will expand slightly and point B will move slightly up along the butter
'
border. The new optimal point B will be at the intersection of the lines 3 x1 +6 x 2=151 and

x 1 + x 2=27 . 5
Solving these equations gives x 1=14 /3 and x 2=137 /6 . Then the criterion function becomes

equal to 20(14 /3)+30(137 /6 )=2335 /3=775+10 /3

So profit has risen by 10/3

If the baker receives an extra kilo of sugar, the admissible set will expand, and the optimal point
is still at B (Recall, at the optimum in the original problem, the baker had 5.75kilos of unused
sugar. Hence there is no extra profit)

An extra kilo of butter would give a new optimal point at the intersection of the lines
3 x1 +6 x 2≤150 and x 1 + x 2=28 . 5 . Solving these equations gives x 1=7 , x 2 =21. 5 with

20 x1 +30 x 2 =775+10 . Profit rises by 10.

These results can be summarized as follows:

 An extra kilo of flour would increase optimal profit by 10/3

 An extra kilo of sugar would increase optimal profit by 0
 extra kilo of butter would increase optimal profit by 10

The three numbers u¿1 =10/3 ,u ¿2=0 ,u ¿3=10 are related to the flour, sugar and butter
constraints respectively. They are the marginal profits from an extra kilo of each ingredient.

Suppose ( x1 , x 2 ) is an admissible pair in the problem, so that constraints are satisfied

Multiply (1) by 10/3, (2) by 0 and (3) by 10. Because the multipliers are all ¿ , the inequalities

(10/3)(3 x 1 +6 x 2 )≤(10 /3 )∗150

0( x1 +x 2 )≤0∗22
are preserved, i.e. 10 (x 1 +x 2 )≤10∗27 . 5

Add the inequalities, using the obvious fact that if A<B, C<D, and E<F, then A+C+E<B+D+F.

the result reduces to 20 x1 +30 x 2 ≤775

 1 ¿ ¿
Thus using the ‘magic numbers’ u1 , u2 , u3 defined before, we have proved that if ( x1 ,x 2 ) is
any admissible pair, then the criterion function has to be less than or equal to 775. Because
x 1=5 and x 2=22 . 5 give the value 775, we have in this way proved algebraically that

(5,22.5) is the solution.

NB: the pattern revealed in this example turns up in all linear programming problems

1 ¿ ¿
In fact the numbers u1 , u2 , u3 are solutions to a new LP problem called the dual of the original
problem.