Wollega University

College of Business and Economics

Department of Economics

Linear Algebra for Economists

(Econ 1022)

A Teaching Material for Distance Students

Prepared by:
ShumetAsefa
FirdisaBirru
KidusMarkos

Nekemte, Ethiopia
June, 2013
Course Description
This course is designed to introduce the students the basic concepts of matrices and
determinants: matrix operations, basic properties of determinants, matrix representation
of linear equations, methods of solving systems of linear equations, special determinants
and matrices and their use in Economics, input-output model and Linear Programming.
The course covers the basics to understand a wide variety of applications that use the
ideas of linear algebra, linear programming and their economic applications. The
emphasis is on concepts and problem solving.

Objectives:
The aim of this course is to equip students with a fundamental understanding of basic
concepts in Linear Algebra and its applications in economics. Up on the successful
completion of this course students will be able to:
 Solve linear equations and apply them in real – life problems in
economics
 Apply matrices to solve applied system linear equations in
economics and business
 Formulate and solve applied linear programming problems either
graphical or simplex methods for both maximization and
minimization problems
 Formulate and describe transportation problem
 Formulate optimization and linear programming problems in
microeconomics and macroeconomics matrix algebra.
 Explain and/or use Leontief Input-output model
 Solve linear programming problems using graphical and / or simplex
method
 Solve economic problems using different kinds of matrix operations.

Linear Algebra for Economists Lecture note, Wollega University, 1

Department of Economics
By:Kumadebis Tamiru
 CHAPTER ONE
1.0 MATRIX ALGEBRA
1.1Definition
A matrix is a rectangular array of numbers enclosed by a pair of parentheses [ ] or
a11 ,a12, a13 , ... a mn ,
brackets ( ) subject to certain rules of operation. A set of numbers
arranged in a rectangular array m rows and n columns like

a11 a12 a13 . . . a1n

a21 a22 a23 . . . a2n
a31 a32 a33 . . . a3n
, , , ,
, , , ,
, , , ,
am1 am3am3amn

and subject to certain rules of operation is called an mxn (read “mbyn”) matrix. The
numbers a11, a12,a13 - - - amn are called the elements (entries) of the matrix. We denote this
matrix by Aijor aij where i = 1, 2, 3… m…..rows
j = 1, 2, 3… n…..columns
The matrix above has m rows and n columns. A ij denotes the elements of the matrix lying
in the ith row and jth column and we call this element as the (i,j)th element of the matrix.

The Setting and Overview

Matrices are mathematical tools which are useful in the study of systems of linear
equations. Matrices are arrays of numbers, and they can be manipulated and combined in
much the same way as we handle numbers. In this chapter we are going to develop the
“algebra” of matrices, i.e amethod of adding, subtracting and multiplying matrices when
these operations make sense.
In real life, we do not have to deal with a single magnitude but with a set of magnitudes.
In analysis, we treat such set of magnitudes as a single entity or a single object of
thought, abstracted by numbers.

Linear Algebra for Economists Lecture note, Wollega University, 2

Department of Economics
By:Kumadebis Tamiru
Suppose there are 3 brothers: A, B and C in a family and that,
A has a set of 3 pants, 4 shirts, 2 Trousers and 1 tie
B has a set of 5 pants, 2 shirts, 3 Trousers and 2 ties
C has a set of 4 pants, 3 shirts, 5 Trousers and no tie
Now, we can arrange this data systematically in the following convenient way.

Brother Pant Shirt Trousers Tie

s s s (Tr) (T)
(P) (S)
A 3 4 2 1 1st row
B 5 2 3 2 2nd row
C 4 3 5 0 3rd row

1st column 2nd column3rd column4th column

The above system comprises 3 rows and 4 columns. (Note that we write rows first and
then after columns). Clearly, therefore, 1st column gives us the total number of pants
(3+5+4 = 12) that A, B and C together have in their family. While 2 nd, 3rd and 4th columns
enumerate the total number of shirts, Trousers and Ties those three brothers together have
in their family respectively.

A set of numbers written in a particular form of rows and columns enclosed by [ ]

parentheses or () brackets are called a matrix.

Example 1

The matrix 2 3 8 has two rows and three columns.

2 -1 3 and hence it is a 2x3 matrix.

Similarly, 4 3
2 5 is a 3x2 matrix
7 4

1 2
3 4 is a matrix of order 2. (2x2 matrix)

1 2 3 is a 1x3 matrix and

Linear Algebra for Economists Lecture note, Wollega University, 3

Department of Economics
By:Kumadebis Tamiru
 1
2 is a 3x1 matrix
3

Example 2

In the matrix

1 2 3
A= 5 6 7
8 9 10

a11 = 1 an entry located in the first row and first column

a32 = 9 an entry located in the 3rd row and 2nd column
a23 = 7 an entry located in the 2nd row and 3rd column

Remark1: unless and otherwise stated, we shall consider matrices over the field of real
numbers for simplicity.
Remark2: A matrix is simply an arrangement of elements and has no numerical value,
it is wrong to say 8 9 is greater than

1.2. Various Types of Matrices

We shall now define some useful matrices.
1. Row matrix: a matrix having only one row is called a row matrix (a row vector).
Example: 1 2 3 4 is a 1x4 row matrix and
a11 a12 a13 . . . a1nis a 1xn row matrix
2. Column Matrix: a matrix having only one column is called a column matrix (a
column vector)
1 a11
2 a21
Example: 3 is a 4x1 column matrix and a31 is nx1 column matrix.
4 .
.
an1

3. Square matrix: A matrix in which the number of rows is equal to the number of
columns is called a square matrix.
Linear Algebra for Economists Lecture note, Wollega University, 4
Department of Economics
By:Kumadebis Tamiru
 Example: 1 2
3 4 is a 2x2 square matrix

1 2 3
4 5 6 is a 3x3 square matrix
7 8 9
4. Null (Zero) matrix:A matrix of any order whose all elements are zero is called a
null (zero) matrix and denoted by 0.
Example: 0 0
0 0 is a null matrix of order 2 (2 x 2)

0 0 0
0 0 0 is a2x3 null matrix.
5. Diagonal matrix: A square matrix whose all elements except the main diagonals
are zero is called a diagonal matrix.
Example:
A= 2 0 1 0 0 are diagonal matrices of order
0 3 and B= 0 4 0 2 and 3 respectively.
0 0 9
Main
diagonal

6. Scalar Matrix: A square matrix whose all elements except the main diagonals are
zero and the main diagonal elements are all equal is called a scalar matrix
Example:
2 0 4 0 0
0 2 & 0 4 0 are scalar matrices of
order 2 & 3
0 0 4 respectively.
7. Unit (identity) matrix: A square matrix whose main diagonal entries are all unity
(1) and the other elements are zero is called a unit or Identify matrix. A unit
matrix of order n is denoted by In.
Example:
Linear Algebra for Economists Lecture note, Wollega University, 5
Department of Economics
By:Kumadebis Tamiru
 1 0 1 0 0 1
0 0 0
I2 = 0 1 I3 = 0 1 0 I4 = 0
1 0 0
0 0 1 0
0 1 0
0
0 0 1

are unit matrices of order 2, 3 and 4 respectively.

8. Singular and Non- singular Matrices: with every square matrix A of certain
order, we associate a determinant /A/ or det A of the same order formed by the
corresponding elements of A. If /A/= 0, the matrix A is called a singular matrix
and if /A/  0, the matrix A is called a non – singular matrix
Definition:- if A = a b , then /A/ = ad – bc (details will be
c d
discussed later on)

Example:
i. If A = 4 6
2 3 then by definition /A/ = 4x3 -2x6 = 12 – 12 = 0
and therefore, A is a singular matrix of order 2
ii. If A = 4 5
6 8 , then /A/ = 4x8 – 6x5 = 32 – 30 = 2
/A/  0.Hence, A is a non – singular matrix of order 2.
9. Symmetric and skew – symmetric matrices: A square matrix A = (aij) is said to be
symmetric if aij = aji for all i and j, i.e., a12 = a21 ,a13=a31, a23 = a32, soon.
Example:
2 7 -5 a b c
7 8 -4 and b d a are symmetric matrices of
-5 -4 3 c a e order 3

A square matrix A is said to be skew – symmetric If aij = - aji for all i and
j, i.e., if a21 = - a12, a13 = - a31 , a32 = -a23 and soon and the main diagonal
elements a11 = a22 =a33…ann= 0

Example
0 -2 -4 0 a b
2 0 -6 & -a 0 -c are skew symmetric matrices of
4 6 0 -b c 0 order 3.

Linear Algebra for Economists Lecture note, Wollega University, 6

Department of Economics
By:Kumadebis Tamiru
 10. Triangular matrix: A square matrix whose elements a ij = 0 whenever i<j is called
a lower triangular Matrix. Similarly, a square matrix whose elements a ij = 0
whenever i> j is called an upper Triangular Matrix.
Example:
2 0 0
8 10 0 and 4 0
14 16 18 10 15 are lower triangular matrices

2 4 6 5 10
0 8 10 and 0 15 are upper triangular matrices.
0 0 12

11. Augmented matrix: A matrix which includes both coefficients and constants of a
system of linear equations is called an augmented matrix.
Example
Given: 3x+ 2y +7z = 18
9x+ 3y +3z = 28
3x+6y + 5z = 20
The above system of linear equations can be written in matrix form as
3 2 7
9 3 3 = coefficient matrix (A)
3 6 5
x
y = Variable matrix (X)
z

18
28 = constant matrix. (C)
20

3 2 7 18
9 3 3 28 thus, this is an augmented
matrix
3 6 5 20

Linear Algebra for Economists Lecture note, Wollega University, 7

Department of Economics
By:Kumadebis Tamiru
1.3 Rules of Operations on Matrices (Algebra of Matrices)
Under the operations on matrices, we shall discuss scalar multiplication of a matrix,
addition of matrices, subtraction and multiplication of a matrix by a matrix.
a. Equality of matrices
Two matrices A and B are said to be equal if and only if
i. A and B have the same order (dimension) and
ii. Each element of A is equal to the corresponding element of B

Example

3 5 7 3 5 7

1. Let A = 4 8 9 2 x3 and B= 4 8 9 2 x3 , then A = B

as they satisfy the conditions for equality having the same dimension (2x3) and
a11 b11 , a 22 b22
corresponding entries are equal
, and so on.

1 2 3 1 2 3
Let A= 4 5 6 and B= 7 5 6
A  B even if matrix A& B have the same order(2x3) due to 4 different from 7.

2. The following two matrices are also not equal

A= 1 2 3 B= 1 2 3
5 6 7 and 4 6 7

A  B as a 21 = 5 and b21 = 4 are different corresponding entries.

3. Suppose the following two matrices are equal. Find the value of the variable t, a, b
and m.
4t b-2 t2+4 -7
log 81
A= a 9 2x2 and B = 4 m 2x2

Solution:
Since the matrices are equal they must have equal dimension and corresponding
entries of the matrices should be equal.
81
a) 4t = t2 +4 b) b- 2 = -7 c) log a = 4 d) 9 = m
Linear Algebra for Economists Lecture note, Wollega University, 8
Department of Economics
By:Kumadebis Tamiru
 t2 -4t+4 = 0 b = -7+2 a4 = 81
4
(t – 2) 2 = 0 b = -5 a4 = 3
t=2 a=3
Thus, A = B if and only if t=2, b = -5, a =3 and m = 9.

b. Scalar multiplication of a matrices

Let be a scalar number and A be a matrix. Then the product of the scalar number and
the matrix A, written as A, is the matrix obtained by multiplying each element of A by.

Example

If  = 4 and A = 1 2 4
2 -3 4 , then

A= 4 1 2 4 4x1 4x2 4x4 4 8 16

2 -3 4 = 4x2 4x-3 4x4 = 8 -12 16

c. Addition and subtraction of two matrices

If two matrices A and B are of the same order, i.e., if they have same number of rows and
columns, then the sum of A and B or subtraction of B from A (A+B) or (A – B) is the
matrix obtained by adding or subtracting the corresponding elements of A and B. Note
that A – B = A+ (-B), adding the negative of B to A.

Example

If A = 1 2 3 B= 2 3 4

4 5 6 2 x3 and 5 6 7 2 x3
,

Then A + B = 1+2 2+3 3 -4 3 5 7

2 x3
4+5 5+6 6+7 = 9 11 13

And A – B = 1-2 2-3 3+4 -1 -1 -1

Linear Algebra for Economists Lecture note, Wollega University, 9

Department of Economics
By:Kumadebis Tamiru
 4-5 5-6 6-7 = -1 -1 -1

d. Matrix multiplication (multiplication of a matrix by a matrix)

The product AB of two matrices A and B is defined only when the number of columns of
the first matrix A is the same as the number of rows of the second matrix in B and by
definition the product AB is a matrix C of order mxp if A and B were of order mxn and
nxp respectively. The following example will give the rule to multiply two matrices.

a11 a12 a13 b11 b12

Let A = a21 a22 a23 and B = b21 b22
2 x3 b31 b32 3.. x 2

Order of A is 2x3 while it is 3x2 for B

 The product AB is defined since the numbers of the pre multiply matrix
columns of A is (3) and numbers of rows of matrix B (3) are the same.
 When the numbers of the post multiply columns of the first matrix and the
numbers of rows of the second matrix (B) are equal, then the two matrices A
and B are conformable for the product AB.

Let us see now how to multiply the above two matrices A and B.
C = AB = a11 a12 a13 b11 b12
a21 a22 a23 x b21 b22
b31 b32
C = AB = a11 b11+ a12b21+ a13 b31 a11 b12 + a12 b22 + a13 b32
a21 b11+ a22b21+ a23 b31 a21 b12 + a22 b22 + a23 b32
n

a ik b kj
C = ail blj + ai2 b2j + - - - + ain bnj = k 1

C= C11 C12 Where C11 = a11 b11 + a12 b21 + a13b31

C21 C22 C12 = a11 b12+a12 b22 + a13 b32
C21 = a21 b11+ a22 b21 + a23b31
C22 = a21 b12+ a22 b22 + a23 b32

Linear Algebra for Economists Lecture note, Wollega University, 10

Department of Economics
By:Kumadebis Tamiru
Entries of the product matrix C = AB is obtained as follows,
i. C11 is obtained by multiply elements of the first row of A with corresponding
elements of the first column of B and add
ii. C 12 is obtained by multiply elements of the first row of A with corresponding
elements of the second column of B and add.
iii. C21 is obtained by multiply elements of the second row of A with
corresponding elements of the first column of B and add.
iv. C22 is obtained by multiply elements of the second row of A with
corresponding elements of the second column of B and add.

Remark:
1. If A and B are two matrices, then AB may not be equal to BA (even though both
are defined).

Example

A= 2 2 B= 2 0
0 0 2x2 & 0 0 2x2

AB = 2 2 x 2 0 2x2 + 2x0 2x0 + 2x0 4 0

0 0 0 0 = 0x2 + 0x0 0x0 + 0x0 = 0 0

BA = 2 0 2 2
0 0 x 0 0 =
2x2 + 0x0 2x2 + 2x0 4 4
0x2 + 0x0 0x2 + 0x0 = 0 0
Thus, this example confirm thatAB  BA.
2. If product AB is defined, then it is not necessary that BA must also be defined.

Example

If A is of order 2x3 and B is of order 3x1 then clearly AB is defined, but BA is

not defined as the number of columns of B is different from the number of rows

Linear Algebra for Economists Lecture note, Wollega University, 11

Department of Economics
By:Kumadebis Tamiru
 of A (commutative property doesn’t hold true in matrix multiplication): AB 
BA.
3. Note the following properties
i. Associative property: A (BC) = (AB)C
ii. Distributive property: A(B+C) = AB+ AC
(A+B) C = AC + BC

Example

Verify remark 3 i for the following matrices

A= -1 0 5 ,B = 1 7 0 -1 -1
7 -2 0 2x3 -3 -1 0 , C= 2 0

1 0 5 3x 3 0 4 3x2

Solution:
i. Checking for the right side of equality: (AB)C
Since the number of columns of A is equal to the number of rows of B, we can
obtain the product AB as follows:

AB = (-1x1) + (0x-3) +(5x1) (-1x7) + (0x -1) +(5x0) (-1x0)+(0x0) + (5x5)

(7x1) + (-2x-3) + (0x1) (7x7)+(-2 x-1) +(0x0)(7x0)+(-2x0)+(0x5)
AB = 4 -7 25
13 51 0 2x3

So, (AB) C = 4 -7 25 -1 -1 -18 96

13 51 0 x 2 0 = 89 -13 2x2
0 4
4 -7 0 -1 -1 13 -1
Again BC = -3 -1 0 x 2 0 = 1 3
1 0 5 0 4 -1 19

Linear Algebra for Economists Lecture note, Wollega University, 12

Department of Economics
By:Kumadebis Tamiru
So, A (BC) = -1 0 5 13 -1 -18 96
7 -2 0 x 1 3 = 89 -13 2x2
-1 19
Therefore, A (BC) = (AB) C and condition of associative property.

Example

If A = 1 -2 3 2 3 -1
2 4 -1 and B = 1 0 2
-3 4 5

Find AB and BA if possible

Solution:

Matrix A is a 2x3 matrix and B is a 3x3 matrix so that the product AB is

defined and it is a 2x3 matrix.

AB = 1 -2 3 2 3 -1
2 4 -1 x 1 0 2
-3 4 5

(1x2+(-2x1) + (3x-3) (1x3) (-2x0)+(3x4) (1x-1) +(-2x2) +3x5)

= (2x2)+(4x1) +(-1x-3) (2x3)+(4x0)+(-1x4) (2x-1) +(4x2) +(-1x5)

= -9 15 10
11 2 1

Note: that the product BA is not defined in this case, since the number of
columns in B  the number of rows in A. since the number of columns of the
first matrix B is different from the number of rows of the second matrix A we
cannot multiply B with A.)
Linear Algebra for Economists Lecture note, Wollega University, 13
Department of Economics
By:Kumadebis Tamiru
Some more illustrative examples

Attempt the following questions:

1. If A = 0 2 3 7 6 3
2 1 4 2x3 and B = 1 4 5 2x3

Find the value of 2A +3B

2. Find x and y, if x+y = 7 0 3 0

2 5 & x-y = 0 3

3. Find the product AB of two matrices A& B

1
Where A = 0 and B = 2 4 9 6 5 0
7 1x6
8 4x1

4. A man buys 8 kgs of Mangoes, 10 kgs of apples, and 4kgs of bananas. Managoes
cost birr 4 per kg, apples 9 birr per kg and bananas cost birr 6 per kg, represent the
quantities bought by a row matrix and the prices by a column matrix and hence
obtain the total cost.
5. If A = 2 -1 and B = 2 , find the matrix C of suitable
4 -3 -3 order such that AC = B.

here are solution for the above questions:

1. We have 2 A = 2x 0 2 3 0 4 6
2 1 4 = 4 2 8

And 3B = 3 x7 6 3 21 18 9
1 4 5 = 3 12 15
Hence, 2A +3B = 0 4 6 + 21 18 9
4 2 8 3 12 15
= 0+21 4+18 6+9
4+3 2+12 8+15
Linear Algebra for Economists Lecture note, Wollega University, 14
Department of Economics
By:Kumadebis Tamiru
 = 21 22 15
7 14 2

2. If x +y = 7 0 3 0
7 5 and x – y = 0 3

We have x +y = 7 0 1
2 5

And x-y = 3 0
0 3 2

Adding 1 and 2, we get

( x+ y) + ( x-y) = 7 0 3 0
2 5 + 0 3

10 0
2x = 2 8 => multiplying both sides by ½

10 0
x= ½ 2 8

x= 5 0
1 4

Now we can substitute the value of x = 5 0

1 4 to obtain the value of y.
From 1, x +y = 7 0
2 5

5 0 7 0
1 4 +y = 2 5

y= 7 0 5 0
2 5 - 1 4

y= 2 0
1 1

1
0 xB= 2 4 9 6 5 0
3. A = 7 1x6
8 4x1

Linear Algebra for Economists Lecture note, Wollega University, 15

Department of Economics
By:Kumadebis Tamiru
 1
AB = 0 x 2 4 9 6 5 0
7
8

1x2 1x4 1x9 1x6 1x5 1x0

AB = 0x2 0x4 0x9 0x6 0x5 0x0
7x2 7x4 7x9 7x6 7x5 7x0
8x2 8x4 8x9 9x6 8x5 8x0 4x6

AB = 2 4 9 6 5 0
0 0 0 0 0 0
14 28 63 42 35 0
16 32 72 48 40 0

4. Let A be the row matrix representing the quantities bought and B be the column
matrix representing the price
4

A= 8 10 4 1x 3 and B = 9
6 3x1

The total cost is given by the element of the product AB which is 1x1 matrixes.

4
AB = 8 10 4 9 x
6
AB = 8x4 + 10x9 + 4x6

AB = 32 + 90 + 24

AB = 146

Hence, the required total cost is birr 146.

5. If A = 2 -1 2
4 -3 and B = -3 , find C such that AC = B

Since A is a 2x2 matrix and AC i.e B Is a 2x1 matrix

Therefore, C must be a 2x1 matrix and let C = x

y

Linear Algebra for Economists Lecture note, Wollega University, 16

Department of Economics
By:Kumadebis Tamiru
Then, AC = B gives = 2 -1 x 2
4 -3 x y = -3

2x –y 2
4x-3y = -3

Thus, by equality of two matrices, we get

2x – y = 2
4x – 3y = -3
-4x + 2y = -4 Multiplying the 1st equation by - 2
4x – 3y = -3
-y = -7
y=7
Substituting y = 7 in the first equation, we get:
2x – 7 = 2
2x = 9
x= 9/2
Hence, the required matrix is C = x 9/2
y = 7
Self help questions
1. A man buys 5 dozens of oranges, 6 dozens of apples and 3 dozens of bananas.
Oranges cost birr 15 per dozen, apples birr 12 per dozen and bananas birr 8 per
dozen. Represent the quantities purchased by row matrix and prices by column
matrix and obtain the total cost.
2. Three persons buy cold drinks of different brands A,B and C. The first person
buys 12 bottles of A, 5 bottles of B and 3 bottles of C. The second person buys 4
bottles of 4,6 bottles of B and 10 bottles of C. The third person buys 6 bottles of
A, bottles of B and 9 bottles of C. Represent these information in the form of a
matrix. If each bottle of A costs birr 4, each bottle of B costs birr 5 and each bottle
of C costs birr 6, then using matrix operations find the total sum of money spent
individually by the three persons for the purchase of cold drinks.
3. A manufacturer produces three products A, B, C which are sold in Nekemte and
Gimbi cities. The annual quantity sold of these products are given below:
Products
Linear Algebra for Economists Lecture note, Wollega University, 17
Department of Economics
By:Kumadebis Tamiru
 City
A B C
Nekemte 9000 12,000 8700
Gimbi 5000 7500 15000

If the selling prices of A, B, C per unit be birr 2, 3 and 4 respectively. Calculate

the total revenue in each city by using matrix.
1.4. Transpose of a Matrix
Let A be a matrix. The matrix obtained from A by interchanging its rows and columns is
called the transpose of A, i.e., as the name implies rows become columns or columns
become rows.
The transpose of matrix A is denoted by At.
Let us see the following examples on row to find transpose of a matrix.

i. If A = 1 0 2 then transpose of A = At = 1 2
2 3 4 2x3 0 3
2 4 3x2

ii. If B = a b c a d g
d e f , then Bt = b e h rows become columns.
g h i 3x3 c f I 3x3

It can be verified that for two matrices A and B,

a) (At)t = A
b) (A+B)t = At +Bt
c) (AB)t = Bt At

Example:

For the following two matrices verify that (At)t = A and (A+B)t = At+Bt
A= 1 2 3 2 3 4
4 5 6 , and B = 1 8 6

Solution:

a.) At = 1 4 2 1
t
2 5 and B = 3 8

Linear Algebra for Economists Lecture note, Wollega University, 18

Department of Economics
By:Kumadebis Tamiru
 3 6 4 6

(At)t = 1 2 3
5 5 6 =A

Thus, (At)t = A

b) A+B = 3 5 7
5 13 12

(A+B)t = 3 5 3 5
5 13 and At +Bt = 5 13
7 12 7 12

Hence, (A+B)t = At +Bt

Note that (AB)t AtBt for two comfortable matrices

Example

Verify that (AB)t=Bt At and (AB)t AtBt

Give A = 1 2 3 and B = 4 7
4 5 6 2x3 5 8
6 9 3x2

Solution:
At = 1 4 4 5 6
2 5 Bt = 7 8 9 2x3
3 6 3x2

So, AtBt = 1x4+4x7 1x5+4x8 1x6 +4x9 32 37 42

2x4+5x7 2x5+5x8 2x6+5x9 = 43 50 57
3x4+6x7 3x5+6x8 3x6+6x9 54 63 72 3x3

andBtAt = 4 5 6 1 4 4x1+5x2+6x3 4x4+5x5+6x6

7 8 9 x 2 5 = 7x1+8x2+9x3 7x4+8x5+9x6
3 6
Linear Algebra for Economists Lecture note, Wollega University, 19
Department of Economics
By:Kumadebis Tamiru
BtAt = 32 77
50 122 2x2

1 2 3 4 7 1x4+2x5+3x6 1x7+2x8+3x9
Again (AB)t= 4 5 6 x 5 8 = 4x4+5x5+x6x 4x7+5x8+6x9
6 9

(AB)t = 32 50t
77 122

So, (AB)t = 32 77
50 122 2x2

Hence, (AB)t = BtAt = 32 77

50 122

But, (AB)t AtBt, having different orders which is 2x2 and 3x3 respectively.
Note the following points:
A matrix ‘A’ with a property A = At is called a symmetric matrix.
Example:
If A = 2 8 18 2 8 18
8 14 10 , then At = 8 14 10
18 10 16 18 10 16

Write any three symmetric matrices. Based on the above example.

i. A square matrix with the property of A = A2 is known as idempotent matrix.
ii. A square matrix with the property of A2 = I is called involuntary matrix.

Linear Algebra for Economists Lecture note, Wollega University, 20

Department of Economics
By:Kumadebis Tamiru
Self Help question 2

1. Let A = 1 1
0 1

a. Find a matrix B such that AB = 1 0

0 1

b. Find a matrix D such that AD = 2 0

4 3
2. For each of the following matrix, show that A+At = 0

a. 0 1 4 0 6 -4 0 -1 2
-1 0 7 b. -6 0 8 c. 1 0 5
-6 -7 0 4 -8 0 -2 -5 0

3. If A = ( 1 2 3) and B = 4
5 , then
6
Verify that (AB)t = Bt At.
4. If A = 2 1 1 -2
3 4 and B = -1 1 , then verify (AB)t = Bt At
5. Let A = 1 0 0
4 2 3 , show that A2 – 3A+2I = 0
2 0 1

6. Find the values of x,y,z if the matrix A= 0 2y z

x y -z
x -y z
Linear Algebra for Economists Lecture note, Wollega University, 21
Department of Economics
By:Kumadebis Tamiru
obeys the law At. A = I where I is an identity matrix.

1.5 Determinants
If A is a square matrix with entries (elements) from the field of complex numbers, then
determinant of A is some complex number. This will be denoted by det A or /A/.
Determinant is a numerical value of a square matrix.
a11 a12 a13 . . . a1n
a21 a22 a23 . . . a2n
If A= , , , ,

, , ,

An1 an3an3ann

a11 a12 a13 . . . a1n

a21 a22 a23 . . . a2n
, , , ,

thendet A = /A/= , , , ,

an1 an3an3ann

Note: the following points.

i) det A or /A/ is defined for square matrix A only
ii) det A or /A/ will be defined in such a way that A is invertible iffA0
iii) The determinant of an nxn matrix will be called determinant of order n.
1.5.1. Determinantof Order One
Let A = (a11) be a square matrix of order one. Then detA = a11
Example: letA = (3), then detA = /A/ = 3 by definition.
1.5.2. Determinant of Order Two
Let A = a11 a12 be a square matrix of order two.
a21 a22

then we define det A as: /A/ = a11 a22 – a12 a21

Linear Algebra for Economists Lecture note, Wollega University, 22

Department of Economics
By:Kumadebis Tamiru
If B = a b
c d , then B= ad – bc

Example: find det A or /A/ if A= 3 4

5 6
Solution: Det A = /A/ = (3x6)-(4x5) = 18 -20 = -2
1.5.3. Determinantof Order Three or More
The determinant of order three or more is determined using two methods (Laplace’s
expansion method and Sarru’s method).
1.5.3.1. Laplace’s Expansion Method
Laplace’s expansion method to determine the determinant of order three or more requires
the knowledge of minors and cofactors which will be covered in later topics. Thus, I will
use the formula which defines the determinant of a square matrix.
+ - +
Let A= a11 a12 a13
a21 a22 a23 be a 3x3 square matrix.
a31 a32 a33

Then we define determinants of A as,

det A = a11 (a22 xa33 -a23 xa32 ) - a12 (a21 xa33 - a23 a31 ) +a13 (a21 xa32 -a22 xa31)
The above definition may be explained as follows:
 The first bracket is determinant of matrix obtained after removing first row and
first column. (a22 a33 -a23 a32 )
 The second bracket is determinant of matrix obtained after removing first row and
second column. (a21 a33 -a23 a31)
 The third bracket is determinant of matrix obtained after removing first row and
third column (a21 a32 -a22 a31)
 The elements before three brackets are first, second and third element respectively
of first row with alternate positive and negative signs.

Linear Algebra for Economists Lecture note, Wollega University, 23

Department of Economics
By:Kumadebis Tamiru
Example1

Find the determinant of the following matrix

1 2 3
A= 4 5 -6
7 8 9
Solution:
The first bracket in the definition of det A is determinant of the matrix after deleting the
first row and first column. Of matrix A

1 2 3 5 -6
4 5 -6 = 8 9 = (5x9) – (-6x9) = 45+54 = 99
7 8 9

The second bracket is determinant of the matrix after removing the first row and the
second column of matrix A.

1 2 3 4 -6
4 5 -6 = 7 9 = (4x9) – (-6x7) = 36 +42 = 78
7 8 9

The third bracket is determinant of the matrix after removing the first row and third
column of matrix A.

1 2 3
4 5 -6 4 5
7 8 9 = 7 8 = (4x8) – (5x7) = 32 -35 = -3

Linear Algebra for Economists Lecture note, Wollega University, 24

Department of Economics
By:Kumadebis Tamiru
So, det A = 1 (99) -2 (78) +3 (-3) = - 66

Example 2

Find the determinants of the following matrices.

2 -1 3 -4 -5 6
i) A= 1 2 -4 ii) B = 7 8 9
-3 1 5 1 2 3

Solution:
i.det A = 2 2 -4 1 -4 1 2
1 5 - (-1) -3 5 + 3 -3 1
= 2 (2x5) – (1x-4) +1 (1x5) – (-3x-4) +3(1x1) – (2x-3)
=2 (10+4) + (5-12) +3 (7)
= 28 -7 +21
= 42
1.5.4 Properties of Determinants
List of some important properties of determinants
1. If two rows (or columns) of a matrix are interchanged, its determinant retains its
absolute value but changes its sign,
Let A = a b
c d , then /A/ = ad – bc

and B= c d
a b , interchanging the two rows of A

/B/ = bc – ad = - (ad-bc) = -/A/

2. If rows are changed in to columns or columns into rows, the determinate remains
unchanged.
Let A = a b
Linear Algebra for Economists Lecture note, Wollega University, 25
Department of Economics
By:Kumadebis Tamiru
 c d , then /A/ = ad – bc

B= a c
b d rows are changed into columns

/B/ = ad – bc = /A/

Hence, the result follows.

3. If two rows (or columns) are identical, the determinant is zero (vanishes).
Let A = a b
a b , then /A/ = ab – ab = 0

Similarly if A = a1 a2 a3
a1 a2 a3 , then /A/ = 0
c1 c2 c3

4. If any row (or column) is multiplied by a real number k, then the determinant
obtained is k times the original determinant.
Let A = a b
c d , then /A/ ad – bc

and let B = ka kb
c d = multiplying the first row by K.

then /B/ = kadkbc = k (ad – bc) = k /A/

5. If to any row (or column) is added k times a the corresponding elements of

another row cor column), the determinant remains unchanged.
i.e if A= a b
c d and

B= a+kc b+kd
c d adding k times of the second row to the first row

Now. /B/ = (a+kc) d- (b+kd) c

Linear Algebra for Economists Lecture note, Wollega University, 26

Department of Economics
By:Kumadebis Tamiru
 = ad + kdc – bc – kdc
= ad –bc = /A/

Self test

Prove (show) the above five properties of determinants by taking your own examplesof
having two and three orders.

1.6 Minors and Cofactors of the Element of a Determinant

1.6.1 Minors
A minor is a sub determinant of a given order determinant obtained by deleting
(canceling) the ith and jth column of the required element and evaluating the n-1 order
determinant Minors are denoted by Mij. Mll is a minor of the element located in the first
row and first column and obtained by canceling the first row and first column.
Similarly, M23 is a minor of the element located in the 2 nd row and 3rd column and
obtained by canceling the 2nd row and 3rd column of the given determinant.

Example1

If A = a b
c d , then find the minors of each element.

Solution:
/A/ = a b= ad-bc
c d
M11 (minor of a) = a b =d
c d
M12 (minor of b) a b
c d =c

M21 = (minor of c) a b
c d =b

Linear Algebra for Economists Lecture note, Wollega University, 27

Department of Economics
By:Kumadebis Tamiru
M22 (minor of d) a b
c d =a

Example 2

If A = 1 -2
-3 -4 , then find minors of each element

Solution:

/A/ = 1 -2 M11 = -4 M21 = -2

-3 -4 M12 = -3 M22= 1

Example 3

If A = a1 b1 c1
a2 b2 c2 , find the minors of each
a3 b3 c3 element and evaluate then.
Solution:
The minor of any element in /A/ is a determinant of second order obtained by omitting
from /A/ the row and column in which that element occurs. Thus minors by a 1, b1, a2, b2,
c2
M11 = b2 c2 a2 c2 a2 b2 a2 b2
b3c3 , m12 = a3 c3 , M13 = a3 b3 , M21 = a3 b3, so on.

Example 4

Write the minors of the elements of the first row of the following determinant and hence
evaluate them.
/A/ = 1 2 3
-4 3 6

Linear Algebra for Economists Lecture note, Wollega University, 28

Department of Economics
By:Kumadebis Tamiru
 2 -7 9

Solution:
M11 = 3 6
-7 9 = 27 – (-42) = 69

-4 6
M12 = 2 9 = -36 -12 = -48

-4 3
M13 = 2 -7 = 28 -6 = 22

Use the same procedures to find minors of the 2nd and 3rd row elements
1.6.2. Cofactors
The cofactor of any element in /A/ is the minor of that element in /A/ with proper sign
depending on the numbers of the row and the column in which the element occurs. If an
element occurs in ith row and jth column in /A/, then the cofactor of the element = (-1) i+j
x
(Minor of the element), cofactor of the element is denoted by Cij.
Thus, Cij = (-1) i+j x Mij

Example 1

If A = 2 -3
4 3 , find the cofactors of each element.
Solution:
Recalling the method of finding minor of an element, we obtain
M11 = 3, M21 = -3 and
M12 = 4 M22 = 2
After obtaining the minors of the elements, we simply substitute them in the formula to
find the corresponding cofactors.
C11 = (-1) 1+1 .M11 = 1x3=3
C12 = (-1) 1+2 .M12 = -1x4 = -4
Linear Algebra for Economists Lecture note, Wollega University, 29
Department of Economics
By:Kumadebis Tamiru
 C21 = (-1) 2+1 . M21 = -1x(-3) = 3
C22 = (-1) 2+2 .M22 = 1x2 = 2
In matrix form the cofactors can be put as follows

cA= 3 -4
3 2

Example2

If A = 1 2 3
-4 3 6 , find the cofactors of each element
2 -7 9

Solution:
First find the minors of each element and multiply each minor by (-1) i+j obtain cofactors.
M11 = 3 6
-7 9 = 27 – (-42) = 69
C11 = (-1) 1+1 M11 = M11 = 69

-4 6
M12 = 2 9 = -36 -12 = -48
C12 = (-1) 1+2 xM12 = -1x (-48) = 48
M13 = -4 3
2 -7 = 28 -6 =22
C13 = (-1) 1+3 xM13 = M13 = 22

Self test

Obtain the cofactors of the 2nd and 3rd row elements in a similar manner.
Note: that in the identity Cij = (-1) i+j xMij,

Linear Algebra for Economists Lecture note, Wollega University, 30

Department of Economics
By:Kumadebis Tamiru
 i. The minor and cofactor of an element is equal if the sum of row and column in
which it is located is even and
ii. The cofactor of an entry is the negative of its minor if the sum of row and
column in which the element is located is odd.

Exercise

1. Find the determinants of the following matrices

1 2 3 2 5 7 1 1 1
A= 2 4 6 B= 3 6 10 C= 1 1+x 1
3 5 7 8 4 15 1 1 1+y

2. Solve for x if

1 1 1 x+1 2 3
i. 1 x 1 =0 ii. 3 x+2 1 =0
1 1 x 1 2 3

x 1 1 x-5 2
iii. 1 x 1 = 0 iv. -3 x =0
1 1 x
3. Find the minors and cofactors of each element for the following Matrices
1 2 0 2 5 4 46 32 15
i. A = 0 1 2 ii) B= 1 4 3 iii) C = 27 21 9
1 0 6 8 10 2 17 16 7

1.7. Adjoint of A Square Matrix

Let A = (aij) be a square matrix of order n. Then /A/ is the determinant of A of the same
order n formed by the corresponding elements of A. If C ij be the cofactors of aij in A, then
the transpose of the matrix (Cij) obtained by replacing the elements of A by their
Linear Algebra for Economists Lecture note, Wollega University, 31
Department of Economics
By:Kumadebis Tamiru
corresponding cofactors in A is called the adjoint (or adjugate) of the matrix A and it is
denoted by adj A,

Follow the following steps to obtain the adjoint of a square matrix.

Step 1: Obtain the cofactors of each element of the square matrix A
Step 2: From a cofactor matrix by replacing entries of the original matrix by the
corresponding cofactors
Step3: Transposing the cofactor matrix gives the adjoint matrix

Example 1

Let A= 1 2
3 4 , then find adj A.
Solution:
Step1: C11 = 4, C12 = -3, C21 = -2, C22 = 1
Step2: Cofactor of matrix A = CA = 4 -3
-2 1
Step3: Transposing the cofactor matrix, we get
CAt = 4 -2
-3 1 = adjA.
Hence, adj A = CAt = 4 -2
-3 1

Example2

For the following matrix A, prove that A. adjA= 0

1 -1 1
A= 2 3 0
18 2 10

Solution:-
We have A = 1 -1 1
2 3 0
Linear Algebra for Economists Lecture note, Wollega University, 32
Department of Economics
By:Kumadebis Tamiru
 18 2 10
and using the steps above to find adjoint of a matrix we obtain

adj A =

3 0 -1 1 -1 1
2 10 2 10 3 0

2 0 1 1 1 1
- 18 10 18 10 2 0

2 3 1 -1 1 -1
18 2 18 2 2 3

Adj A= 30 22 -3
-20 -8 2
-50 -20 5

Therefore, adj A = 1 -1 1 30 22 -3 multiplying

2 3 0 x -20 -8 2 the conformable
18 2 10 -50 -20 5 matrices

0 0 0
= 0 0 0
0 0 0
where 0 is the null matrix of order 3.
Linear Algebra for Economists Lecture note, Wollega University, 33
Department of Economics
By:Kumadebis Tamiru
1.8. Inverse of A Square Matrix
If A is a square matrix of order n, then a square matrix B of the same order n is said to be
inverse of A if AB = BA = I (unit matrix of order n).

Consider the matrices

2 0 -1 3 -1 1
A= 5 -1 0 , B = -15 6 -5
0 1 3 5 -2 2

Multiplying A and B
It can be easily see that
AB = BA = I (unit matrix)
In this case, we can say that B isinverse of A and A is inverse of B.
Note that:
i. Inverse of a square matrix is defined only for square matrices. But this does
not mean that every square matrix is invertible.
ii. If B is an inverse of A, then A is also an inverse of B.
iii. If a matrix A has an inverse, then A is said to be invertible
iv. Inverse of a matrix is unique
Uniqueness of inverse of a matrix can be easily shown as follows
For, let B and C be two inverses of A.
then, AB = BA = I and AC = CA = I
So B = BI = B(AC) = (BA)C = IC = C
Hence, B = C
This shows that the matrix A cannot have two distinct inverse.
Notation: Inverse of A is denoted by A-1
In this section, we give a method to determine the inverse of a matrix. .
Let us see now the following two ways of finding the inverse of a square matrix,
namely,adjoint method and Gaussian elimination procedure.
a) The adjoint method of finding an inverse
Linear Algebra for Economists Lecture note, Wollega University, 34
Department of Economics
By:Kumadebis Tamiru
Theorem: If A be a square matrix of order n, then A (adjA) = (adj A) A = ǀAǀI, where I is
the Identity matrix of order n.
By this theorem above, we have
A (adjA) = (adjA) A = ǀAǀI

A(adjA) (adjA) A A I
⇒ = ⇒
A A A
(Dividing both sides by ǀAǀ≠0.)
A(adjA) (adjA) -1
= A = AA =I
 Since A-1A= AA-1= I, A A

adj A 1
= (adjA)
 This relation shows that A-1 = A A
dividing again both
sides by A
1
(adjA), where A ≠ 0
A
Hence, A-1 = inverse of A=
We can easily obtain the inverse of a square matrix A using adjoint method following
the steps below:
step 1: finding the determinant of the given matrix A
step 2: finding the adjoint of the given matrix
step 3: dividing the adjoint matrix by the determinant obtained in step1.
adjA CA t
=
A A
Thus, A-1 =
Note that if ǀAǀ= 0, then A-1 does not exist.

Example

Find the inverse of each of the following matrices

i. 2 -1 a b 2 -3 5
-1 2 , ii. c d iii. 5 2 -7
-4 3 1
Solution:
We give below the solution to i and ii, and try iii by your own.

Linear Algebra for Economists Lecture note, Wollega University, 35

Department of Economics
By:Kumadebis Tamiru
i. A = 2 -1
-1 2
Step1:ǀAǀ= 4-(1) = 3
Step2: C11 = 2, C12 = -(-1) = 1, C21 = -(-1) = 1, C22 = 2

Remember how to find cofactors of each element,

2 1
CA = -1 2
2 1
CAt= 1 2 = adj. A

  2 1
1 1  2 1 3 3
.adj A   
A 3   1 2
-1
Step3: A =  1 2  = 3 3

2 1
3 3
1 2
Thus, A-1 = 3 3

ii. B = a b
c d
Step 1:ǀBǀ= ad – bc
Step 2: C11 = d, C12 = -C, C21 = -b, C22 = a
CB = d -c
-b a

CBt = d -b
-c a = adj B

d b
d b ad bc ad bc
1
1 x c a
adj B ad bc
B c a = ad bc ad bc
Step 3: B-1 = =

Linear Algebra for Economists Lecture note, Wollega University, 36

Department of Economics
By:Kumadebis Tamiru
 b) The Gaussian Elimination Procedure
The following elementary operations are needed to use Gaussian method of finding an
inverse of a matrix
i. Interchanging two rows
ii. Multiplying a row by a non-zero scalar
iii. Adding a scalar multiple of one row to another.

Briefly, the Gaussian method starts by writing the given matrix at the left and the
corresponding identify matrix at the left and the corresponding identify matrix next to it
at the right). Then select and carry out row operations that will convert the given matrix
into the identify matrix, and apply the same operations to the matrix at the right. When
the left matrix becomes the identify matrix, the matrix on the right will be the desired
inverse.

Zeros first: For simplicity, the Gauss- Jordan method proceeds to obtain ones on the
main diagonal and zeros on the off- diagonal elements.

 obtain the off –diagonal zeros first and following this, obtain ones on the main
diagonal.
To illustrate, let us find the inverse of the matrix
A= 3 2
1 1
A = AI = 3 2 1 0
1 1 0 1
= 1 1 0 1 interchanging row 1 and row 2
3 2 1 0
= 1 1 0 1  R2 = -3 R1+R2
0 -1 1 -3
R2
= 1 1 0 1  R2 =  1
0 1 -1 3
Linear Algebra for Economists Lecture note, Wollega University, 37
Department of Economics
By:Kumadebis Tamiru
 = 1 0 0 1  R1 = R1 –R2
0 1 -1 3

We now have the identify matrix at the left, so the inverse of

A= 3 2
1 1 is the matrix at the right, namely;

A-1 = 1 -2
-1 3

As a check, we have
3 2 1 -2 1 0
1 1 x -1 3 = 0 1

Example

Find the inverse of the following matrices using Gaussian elimination method.
i. A 7 3 ii. 0 1
2 1 , 2 3
Solution:
Applying the procedure of the above example, we can obtain the inverse of the matrices
A and B as follows.

i. A = AI= 7 3 1 0
2 1 0 1

2 1 0 1
= 7 3 1 0 - Interchanging rows.

Linear Algebra for Economists Lecture note, Wollega University, 38

Department of Economics
By:Kumadebis Tamiru
 = 2 ½ 0 ½ Dividing row 1 by 2 R1=R1 to obtain 1
7 3 1 0 2 in the main
diagonal

= 1 ½ 0 ½ R2 = R2 – 7 R1 converting a21 (7) to zero.

0 -½ 1 0

1 0 1 -3
7
= 0 -½ 1 /2 R1 = R1+R2

= 1 0 1 -3
0 1 -2 7 R2 = -2 R2
Since the left matrix is converted to identify, the inverse of A

-1
A = 1 -3
-2 3 at the right.
ii. B= 0 1
2 3

B= BI = 0 1 1 0 , we have un wanted zero as the upper left

2 3 0 1 diagonal element.

Inter changing rows we have,

= 2 3 0 1 Interchanging rows
0 1 1 0

= 2 0 -3 1 R1 = -3R2 +R1
0 1 1 0
Linear Algebra for Economists Lecture note, Wollega University, 39
Department of Economics
By:Kumadebis Tamiru
 3
= 1 0 /2 ½
0 1 1 0

1
Hence, A = -3/2 ½
1 0

Exercise 4

1. Find the adjoint of the following matrices

i. 1 -2 ii. 4 6 iii. 1 -1 1 iv. 4 -2 1

-3 4 -2 -3 , 2 3 0 1 10 -7
16 2 10 2 -4 1
2. Find the inverse of the following matrices, if it exists,

i. 3 8 ii. 3 1 iii. 0 1 2 iv. 3 2 7

2 1 , 4 0 1 2 3 4 -3 -2
3 1 1 5 9 23

Linear Algebra for Economists Lecture note, Wollega University, 40

Department of Economics
By:Kumadebis Tamiru
 CHAPTER TWO
2.1 SYSTEMS OF LINEAR EQUATIONS
2.1 Introduction
Linear algebra is a study of systems of linear equations and attempts to find a
simultaneous solution for the announce of those equations if such a solution exists.Matrix
algebra permits the concise expression of a system of linear equations. Systems of linear
equations arise in two ways in economic theory. Some economics models have a natural
linear structure. On the other hand, when the relationships among the variables under
consideration are described by a system of nonlinear equations, one takes the derivative
of these equations to convert them to an approximating linear system.

Advantages of Matrix Algebra,

 It permits expressions of complicated systems of equations in a compact and

simplified way,
 It provides a means to check whether a solution exists before an attempt,
 It also permits a means of solving the equation system.
Linear Equations: Any straight line in the XY- plane can be represented algebraically
by an equation of the form
a1x + a2y = b,
Where a1, a2, and b are real constants,
a1 and a2 are not both zero.

Linear Algebra for Economists Lecture note, Wollega University, 41

Department of Economics
By:Kumadebis Tamiru
An equation of this form is called a linear equation in the variables x and y. More
generally, we define a linear equationin the n variables x1, x2, …,xn to be one that can be
expressed in the form
a1x1 + a2x2 +…+ anxn = b,
Where a1, a2, …,an, and b are real constants.

The variables in a linear equation are sometimes called unknowns.

Examples of linear equations:

The equations 3x + y = 7, y = 1/2x +3z + 1 and x 1 + 2x2 – 3x3 + x4 = 90 are linear

equations.

Observe that a linear equation does not involve any products or roots of variables. All
variables occur only to the first power and do not appear as arguments for trigonometric,
logarithmic, or exponential functions.

The equations x + 3y1/2 = 10, x + 3y – z + xz = 14 and y = sinx are not linear.

A solutionof a linear equationa1x1 + a2x2 +…+ anxn = b is a sequence of n numbers s1, s2,
…,sn such that the equation issatisfied when we substitute x 1 = s1, x2 = s2, …, xn = sn. The
set of all solutions of the equation is called its solution set or sometimes the general
solution of the equation.

Example

Find the solution set of 4x – 2y = 1

Solution:

Linear Algebra for Economists Lecture note, Wollega University, 42

Department of Economics
By:Kumadebis Tamiru
To find solution, we can assign an arbitrary value to x and solve for y, or choose an
arbitrary value for y and solve for x. Ifwe follow the first approach and assign x an
arbitrary value t, we obtain x = t, y = 2t – ½

These formulas describe the solution set in terms of an arbitrary number t, called a
parameter. Particular numerical solutions can beobtained by substituting specific values
for t. For example, t = 3, yields the solution y = 11/2 and t = -½ yields the solution y = -3/2.
If we follow the second approach and assign y the arbitrary value t, we obtain x = ½t + ¼,
y=t
Although these formulas are different from those obtained above, they yield the same
solution set as t varies over all possible realnumbers.

Exercise

Find the solution for: x1 – 4x2 + 7x3 =5.

Linear Systems:
A finite set of linear equations in the variables x1, x2, …,xnis called a system of linear
equations or a linear system. A sequence of numbers s1, s2, …,sn is called a solution of the
system if x1 = s1, x2 = s2, …, xn = sn is a solution of every equation in the system. For
example, the system
4x1 – x2 + 3x3 = -1
3x 1 + x2 + 9x3 = -4 has the solution x1 = 1, x2 = 2 and x3 = -1 since
these values satisfy both equations.

However, x1 = 1, x2 = 8, x3 = 1 is not a solution since these values satisfy only the first
equation in the system.Not all systems of linear equations have solutions. For example, if
we multiply the second equation of the system
x+y=4
2x + 2y = 6 by ½, it becomes evident that there are no solutions since
the resulting equivalent system x+y=4
x+y=3 has contradictory equations.

Linear Algebra for Economists Lecture note, Wollega University, 43

Department of Economics
By:Kumadebis Tamiru
A system of equations that has no solutions is said to be inconsistent; if there is at least
one solution of the system, it is called consistent. To illustrate the possibilities that can
occur in solving systems of linear equations, consider a general system of two linear
equations in the unknowns x andy:

a1x + b1y = c1 (a1, b1 not both zero)

a2x + b2y = c2 (a2, b2 not both zero)

The graphs of these equations are lines; call them l1and l2. Since a point (x, y) lies on a
line if and only if the numbers x and y satisfy the equation of the line, the solutions of the
system of equations correspond to points of intersection of l1and l2. There arethree
possibilities, illustrated in Figure 2.1:

l2 y Il
Y 1 2

l1
x x

1) No solution b) has one solution

l1 and l2
y

c. has infinitely many solutions

 The lines l1and l2 may be parallel, in which case there is no intersection and
consequently no solution to the system.
Linear Algebra for Economists Lecture note, Wollega University, 44
Department of Economics
By:Kumadebis Tamiru
  The lines l1and l2 may intersect at only one point, in which case the system has
exactly one solution.
 The lines l1and l2 may coincide, in which case there are infinitely many points of
intersection and consequently infinitelymany solutions to the system.
Every system of linear equations either has no solutions, or has exactly one solution, or
has infinitely many solutions.

An arbitrary system of m linear equations in n unknowns can be written as

Where x1, x2,…,xn are the unknowns and the subscripted a's and b's denote constants. For
example, a general system of three linear equations in four unknowns can be written as

2.2. Methods of Solving Simultaneous Linear Equations

In this section we shall develop a systematic procedure for solving systems of linear
equations. A system of linear equations can be solved using any of the following three
methods.
i. Inverse method
ii. The Cramer’s rule
iii. Gauss Jordan Elimination Method

2.2.1. Inverse Method

Systems of linear algebraic equations and their solutions constitute one of the major
topics studied in “linear algebra.” The solution of linear system of equations can be given
by the product of the inverse of the coefficient matrix A -1 and the column vector of

Linear Algebra for Economists Lecture note, Wollega University, 45

Department of Economics
By:Kumadebis Tamiru
constants C, i.e., if AX = C and if A is invertible, following the law of conformability we
can write in the following form. AX = C and A -1AX = A-1C, then, X = A-1C, this is
inverse rule.
Considering two linear equations in x and y given by:
a11 x + a12y = b1 ………………..(1)
a21x + a22y = b2

Let A be the coefficient matrix. That is

a11 a12 x b1
A= a21 a22 , X = y and C = b2

Equations (1) above can be written in matrix notation form as follows

AX = C = a11x+a12y =b1
a21x+a22y=b2
The above system of equations is said to be consistent if it has at least one solution. If the
system has no solution, then it is said to be inconsistent (i.e, not consistent). The system
with exactly one solution is said to be consistent and independent.

Example

2x+3y = 8
x+2y =5 has exactly one solution and thus, consistent and independent.

We can see, A = 2 3 and A = 4 – 3 = 1 ≠ 0

1 2

and the system x+3y = 5 has no solution and the two equations are
2x+6y = 7 inconsistent. Determinant of A is exactly zero.

That is, A= 1 3
2 6 andA =0

Furthermore, the system x+3y = 5 has an infinite solutions, in

Linear Algebra for Economists Lecture note, Wollega University, 46

Department of Economics
By:Kumadebis Tamiru
 2x+6y =10 this case, the two equations are consistent
but dependent.
In general,
i. If the determinant of the coefficient matrix is different from zero, the system has
exactly one solution (has a unique solution). Therefore, the system of equations is
consistent and independent.
ii. If the determinant of coefficient matrix is exactly zero, then the system of
equations is inconsistent. Let ≠ 0, then, A-1 exists. From AX=C we get
X = A-1C
IX = A-1 C, then,
X = A-1 C

Thus, to find the solution of a linear system of equations using the inverse method, first
obtain the inverse of the coefficient matrix and then, multiply it by the constant matrix,
which gives the required solution.

Example

The equilibrium condition for three related markets is given by the following systems of
equations.
11P1 – P2 – P3 = 31
-P1 + 6P2 – 2P3 = 26
-P1- 2P2 + 7P3 = 24
Find the equilibrium prices by using the inverse method.

Solution:
11 -1 -1
Let A = Coefficient matrix = -1 6 -2
-1 -2 7
P1 31
P = variable matrix = P2 and C = Constant matrix = 26
P3 24

Linear Algebra for Economists Lecture note, Wollega University, 47

Department of Economics
By:Kumadebis Tamiru
 AP = C, then, P = A-1 C
A = 11(42 – 4) + 1( -7-2) – 1(2 + 6)
= 11(38) – 9 – 8 = 418 – 17 = 401, thus, determinant of the coefficient matrix A is
nonzero implying there is no linear dependence among rows and columns.
P = A-1 C;
A-1 = 1/( A )AdjointA
ǀ
c11 c12 c13
Adj.A = Cǀ = c21 c22 c23 ,
c31 c32 c33

But, c11 = M11 = c11(c22c33 - c23c32)

Likewise for, 11 -1 -1 P1 31
-1 6 -2 P2 = 26
-1 -2 7 P3 24

c11 = M11 = 42 – 4 = 38
c12 = -M12 = -(- 7– 2) = 9
c13 = M13 = 8
c21 = -M21 = -(-9) = 9 and c33 = M33 = 66 – 1 = 65
, 38 9 8 38 9 8
-1
C = 9 76 23 , A = 1/401 9 76 23
8 23 65 8 23 65

Therefore, P1 38 9 8
-1
P= P2 A C= 9 76 23
P3 8 23 65

P1 = 1/401[38(31) + 9(26) + 8(24)] = 1/401[1178 + 234 + 192] = 4

P2 = 1/401[9(31) + 76(26) + 23(24)] = 1/401[279 + 1976 + 552] = 7
P3 = 1/401[8(31) + 23(26) + 65(24)] = 1/401[248 + 598 + 1560] = 6
P1 4
Thus, P= P2 = 7 , these are the equilibrium prices.
P3 6
Linear Algebra for Economists Lecture note, Wollega University, 48
Department of Economics
By:Kumadebis Tamiru
Exercises

i. Given a system of linear equations:

4x1 – 5x3 + x2 = 8
3x2 – x3 – 2x1 = 12
-x2 + 3x1 + 4x3 = 5 solve for unknowns using inverse method.
ii. Given the following system
Q2 + 10K = 0
Q1 + 2K = 0
10Q1 + 2Q2 - 240 = 0
Find the values of Q1, Q2 and K using inverse method
iii. Using inverse method, solve the equations.
x+y+z = 6
x+2y+3z = 16
-x+y-z = -2
iv. A manufacturer produces two types of products A and B. Each product is first
processed in machine M1 and then, sent to another machine M 2 for finishing. Each
unit of A requires 20 minutes time in M1 and 10 minutes in M2 whereas each unit
of B requires 10 minutes time in M1 and 20 minutes in M2, The total time
available on each machine is 600 minutes. Calculate the number of units of two
types of products produced by constructing a matrix equation of the form AX = C
and then, solving it by the matrix inversion method.
2.2.2. The Cramer’s Rule
The rule provides a simplified method of solving a system of linear equations through the
use of determinants. The rule states that:

Let a1 x + b1y = C1
a2 x + b2y = C2
The solution of the system above according to Cramer’s rule is obtained as follows:
Let a1 b1 X= xc1
A= a2 b2 , y and B = c2

Linear Algebra for Economists Lecture note, Wollega University, 49

Department of Economics
By:Kumadebis Tamiru
Dx = c1 b1 Where coefficient of x is replaced by the
c2 b2 constant matrix

Dy = a1 c1 Where coefficients of y is replaced by the constant matrix.

a2 c2

c1 b1
Thus, x= c2 b2 = c1 b2 – b1 c2 Dx
A a1 b2 – b1 a1 A

Thus, x= a1 c1
a2 c2 = a1 c2 – c1 a2 = Dy
A a1 b2 – b1 a1 |A|

Example

Using the Cramer’s rule, solve the following systems of equations

2x -3y = 1
X+5y = 7
Solution:
Let A = 2 -3 x= x &B= 1
1 5 y 7

Dx = 1 -3
7 5 = 5-(-21) =26

Dy = 2 1

Linear Algebra for Economists Lecture note, Wollega University, 50

Department of Economics
By:Kumadebis Tamiru
 1 7 = 14-1 = 13

A = 2 -3
1 5 = 10 – (-3) = 13

Dx 26
= =2
A 13
X=
Dy 13
=1
A 13
Y=
Hence, x = 2 and y = 1

Example

Solve with the help of determinants (Cramer’s rule)

5x- 6y +4z = 15
7x +4y – 3z = 19
2x +y +6z = 46
Solution : Remember how to find a determinant of order three.
Let A = 5 -6 4 x 15
7 4 -3 ,x= y and B = 19
2 1 6 z 46

Dx = 15 -6 4
19 4 -3 = 1257
46 1 6

Dy = 5 15 4
7 19 -3 = 1676
2 46 6

Linear Algebra for Economists Lecture note, Wollega University, 51

Department of Economics
By:Kumadebis Tamiru
 5 -6 15
Dz = 7 4 19 = 2514
2 1 46

5 -6 4
A = 2 4 -3 = 419
2 1 6
Dx 1257
= =3
A 419
Now x=

Dy 1676
= =4
A 419
y=

Dz 2514
= =6
A 419
and z=
Hence, x = 3, y = 4 and z = 6

2.2.3. Gauss Elimination Method

In this section, we will show how to use the Gauss – Jordan method to solve m by n
linear systems of equations where mxn. For ease of understanding, we present the case
where the number of rows (number of equations) m is greater than the number of
columns (number of variables), i.e., m >n.
Consider the following 3 by 2 system.
4x1 +5x2 = 30
3x1 + 2x2 = 19
2x1 + 5x2 = 20
Proceeding in the usual manner, we start with the augmented matrix
4 5 30
3 2 19
2 5 20

Linear Algebra for Economists Lecture note, Wollega University, 52

Department of Economics
By:Kumadebis Tamiru
Convert the main diagonal entries (aij = aji ) to unity and the remaining entries of the
coefficient matrix to zero (0) using row operations. Then, we divide the first row by 4 to

obtain 1 in a11
5 15
1 /4 /2 R 1 = ¼ R1
3 2 19
2 5 20
Next we add – 3 times row 1 to row 2 and -2 times row 1 to row 3.

5 15
1 /4 /2

-7 -7
0 /4 /2 R2 = -3R1+R2
5
0 /2 5 R3 = -2R1+R3

Now we divide row 2 by -7/4

5 15
1 /4 /2 R2 = -7R2 1 0 5
0 1 2 0 1 2
5 -5
0 /2 5 0 /4 5

Then add -5/4 times row 2 to row 1 and -5/2times row 2 to row 3. At this point, the method
terminates since this satisfies the condition. The first, second and third row tells us
x1 +0x2 = 5 then, x1 = 5
x1 +1x2 = 2 then, x2 = 2
0x1 +0x2 = 0 then, 0 = 0
Thus, our solution is given by x1= 5 and x2 =2

Exercises

Inverses and solutions of linear equations

1. Find the adjoint and inverse of the following matrices

A= 3 8 3 1 0 1 2
2 1 , B= 4 0 , C= 1 2 3
Linear Algebra for Economists Lecture note, Wollega University, 53
Department of Economics
By:Kumadebis Tamiru
 3 1 1
2. Solve the following systems using the four methods of finding a solution.
i. 4x – 3y = 5 ii. 3x+ 2y = 6 iii. x – y = 3
3x – 5y = 1 5x + 4y = 8 2x + 3y – 4z = 17
y+2z=7
vi.x + 2y +3z = 6 v. x +y + z = 3
3x – 2y +3z = 2 2x – y +z = 2
4x +2y +z = 7 x – 2y +3z = 2

Applications Questions

1)

INDIVIDUAL ASSIGNMENT
1).A candy manufacturer regularly makes three kinds of candy, each of which
requires milk and butter as follows; candy 1 requires 3 gallons of milk and 1 pound
of butter per dozen, candy 2, requires 2 gallons of milk and 3 pounds of butter per
dozen and candy 3 requires 3 gallons of milk and 2 pounds of butter per dozen. The
purchasing department buys a variable amount of milk and butter each week,
depending upon the market prices; the production department must then determine
each week the amounts of each candy to be made in order to consume the quantities
of milk and butter bought. If in the current week 23 gallons of milk and 26 pounds
of butter are purchased, find the number of dozens of each candy to be made

2). A special food for athletes is to be developed from tow foods. Food X and Food
y the new food is to be designed so that it contains exactly 16 ounces of vitamin A,
exactly 44 ounces of vitamin B and exactly 12 ounces of vitamin C. Each pound of
food x contains 1 ounce of vitamin A, 5 ounces of vitamin B and 1 ounce of vitamin
C- On the other hand, each pound of food y contains 2 ounces of A, 1 ounce of B
and 1 ounce of C Find the number of pounds of each food to be used in the mixture
in order to meet the above requirements.

Linear Algebra for Economists Lecture note, Wollega University, 54

Department of Economics
By:Kumadebis Tamiru
 CHAPTER THREE

3.0 SPECIAL DETERMINANTS AND MATRICES IN ECONOMICS

3.1. A Jacobian Determinant

A Jacobian determinant is special types of matrix used to test functional dependency
between linear & non – linear functions. A Jacobian determinant is composed of all the
first order partial derivatives of a system of equations arranged in ordered sequence. It is
J
denoted by
.

If Jacobian determinant becomes zero then it is to mean that there is linear or non – linear
functional dependency between two or more equations. The reverse is true if the
determinant different from zero.

Example given
y1 = f 1 (x1 , x 2 , x3 )

y 2 = f 2 (x1 , x 2 , x3 )

y3 = f 3 (x1 , x2 , x3 )

Linear Algebra for Economists Lecture note, Wollega University, 55

Department of Economics
By:Kumadebis Tamiru
 ∂y1 ∂y1 ∂y1
∂x1 ∂x 2 ∂x3
∂y ∂y 2 ∂y 2
J = 2
∂x1 ∂x 2 ∂x3
∂y 3 ∂y 3 ∂y 3
∂x1 ∂x 2 ∂x3

- If
J =0, then there is functional dependency among equations and hence no

unique solution exists.

Example

∂y1 ∂y
1 1
y 12 means y11 means 1
y1 =4 x1 x 2 ⇒ y =4 & y = 1
1 2 ∂x 2 and ∂x1

∂y 2
y 2 =16 x12 +8 x1 x 2 +x 22 ⇒ =32 x1 +8 x 2
∂x1

∂y 2
=8 x1 +2 x 2
∂x 2

4 1
J  64 x1  16 x 2 ≠ 0,
32 x1  8 x 2 8 x1 2 x 2

Thus, the system has a solution because the value of the Jacobian determinant is non-
singular, i.e., different from zero.

As stated in the preceding paragraphs, partial derivatives can provide a means of testing
whether there exists functional (linear or nonlinear) dependence among a set of n
functions in n variables. This is related to the notion of Jacobian determinant.

Exercises

Linear Algebra for Economists Lecture note, Wollega University, 56

Department of Economics
By:Kumadebis Tamiru
Use Jacobian determinants to test the existence of functional dependence between the
functions paired below:

(a) 3 (b) 3
9 5

3.2. Hessian Determinant

H
It is special types of matrix which is denoted . The economic importance of hessian

matrix determinant is characterization of the free/unconstrained optimization problem

whether it has relative maximum value or relative minimum value or neither of the two

(inflection value).

A Hessian determinant is composed of all the second order partial and cross partial
derivatives objective function arranged in ordered sequence.

Hessian Determinant Test:

a) n-variable case
z  f ( x1 , x2 , , xn )
dz  f1dx1  f 2 dx2    f n dxn
f11 f12  f1n
f 21 f 22  f 2n
H 
  
f n1 fn2  f nn
Principal minors: H1 , H 2 ,  , H n
F.O.C.: f1  f 2   f n 0
S.O.C.: n principal minors must be positive (minimum)
n principal minors must alternate in sign (maximum)

b) Three-variable case
First order condition:
Linear Algebra for Economists Lecture note, Wollega University, 57
Department of Economics
By:Kumadebis Tamiru
 z  f ( x1 , x2 , x3 )

First partials: f1 , f 2 , f3
2 z
fij 
xi x j
Second partials:
fij  f ji
Young’s theorem:
For max, min, dz 0 (for arbitrary values of dx1, dx2, dx3, not all zero)
dz  f1dx1  f 2 dx2  f 3dx3 0

 f1 0, f 2 0, f 3 0, is the necessary condition for extreme

Second Order Condition:

At stationary value of z, or dz=0,
2
If d z is positive definite, sufficient condition for minimum
2
If d z is negative definite, sufficient condition for maximum

Technique: regard dxi as variables that can take on any values, fij as coefficients
We have a symmetric Hessian Determinant.

f11 f12 f13

H  f 21 f 22 f 23
f31 f 32 f 33

Successive principal minors

f11 f12
H1  f11 , H2  , H3  H
f 21 f 22

Rules:

 maximum  
 H1  0, H 2  0, H 3  0 
  d 2 z is negative definite 
 
z is a   if   or  2 
 minimum   H1  0, H 2  0, H 3  0 
   d z is positive definite 
 

Linear Algebra for Economists Lecture note, Wollega University, 58

Department of Economics
By:Kumadebis Tamiru
Example 1

Optimize the following functions

z = f (x, y )=6 x 2 9 x 3 xy 7 y +5 y 2

Solution:
First order condition:
f x =12 x 9 3 y =0

f y = 3 x 7 +10 y =0

Solving simultaneously yields the critical values x =1 & y=1

Second order condition:

f xx 12 f yy 10 f  3
, and xy
12 3
H 1  12  0, and . H 2  121  0
 3 10
2
Thus, d z >0 (positive definite).i.e. the equation is minimized at the critical values.

Example 2

Find the extreme values of

z 2 x12  x1 x2  4 x2 2  x1 x3  x32  2
f1  4 x1  x2  x3 0
f2  x1  8 x2 0
f3  x1  2 x3 0

Solution:

Linear Algebra for Economists Lecture note, Wollega University, 59

Department of Economics
By:Kumadebis Tamiru
  x1  x1
x2  , x3 
8 2
 x1  x1
4 x1   0
8 2
32 x1  x1  4 x1 0
27 x1 0
x1 0, x2 0, x3 0, z 2

f11 f12 f13 4 1 1

H  f 21 f 22 f 23  1 8 0
f 31 f 32 f 33 1 0 2

H1  f11 4
f11 f12
H2  31
f 21 f 22
H 3  H1 54

H1  0, H 2  0, H 3  0  positive definite, minimum

CHAPTER FOUR
4.0 INPUT –OUTPUT MODEL
4.1. Introduction
Input-output analysis is a technique which was invented by Professor Wassily W.
Leontief in the year 1951. The input-output analysis summarizes the flows of inputs and
outputs in an economy in the form of input-output table. The basic notion of Input-output
analysis depends on the belief that the economy of any country can be divided into a

Linear Algebra for Economists Lecture note, Wollega University, 60

Department of Economics
By:Kumadebis Tamiru
distinct number of sectors (industries, or activities) each consisting of one or more firms
producing a similar but not necessarily homogeneous product.

Each industry requires certain inputs from other sectors in order to produce its own
output. Similarly, each industry (sector) sells some of its output to other industries so that
they too can satisfy their intermediate material needs.

A table summarizing the origin of all the various inputs and destination of all various
outputs of industries in an economy is known as an input-output. The input-output table
provides a convenient framework for measuring and tracing these inter-industry (inter-
sectorial) flows of inputs and outputs among the various sectors of the economy. Thus,
the essence of input-output analysis is that, given certain technological coefficients and
final demand, each endogenous sector would find its output uniquely determined as a
linear combination of multi-sector demand.

Let us assume that an economy consists of 4 producing sectors only, and that the
production of each sector is being used as an input in all the sectors and is used for final
consumption.

4.2 Closed and Open Input-output Models

In an input-output model, we deal with that part of economy in which consumption
equals production, i.e, all that is produced is consumed. Broadly speaking there are two
types of input-output models (i) closed (ii) open
4.2.1 Closed Input-output Model
In a closed input-output model, whatever is produced is supposed to be consumed by the
producers themselves. NO one except those who participate in the production are
benefited.

Example

Linear Algebra for Economists Lecture note, Wollega University, 61

Department of Economics
By:Kumadebis Tamiru
Suppose a society consists of three productions A, B and C and suppose further that

amongthemselves. Let A consume 20 percent of

whatever they produce is consumed
what he produces and B, 60 percent of A’s production and the remaining is consumed by
C (20 percent). Again, let production of B be shared by A,B and C in the ratio, 3:4:3 or as

3 4 3
A' S = , B ' S = and C ' S = .
10 10 10

Finally let A consumes 10 percent of C’S production, B consumes 20percent of C’s and
C consumes the remaining 70 percent. This is an example of a closed model and can be
summarized as follows:

Production

A B C

A 0.2 0.3 0.1

Consumption B 0.6 0.4 0.2

C 0.2 0.3 0.7

Note that the sum of the entries in any column is always one

4.2.2 Open Input-output Model

In the input output model, a part of the production is consumed by those who produce it
and the rest is utilized by external bodies. In fact It is assumed that whatever produced is
consumed. In this model we are interested to find the amount of production needed so
that an expected demand of future is fulfilled when the present demand is known.

Example

Linear Algebra for Economists Lecture note, Wollega University, 62

Department of Economics
By:Kumadebis Tamiru
Suppose there are two industries A and B, each one producing a single product. It is
further assumed that whatever produced by A is consumed by A and B and the rest by
outside agencies (consumers). Similarly, production of B is utilized. The following table
gives the information as to how much is produced by each and how much is consumed by
whom.
________________________________________________________________________
Current consumer Total output
AB Demand
A 30 40 60 130
B 20 10 40 70

The input-output matrix is then

A B A B
A or A
B B

Notice that the sum of entries in each column is strictly less than 1 as some outside
agency is also using the product.
We can now translate the distribution of total output (product) of four producing sectors
in the following way.

Table: 4.1 Input-output Translation Table

Producing Total output Input requirement of producing Requirements for
sector no of the sector sector final use (output)

Linear Algebra for Economists Lecture note, Wollega University, 63

Department of Economics
By:Kumadebis Tamiru
 (Co1.1) Co1.2 X1 X2 X3 X4 Co1.7
(Co1.3 (Co1.4 (Co1.5 (Co1.5
) ) ) )
1 X1 X11 X12 X23 X14 F1
2 X2 X21 X22 X12 X24 F2
3 X3 X31 X32 X33 X34 F3
4 X4 X41 X42 X43 X44 F4
Primary Total
inputs primary L1 L2 L3 L4 -
(labour) input = L

From the above table we can derive two important equations

1. Columns 3,4,5 and 6 of the above table give us total inputs from all sectors
utilized by each sector for its production. In other words column 4 gives the
production function of sector 2 and column 3 represents the production function
of sector 1.
Thus,
X1 =f (X11, X21, X31, X41, L1)
X2 =f (X12, X22, X32, X42, L2)
X3 =f (X13, X23, X33, X43, L3) X2 = f(X14, X24, X34 , X44, L4)

Xn = f(X1n, X2n, X3n, X4n, Ln)

2. Rows of the table give us the equality between the demand and supply of each
product:

X1 = X11, + X21, + X31, + X41, + F1

X2 = X21, + X22, + X23, + X24, + F2
X3 = X31,+ X32, + X33, + X34, + F3
X4 = X41, + X42, + X43, + X44, + F4
Linear Algebra for Economists Lecture note, Wollega University, 64
Department of Economics
By:Kumadebis Tamiru
 L = L 1 + L2 + L3 + L4
In general Terms, If there are n producing sectors:

X1 = X12, + X12+ X13, + X14,+ - - - +x1n +F1

X2 = X21,+ X22+ X23, + X24,+ - - - +x2n +F2

Xn = Xn1,+ Xn2+ Xn3, + Xn4,+ - - - +xnn+Fn

L = L 1 + L2 + L3 + L4 + - - - L n
n n
Xi = ∑x ij +Fi, and L =∑ Li
That is, j=1 i=1

Here, Xi = Total output of the ith sector

Xij = Output of ith sector used as input in jth sector, and
Fi = Final demand for ith sector

The above identity states that all the output of a particular sector could be utilized either
as an input in one of the producing sectors of the economy and/or as a final demand.
Basically, therefore, Input-output analysis is nothing more than finding the solution of
these simultaneous equations.

4.3 Assumptions of the I-O Model

The economy can be meaning fully divided into a finite number of sectors (Industries).
And the following assumptions are essential to apply the model.
1. Each industry produces only one homogeneous output No two products are
produced jointly; but if at all there is such a case then it is assumed that they are
produced in fixed proportions.
2. Each producing sector satisfies the properties of linear homogeneous production
function – in other words, production of each sector is subject to constant returns
to scale so that a percentage change in every input will result in an exactly a
percentage change in the output.
3. A far stronger assumption is that each industry uses a fixed input ratio for the
production of its output j in other words, requirements per unit of output in each

Linear Algebra for Economists Lecture note, Wollega University, 65

Department of Economics
By:Kumadebis Tamiru
 sector remain fixed and constant. The level of output in each sector (industry)
uniquely determines the quantity of each input which is purchased.

The Technological coefficient Matrix

From the assumptions of fixed input requirements we see that in order to produce one
unit of jth commodity, the input used of ith commodity must be a fixed amount, which we
denoted by aij thus
X ij
Xj
aij = (1) where xj is total output of the ith commodity
aij istechnical coefficient in the ith row and jth column.
From the above equation, we can obtain Xij = a ijxj where xij is input requirement in the i th
commodity. As such we can now put the input-output transaction table interims of
technical coefficients as follows:

Table: 4.2 input output transaction interims of technical coefficients

Producing Total output Input requirements of Requirements
sector no of the sector producing sectors for final uses
sales X1 X2 X3 X4
1 X1 a11 x1 a12 x2 a13 x3 a14 x4 F1
2 X2 a21 x1 a22 x2 a23 x3 a24 x4 F2
3 X3 a31 x1 a32 x2 a33 x3 a34 x4 F3
4 X4 a41 x1 a42 x2 a43 x3 a44 x4 F4
Primary Total primary
input input
Labour Li  1 x1  2 x2  3 x3  3 x3

Note that all the coefficients are positive (aij0). The above table gives us the total output
of each sector interms of technical coefficients; and if there are “n” producing sectors,

X1 = a11 x1+ a12 x2+ a13 x3+ a14 x4+ - - - a1n xn+F1

Linear Algebra for Economists Lecture note, Wollega University, 66

Department of Economics
By:Kumadebis Tamiru
 X2 = a21 x1+a22 x2+a23 x3+ a24 x4+ - - - a2n xn+F2

Xn = an1 x1+ an2 x2+ an3 x3+ - - - + annXn+ Fn

L =  1 x1+  2 x2+  3 x3+ - - -  n xn
We can put the total outputs above using a summation nation as follows
n

∑ aij xj +Fj
Xi = j==1
Where (I = 1,2,3 - - -, n and j 1,2,3, - - -,n)
n

∑ ixj
And L = j=1

The equations may be put in matrix notations as under.

X1 a11 a12 a13…a1n X1 F1
X2 a21a22 a23…a2n X2F2

=
Xnan1 an2 an3… annXnFn
Hence, X = Ax+F From this, we can derive a relationship in which the total output that
must be produced in each sector using Identity matrix (I), coefficient matrix A ( a ij) and
the final demands (Fi), i.e
X = Ax+F
X – Ax = F
X (I-A) = F
-1 -1
X ( I - A) ( I - A) =( I - A) F
- 1
X = ( I - A) F
-1
The expression ( I - A) is known as the inverted Leontief matrix or the Leontief inverse
and it indicates that the total output of each sector is obtained by multiplying the inverted
Leontief matrix by the total final demand of the sectors. Where A is the given matrix of
input coefficients, while X and F are the two vectors of output and final demand of each
-1
producing sector. If ǀI-Aǀ=0, then ( I - A) exists, we can then estimate for either of the
two matrices x and F by assuming one of them to be given exogenously.
Linear Algebra for Economists Lecture note, Wollega University, 67
Department of Economics
By:Kumadebis Tamiru
It is here that we observe that assumptions made in input-output analysis go a long way in
making the problem simplified. For example, with the assumption of linear homogeneous
function, it is possible to write a linear equation of each producing sector which then can
be easily transformed into matrix notation. On the other hand, as long as the input
coefficients remain fixed (as assumed) the matrix A will not change or (I-A) will not
-1
change. Therefore, in finding the solution of X = ( I - A) only one matrix inversion
needs to be performed even if we are to consider thousands of different final demand
vectors according to alternative development targets. Hence, such an assumption of fixed
technical coefficients has meant considerable saving in computational efforts. Now lets
us consider a hypothetical Input –output table to understand the model nicely.

Table: 4.3 Input – output table in units

Purchasing sector
Sectors Input to Input to Input Tota Final Gross
Agricultur Manufactur to l deman outpu
e e Service d (F) t xi
Agriculture 15 0 20 35 80 115
Manufacturing 10 0 35 45 40 85
Service 5 10 15 30 45 75
Total 30 10 60 110 165 285
Selling Sector

intermediate
purchase
House hold (h) 40 35 5 80 50 130
Capital (k) 45 50 10 105 60 165
Gross input 115 85 75 285 110 295
Xj
Reading the I-O table:
1. Reading a row (example row1 = Agriculture): Of the total available output of
agricultural products (115) only 15 units are used by agriculture itself, 0 units by

Linear Algebra for Economists Lecture note, Wollega University, 68

Department of Economics
By:Kumadebis Tamiru
 manufacturing, 20 units are used by the service sector, and the remaining 80 units
goes to final consumption.
2. Reading a column (Example column 1): Reading down this column, we see that in
order to produce a total output of 115 units, the agricultural sector has to use 15
units of its own output, 10 units of manufactured outputs, and 5 units of outputs
of the service sector.

Transaction Matrix:
A transaction matrix shows how the total output of one industry is distributed to all other
industries as inputs and to the final demand sector. For instance in the above table, 115
units of Agricultural output is required to distribute 15 units Agricultural, no units to
manufacturing, 20 units to service sector and 80 units to the final demand sector.

Technological (Technical) Coefficient Matrix:

The assumption of the single process production function (fixed input requirements)
enables us to calculate the input requirements of various output levels from the data
Xij
above to contain in one I – 0 table. The matrix A = ( a ij) where ( a ij) = Xj is called a
technological (technical) coefficient matrix and indicates the input requirement for the
production of one unit of output.
For the given input-output table above in the three – sector economy, the technical
coefficients can be obtained as follows.
x11 15 x 0 x 20
a11 = = =0.13, a12 = 12 = =0, a13 = 13 = =0.27
x1 115 x2 85 x3 75

x 21 10 x 0 x 35
a 21   0.04, a 22  22  0, a 23  23  0.47
x 115 x2 85 x3 75

x31 5 x 10 x 15
a31 = = =0.04 a32 = 32 = =0.12 a33 = 33 = =0.20
x1 115 x2 85 x3 75
xij
aij =
xj
Thus using the equation , the matrix of technical coefficients for our hypothetical
example is given as follows.
Linear Algebra for Economists Lecture note, Wollega University, 69
Department of Economics
By:Kumadebis Tamiru
Agriculture Manufacturing Service
Agriculture 0.13 0 0.27
Manufacturing 0.09 0 0.47
Service 0.04 0.12 0.20

Note that: the sum of entries is lest column would always be one in closed model but the
sum in less than one if we assume an open model.

Interpretation:
Any column in the technical coefficient matrix indicates the different materials (inter-
industry) input requirements to produce one unit of output. For instance, column1
indicates different material input requirements needed by the Agricultural sector from all
the sectors to produce one unit of output. More specifically, to produce one unit of
output, the Agricultural Sector needs 0.13 units of inputs from Agricultural sector, 0.09
units of inputs from Manufacturing sector, and 0.04 units of inputs from service sector.
Now we can put the input –output follows among different sectors using a matrix
Algebra.
X =( I - A) - 1 F
-1
x11 0 0 0.13 0 0.27 80
x2 = 0 1 0 - 0.09 0 0.47 x 40
x3 0 0 1 0.04 0.12 0.20 45

Here, you need to recall back how to find an inverse.

By using earlier methods we find the inverse of (I-A)
-1
115 0.87 0 -0.27 80
85 = -0.091 -0.47 40
75 -0.04 -0.12 0.80 45
By using earlier methods to find an inverse and multiplying it with the final demand we
get
115 115
Linear Algebra for Economists Lecture note, Wollega University, 70
Department of Economics
By:Kumadebis Tamiru
 85 = 85
75 75

Use of input – output model for planning

The input-output analysis developed by Leontief is very essential to solve different
problems in planning. Important question which require solution in planning includes the
following.
 What level of output (xi) will be required in each of the producing sectors of the
economy to achieve the demand targets?
 What effect will the demand targets have on the level of primary inputs that will
have to be utilized in the economy in order to carry output successfully the aims
of the overall plan?
 What will be the magnitude and direction of inter industry input flows to satisfy
the resultant material input demands in the economy?

We apply the expression X = AX+F in using I – 0 model for planning.

X = Ax+F
X-Ax = F
(I-A x = F
-1
 X = ( I - A) .F
-1
Where ( I - A) is the inverted Leontief matrix.
A = technical coefficient matrix
I = identify matrix having the same order with A.
(I-A) = difference b/n the Identify matrix and the technical coefficient

Interpretation of the inverted Leontief matrix

Each element of the Leontief inverse indicates the amount of commodity “i” that must be
produced for each unit of final demand for the goods sector “j”

Illustrated Examples

Linear Algebra for Economists Lecture note, Wollega University, 71

Department of Economics
By:Kumadebis Tamiru
 1. Suppose there are only three industries in an economy and estimate the output of
each sector (industry) with the given input coefficient matrix and final demand as
follows: (the coefficient matrix is in value terms)

A= 0.3 0.5 0.2 100

0.2 0 0.5 and F = 40 in millions (birr)
0.1 0.3 0.1 50

Solution :
Since A= 0.3 0.5 0.2 1 0 0 0.3 0.5 0.2
0.2 0 0.5 , I–A= 0 1 0 - 0.2 0 0.5
0.1 0.3 0.1 0 0 1 0.1 0.3 0.1

0.7 -0.4 -0.2

I–A= -0.2 1 -0.5
-0.1 -0.3 0.9

1 0.5 0.2 0.5 0.2 1

0.3 0.9 0.1 0.9 0.1 0.3
ǀI-A ǀ= 0.7 + (0.4) + ( - 0.2)

ǀI-A ǀ= 0.401
Again (I-A)-1 = 1 adj (I-A) = 1 0.75 0.42 0.40
ǀI-A ǀ 0.401 0.23 0.61 0.39
0.16 0.25 0.62
-1
Substituting these values in X = ( I - A) F , weget
X= X1

X2 =
(I - A) .
-1
F

X3

Linear Algebra for Economists Lecture note, Wollega University, 72

Department of Economics
By:Kumadebis Tamiru
 X1 0.75 0.42 0.40 100
X2 = 1 0.23 0.61 0.39 x 40
X3 0.401 0.16 0.25 0.62 50

1
(0.75 x100 +0.42 +0.401(50))
x1 = 0.401
x1 = birr 270 million
1
(0.23x100 +0.6(40) +0.39(50))
x2 = 0.401
x2 = birr 167 million and
1
(0.16 x100 +0.25 x 40 +0.62(50))
x3= 0.401
x3 = birr 142 million

Thus, the three industries should produce outputs amounting birr 270 million, 167 million
and 142 million respectively to satisfy the given final demands.

Note that: the three columns sums of A are (0.3+0.2+0.1) = 0.6, (0.5+0) = 0.8 and
(0.2+0.5+0.1) = 0.8 which are less than 1 in each case. In other words (1-0.6) = 0.4, (1-
0.8) = 0.2 and (1-0.8) = 0.2 is the maximum amount of primary input which can be used
for producing a birr worth of the three commodities (P,Q,R) respectively

Hence, the amount of primary input would be needed to get the above amount of output
mix (279 mill, 167 mill & (142 mill) in three producing sectors.

Therefore, the total primary input requirements of the three sectors will be equal to:0.4
(270 mill) + 0.2 (167 mill) +0.2 (142 mill) = birr 190.1 mill.

Linear Algebra for Economists Lecture note, Wollega University, 73

Department of Economics
By:Kumadebis Tamiru
  0.8 0.2 

 0.7  , does the case have a solution?
2. Suppose A =  0.9

Solution

 1. 0   0.8 0.2 
    0 .2 0 .2 
 0.9  
(I- A = )
0 1 
 -  0.7  =  0 .9 0 .3 

I-A
= 0.2 (0.3) - (0.2 x0.9) =0.06 - 0.18 = - 0.12

Since the determinant of (I - A) is less than zero, the system will not have a possible
solution.

3. Given the technical coefficient

 0.1 0.3 0.1
 
 
 0 0.2 0.2 
 
 0 0 0.3 
A=  
And final demands are F1,F2 and F3 Find the output levels consistent with I-0 model what
will be the output levels if F1 = 20, and F3 = 100?

Solution:
We know that x1 F1
-1
x2 = (I- A) F2
x3F3
1
adj ( I - A)
and (I- A)
-1
= ( I - A)

(I – A) = 0.9 -0.3 -0.1

0 0.8 -0.2
0 0 0.7

Linear Algebra for Economists Lecture note, Wollega University, 74

Department of Economics
By:Kumadebis Tamiru
Cofactors of (I- A) are as follows.

0. 8 0 .2 0 .3 0.1 0 .3 0. 1
C11 = =0.56 C21 = =0.21 C31 = =0.14
0 0 .7 0 0.7 0.8 0 .2

0. 0. 2 0. 9 0. 1 0.9 0.1
C12 = =0 C22 = =0.63 C32 = =0.18
0 0 .7 0 0 .7 0 0 .2

0. 0.8 0 .9 0.3 0.9 0. 3

C13 = =0 C22 = =0 C33 = =0.72
0 0 0 0 0 0 .8

C (I –A) = 0.56 0 0
0.21 0.63 0
0.14 0.18 0.72

t
C (I – A) = adj (I-A) = 0.56 0.21 0.14
0 0.63 0.18 and
0 0 0.78

0.8 0.2 0 0. 2 0 0.8

+0.3 +( 0.1)
E( I A)
t
=dj ( I A) = 0 0 .7 0 0 .7 0 0

-1
I-A
= 1.11 0.42 0.28
0 1.25 0.36
0 0 1.43

x1 = 1.11 0.42 0.28 20

x2 0 1.25 0.36 * 0
Linear Algebra for Economists Lecture note, Wollega University, 75
Department of Economics
By:Kumadebis Tamiru
x3 0 0 1.43 100

x150.2
x2 = 36
x3143
Thus, the total outputs in the three sectors are x1 = 50.2, x2 = 36 and x3 = 143 units.

In the above example, if final demands change by 10 units each, then what will be the
change in sector outputs? (this is a kind self test question).

4. An economy produces only coal and steel. The two commodities serve as
intermediate inputs in each other’s production. 0.4 tons of steel and 0.7 tons of
coal are needed to produce a ton of steel. Similarly 0.1 tons of steel and 0.6 tons
of coal are required to produce a ton of coal. No capital inputs are needed. Does
the system viable?Besides, 2 and 5 labor days are required to produce a tone of
coal and steel respectively. If the economy needs 100 tons of coal and 50 tons of
steel, calculate the gross output of the two commodities and the total labor
required.

Solution:
The technology matrix is given as under
Steel Coal Final demand
Steel 0.4 0.1 50
Coal 0.7 0.6 100
Labor 5 2

I–A= 0.6 -0.1 1 0 0.4 0.1

-0.7 0.4 = 0 1 - 0.7 0.6

The system will be viable if the following conditions are satisfied.

i) The diagonal elements of (I-A) are all positive
Linear Algebra for Economists Lecture note, Wollega University, 76
Department of Economics
By:Kumadebis Tamiru
 ii) The determinant should be positive

In present problem it is:

0.6 and 0.4 > 0 = first condition
ǀI – A ǀ= 0.17 >0 = second condition
Hence, this indicates that the system is viable.

The required total output is given by:

-1
X = ( I – A) .F
-1
x1= 0.6 -0.1 50
x2 -0.7 0.4 100

By with the final demand, we get using earlier methods to find the inverse of (I-A) and
multiplying
x1 = 176 .5
x2 558 .8
Hence, total output of steel is 176.5 times and total output of coal is 558.8 units.
Total labor days required is: 5x steel output +2times coal output
= 5x176.5 +2x558.8 = 2000 labor days

Some limitations of input –output analysis:

i) Errors in forecasting final demand will have grave consequences.
ii) The assumption of linear homogenous production function may not be valid. The
technical coefficients will not remain constant even if input price ratios are held
constant in such circumstances.
iii) The assumption of constant technical coefficients goes counter to the possibility
of substitution of inputs and factors.
iv) Sectorial division is limited for practical purposes sectorisation is not good
enough for many forecasting purposes. Input – output model building is highly
costly interims of time and money,

Linear Algebra for Economists Lecture note, Wollega University, 77

Department of Economics
By:Kumadebis Tamiru
 v) Sectorisation (grouping of commodities in sectors) is often arbitrary. The intra –
sectoral heterogeneity with respect to technologies, efficiency and is not invariant
over time.

Exercises

1. The following table gives the input-output coefficients for a two sector economy
consisting of Agriculture and manufacturing industry.
input A B
A 0.10 0.50
B 0.20 0.25

The final demand for the two industries is 300 and 100 units respectively. Find
the gross (total) outputs of the two industries. If the input coefficients for labor for
the two industries are respectively 0.5 and 0.6 find the total units of labor
required.

1 1 1
2. Let A= 8 3 4

1 1 1
2 6 4

1 1 1
4 6 4

And the final demand vector be 10

F = 28
14
Find the total output of the 3 sector. What will be the total outputs if the final
demand for sector 1 has increased by 4 units

3. Find the demand vector F which is consistent with the output vector
X= 10
28 , when the input – output coefficient matrix is

Linear Algebra for Economists Lecture note, Wollega University, 78

Department of Economics
By:Kumadebis Tamiru
 14
0.2 0.3 0.2
A= 0.4 0.1 0.2
1.4 0.3 0.3

4. From the following matrix find the final output goals of each industry assuming
that the consumer output targets are $ 85 million in steel, 25 million in coal and
55 million in railway transport.
Steel Coal Railway transport
Steel 0.4 0.1 0.2
Coal 0.2 0.2 0.4
Railway transport 0.2 0.3 0.3
Labor 0.2 0.4 0.1
What would be the labor requirements of the final outputs of the three industries?
5. Given the following transaction matrix, find the total output to meet the final
demand of 200 units of Agriculture and 800 units of industry.
Producing sector purchasing sector current
Agriculture Industry Demand
Agriculture 300 600 100
Industry 400 1200 400

6. Suppose the inter-relationships between the production of three industries A, B

and C in a year is given by

A B C Current demand Total output

A 50 20 40 70 180
B 20 30 20 90 160
C 30 20 20 50 120
Find the total output of each industry required to meet future demand of
60
F= 110
Linear Algebra for Economists Lecture note, Wollega University, 79
Department of Economics
By:Kumadebis Tamiru
 CHAPTER FIVE
5.0. LINEAR PROGRAMMING PROBLEMS
Linear programming is a mathematical device that developed by the mathematician
George Dantzig in 1947 for planning the diversified activities of US Air Force with the
problem of suppliers and labor were limited by war time conditions. It is a mathematical
technique for the analysis of optimum decisions, subjected to certain constraints in the
form of linear inequalities. It applies to those problems subjected to a system of linear
inequalities stated in terms of certain variables.

In economics not all relationships between variables are represented by equality and that
is why we are now consider the case when the constraints are given by inequalities.
Provided that the function to be optimized is linear and inequalities are all linear, the
problem is said to be linear programming problem.

We have four subsections under this chapter that we are going to consider.Subsection 5.2
describes the graphical method to solve linear programming problems using graphical
approach. For simplicity we concentrate on problems involving just two unknowns to
describe the graphical methods of solution.

Subsection 5.3, describes the simplex method to solve the linear programming problems.
It is an iterative algorithm which is a systematic repetitive solution procedure to obtain
the best possible solution under the given constraints. Under this section we shall discuss
the whole procedure step by step for solving linear programming problems.
At the end of this section students should be able to:
 Define linear programming problem
 Identify the region defined by a linear inequality and sketch the feasible region
defined by simultaneous linear inequalities
 Solve linear programming problems graphically
 Use the simplex method to find the solution for linear programming problems
 Form dual of linear programming model

Linear Algebra for Economists Lecture note, Wollega University, 80

Department of Economics
By:Kumadebis Tamiru
5.1 Simple Examples of Linear Programming Problems
Under this subsection, as an introduction, we describe the linear programming problems
using some simple example. The discussion of the linear programming model using
simple example help us state the general form of linear programming problem at the end
of this subsection.

Example 5.1

Using Production Problem

Lets’ consider the problem that a given company faces in production process. Suppose
that a given food company produce two products () using two ingredients (). To produce
one unitof product it requires units of ingredient. The maximum amounts of ingredients,
available for food production for company are respectively. The company makes a profit
on each unit of product sold to the market. Hence, the problem of the company under
this state of conditions is to maximize its profit without running out of ingredients.

Further, suppose that the company decides to produce units of product. Then the profits
that obtained from marketing of the products will be given by

On the other hand, the production process will use the following amounts of ingredients,
of ingredient , and
of ingredient

Therefore and most satisfy the following constraints on ingredients use

and .
Also are assumed non negative.

We therefore have to solve the following linear programing problem:

Linear Algebra for Economists Lecture note, Wollega University, 81

Department of Economics
By:Kumadebis Tamiru
Example 5.2

Using Transportation Problem

Now lets’ consider for more than two variables cases using the transportation problem.
Suppose that a company has m factories () and n warehouses. Factory can at most units
of a certain product per week and warehouse must be able to supply at least units per
week. The cost of shipping one unit from factory to warehouse is. Hence, the question
here is that how many units should be transported from each factory to each warehouse
per week in order to minimize the total transportation cost and yet still satisfy the
requirements on the factories and warehouses. Further, lets’ suppose that be the number
of units to be transported from factory to warehouse per week. Then the total
transportation cost per week is

The condition on factory is while it is. We are, therefore, faced with the following linear
programing problem:

Thus from the above examples it is possible to understand the maximization and
minimization problem cases of linear programming for the case of bivariate and
multivariate cases. Under the first example we have two unknown variables to be solved.
But the second case is more general it considers more than two unknown cases.

After these examples we are ready to describe the general form of linear programming
problems. As it is possible to understand from the above example the general form of
linear programming must have at least two parts. The first part is a function that to be
maximized or minimized and the second part includes subject functions.

The function to be maximized or minimized is termed as the objective function and the
subject functions are termed as constraintsof linear programming problem. Given be
Linear Algebra for Economists Lecture note, Wollega University, 82
Department of Economics
By:Kumadebis Tamiru
variables the objective function of linear programming problem can be stated as a linear
function of variables which has be maximized or minimized.

These variables are subjected to a number of linear conditions, constraints, which take the
form

In addition, certain variables take non-negativity assumption. In general linear

programming problem, therefore, takes the following forms:

Thus, in the above general formulation of the linear programming problem the
constants () are all known quantities. The objective here is to find which optimize the
objective function z, while satisfying the constraints.

If X satisfies all constraints including the non-negativity assumption, then it is called a

feasible solution of a particular linear programming problem. These geometrical
represented as a feasible region in Euclidean space. A feasible solution for which the
objet function is maximum or minimum is said to be an optimal solution. The solution
for the linear programing problem can be obtained using either the geographical or
simplex approaches. These approaches are the concern of the next subsections.

Activity 5.1

Provide at least two examples of problems in your area that can be describe and solved
by using the general formulation of linear programing model.

5.2 Graphical Method of Linear Programming Problem

Valuable insight into the nature of the linear programmingproblem is gained by using a
geometrical approach for problem solving on n-dimensional space. For simplicity we
concentrate on problems involving just two unknowns to describe the graphical methods
of solution.

Linear Algebra for Economists Lecture note, Wollega University, 83

Department of Economics
By:Kumadebis Tamiru
Linear InequalitiesSketching:
Before considering linear programming problem it is essential to know how to sketch
linear inequalities on. A linear equation of form x=y can be represented by a straight line.
We can give a similar graphical interpretation for linear inequalities involving two
variables when the equal sign is replaced by one of the following forms of inequalities.

To illustrate this consider the simple inequality. Here, we would like to identify those
points with coordinates for which this inequality holds true. Clearly this has same thing
to do with the straight line initially. Thus we have to start with sketching of straight line
Figure 5.1 Sketching Linear Inequalities

●P
●R

●Q

If a point P lies above the line, then y coordinate is greater than x coordinates, so that.
Similarly, if a point Q lies below the line then the y coordinate is less than x coordinate
so that. The coordinate of point R lies on the lineand satisfying
Linear Algebra for Economists Lecture note, Wollega University, 84
Department of Economics
By:Kumadebis Tamiru
Hence, we can see that the inequality holds for any point that lies on or above the line. It
is useful to be able to indicate this region pictorially. We do this by shading one half of
the coordinate plane.

Figure 5.2 Sketching Linear Inequalities

●P

In general, to sketch any inequalities of form, , we first need to sketch the corresponding
line and then deciding which side of the line to deal with. An easy way of doing this is to
pick a ‘test point’, . It does not matter what point is chosen, provided that it does not
actually lie on the line itself. The numbers of x and y are then substituted into the original
inequality. If the inequality satisfied then the side containing the test point is the region of
interest. If not, then we go for the region on the other side of the line.

Example 5.3

Sketch the region for inequality:

In order to sketch this inequality first we have to sketch the line. This requires finding
and intercepts. When we get y intercept, 4. Similarly, when we get which is . Thus, the
line pass through (0,4) and (2, 0) as shown in Figure 5.3.

Figure 5.3 Sketching Linear Inequalities

Linear Algebra for Economists Lecture note, Wollega University, 85

Department of Economics
By:Kumadebis Tamiru
(0, 4)

● (3, 2)
X
(2, 0)

For a test point let us take (3, 2) which lies above the line. Substituting and in to equation
for gives 2(3) +2=8. This is not less than 4, so the test point does not satisfy the
inequality. It follow that the region of interest lies below the line. This is illustrated in
Figure 5.4. In this example the symbol < is used rather than ≤. Hence, points on the line
itself are not included in the region of interest. This technical can be represented using a
broken line for the boundary.

Figure 5.4 Sketching Linear Inequalities

(0, 4)

● (3, 2)
X
(2, 0)

Activity 5.2

Linear Algebra for Economists Lecture note, Wollega University, 86

Department of Economics
By:Kumadebis Tamiru
Sketch the linear inequality- x+3y> 6 on the graph paper and considering the test point
(1,4) indicate the region over which –x+3y>6 is satisfied.

Feasible Region Sketching:

We consider the region defined by simultaneous linear inequalities. This is known as a

feasible region. It consists of those points (x, y) which satisfy several inequalities at the
same time. We can find it by sketching the regions defined by each inequality. The
feasible region is then the un-shaded part of the plane corresponding to the intersection of
all individual regions of inequalities.

Example 5.4

Sketch the feasible region for the following set of inequality

In this problem the easiest inequalities to handle are the last two non-negativity
inequalities so the feasible region for these two inequalities is in the top right-quadrant of
the plane, as shown in Figure 5.5 below.

For the inequality we need to sketch the line. For this purpose we need to find and
intercepts. When we get and when we get. Thus, the required line pass through
coordinates and.

Figure 5.5 Sketching Feasible Region for Non-negativity Assumptions.

Y

Linear Algebra for Economists Lecture note, Wollega University, 87

Department of Economics
By:Kumadebis Tamiru
 X

For testing point let us take, since such a choice minimizes the amount of arithmetic that
we have to do. Substituting the test point in to the inequality gives, which is obviously
true. Now that the origin lies below the line, so we shade the region that lies below it.
This indicated in Figure 5.6.

Figure 5.6 Sketching Linear Inequalities

(0, 6)

●
X
(12, 0)
Y

For the inequality we need to sketch the line using coordinate as Y-intercept and
coordinate as X-intercept of the line. At the test point,, the inequality is satisfied because ,
which is obviously true. Thus, we are interested in the region below the line, since it
contains the test point, the origin. As usual this indicated by shading the area below the
line.

Linear Algebra for Economists Lecture note, Wollega University, 88

Department of Economics
By:Kumadebis Tamiru
The complete picture of feasible region is shown in the Figure 5.7. The coordinates in the
region, , satisfy all four inequalities, which is the most shaded part of the region.

Figure 5.7 Sketching the Feasible Region of Set of Inequalities

Y

(0, 6)

(0, 3)

X
(-3,0) (12, 0)

Activity 5.3

Sketch the feasible region for the following set of inequalities:

The Geometry of Linear Programming Problems:

We know in a better position to describe and solve linear programming problem using
graphical approach. Valuable insight into the nature of the linear programmingproblem is
gained by adopting a geometrical approach on n-dimensionalspace.We shall know
introduce graphical approach to describe and solve the linear programming problem
using solved problem using appropriate example.

Example 5.5

Solve the following linear programming problem using graphical approach

Minimize:
Subjected to constraints:

Linear Algebra for Economists Lecture note, Wollega University, 89

Department of Economics
By:Kumadebis Tamiru
 3

Geometrically, points which satisfy simultaneous linear inequalities define a feasible

region in the coordinate plane. In fact, for this particular problem, the feasible region has
already been sketched in figure 5.7. Now, the problem is to try to identify that point
inside the feasible region which minimizes the value of object function.

One way of finding the required point is using the trial and error approach. In this
approach the objective function evaluated as every point within the feasible region and
the point that produce the smallest value of the objective function is going to be chosen.
That is why it is considered as a trial approach. For example, point (1, 1) lies in the
feasible region and when these values are substituted in the objective function they yield -
1 (i.e., -2(1) +1=-1). Similarly, we might try point (3.4, 2.1) in the feasible region, which
produces -4.7 (-2(3.4) +2.1=-4.7) which is an improvement, since -4.7<-1.

The drawback of this approach is that there are infinitely many points inside the region,
so it is going to take a very long time before we can be certain of the solution. This
approach geometrical described as the superimposing the family of straight lines on the
top of feasible region. The family of straight lines is represented by the objective function
in the linear programming problem. In our particular example the family of straight lines
represented equation with slope of a line of 2.

A selection among the family of straight lines is represented in Figure 5.8 below. The
values of c range from 0 to -24. These lines have been sketched using the information
that they pass through (-c/2, 0) and have a slope of 2. It is possible to note that as c
decreases from 0 to -24 the lines sweep across the feasible region from left to right. Also,
once c goes below -24 the lines no longer intersect this region. The minimum value of c

Linear Algebra for Economists Lecture note, Wollega University, 90

Department of Economics
By:Kumadebis Tamiru
is therefore -24. Moreover, when c=-24, the line intersects the feasible region in exactly
one point, namely (12, 0). This then must be the solution of our problem.

Figure 5.8 Sketching the Family of Straight Lines over the Feasible Region
Y
c=0 c=-8 c=-16 c=-24

(0, 3)

X
2 4 6 8 10 (12, 0)

For the above example it is possible to understand that the optimal solution is attained at
one of the corners of the feasible region. This is not simply a coincidence. It can be
shown that the solution for any linear programming problem always occurs at one of the
corners. Consequently the trial and error approach suggested that the only candidates for
answer are the corners and so only a finite number of points need ever be examined.

Given the above example the general procedures for finding the solution for linear
programming problem may be summarized as:
Step 1: sketch the feasible region
Step 2: identify the corners of each feasible region and find their coordinates.
Step 3: evaluate the objective function at the corners and choose the one
which has the maximum of the minimum value

Returning to our pervious example, we work as follows:

Step 1: the feasible region has already been sketched in Figure 5.7.
Step 2: the feasible region has a corners with coordinates (0, 0), (0, 3), (2, 5), and
(12, 0).
Step 3: evaluate the objective function at the corners and choose the one which
has the minimum value
Linear Algebra for Economists Lecture note, Wollega University, 91
Department of Economics
By:Kumadebis Tamiru
 Corner Objective function value
(0, 0) -2(0)+0=0
(0, 3) -2(0)+3=3
(2, 5) -2(2)+5=1
(12, 0) -2(12)+0=-24

From this we can see that the minimum value of the objective function attained the
coordinate (12, 0). Incidentally, if we also require the maximum value then this can be
deducted without further effort. From the table the maximum is 3, which occurs at corner
(0, 3).

Activity 5.4

Solve the following linear programming problem using the general procedures of
graphical approach
Maximize:
Subjected to constraints:

It is possible for the linear programming problem to have either has no unique solution or
no solution. Next we illustrate such special cases by considering two appropriate
examples.

Example 5.6

No unique solution case:

Solve the linear programming problem
Maximize:
Linear Algebra for Economists Lecture note, Wollega University, 92
Department of Economics
By:Kumadebis Tamiru
 Subjected to:

Step 1:
The non-negativity constraints and indicate that the region is bounded by the coordinate
axes in the positive quadrant. The remaining constraint passes through the coordinates
(0, 2) and (4, 0). At the test point (0,0) the inequality is satisfied. Therefore, we are
interested in the region below the line, sine this region contains the test point, (0, 0). This
feasible region sketched in the Figurer 5.9 as below.

Figure 5.9 Sketching the Feasible Region of Set of Inequalities in Example 5.6

(0,2)

X
(4, 0)
Y
Step 2:
The feasible region is a triangle with three corners, (0, 0), (0, 2), and (4, 0).

Step 3:
Corner Objective function value
(0, 0) 0+2(0)=0
(0, 2) 0+2(2)=4
(4, 0) 4+2(0)=4

Linear Algebra for Economists Lecture note, Wollega University, 93

Department of Economics
By:Kumadebis Tamiru
Here, the maximum value of the object function is 4, which is attained at two corner
coordinates (0, 2) and (4, 0). This shows that the problem does not have a unique
solution.

If we superimpose a family of straight lines over the feasible region by setting the
objective function equal to some constant, c. The parallel lines pass through the points
(0, c/2) and (c, 0).

A selection among the family of straight lines is represented in Figure 5.10 below. The
values of c range from 0 to 4. As c increases, the lines sweep across the region from left
to right. Moreover, when c goes above the 4 the lines no longer intersect the feasible
region. The maximum value that c can take is therefore 4. However, instead of the line =4
intersecting the region at only one point, it intersects along a whole line segments of
points. Any point on the line segment lies in the feasible region and the corresponding
value of objective function on this line is 4, which is the maximum value.

Figure 5.10 Sketching the Family of Straight Lines over the Feasible Region
Y

(0,2)
1
X
1 2 3 (4, 0)
c=1 c=2 c=3 c=4
This example suggests that if the maximum (or the minimum) occurs at two corners then
the problem has infinitely many solutions. Any point on the line segment joining these
corners, including the two corners themselves, is also a solution.

Example 5.7

No solution case:
Solve the linear programming problem
Linear Algebra for Economists Lecture note, Wollega University, 94
Department of Economics
By:Kumadebis Tamiru
 Maximize:
Subjected to:

Step 1:
The non-negativity constraints and indicate that the region is bounded by the coordinate
axes in the positive quadrant. The remaining constraints:
The line passes through the coordinates (0, 2) and (8, 0).
The line passes through the coordinates (0, 5) and (5, 0).
The line passes through the coordinates (0, 6) and (3, 0).
A test point (0, 0) does not satisfy any of the corresponding constraints. We are therefore
interested in the region above all of these lines, as shown in Figure 5.11.

Figure 5.11 Sketching the Feasible Region of Set of Inequalities in Example 5.7
Y
6 Unbounded
feasible
5
region
4
3
2
1
1 2 3 4 5 6 7 8 9 X

Step 2:
The feasible region has four corners, (0, 6), (1, 4), (4, 1) and (8, 0).
Step 3:
Corner Objective function value
(0, 6) 3(0)+2(6)=12
Linear Algebra for Economists Lecture note, Wollega University, 95
Department of Economics
By:Kumadebis Tamiru
 (1, 4) 3(1)+2(4)=11
(4, 1) 3(4)+2(1)=14
(8, 0) 3(8)+2(0)=24

From the table the maximum value of the objective function is 24, which occur at (8, 0).
However, we do have slightly unusual situation in that feasible region is not enclosed on
all sides. The feasible region is unbounded in this case. It is opened at the top and, strictly
speaking, it does not make sense to talk about the corners of such a feasible region. We
can substitute huge values of x and y in to and get an ever-increasing result.
Consequently, if the problem is one of maximization it is possible to conclude that it does
not have a finite solution. On other hand, if the problem is one of minimization then it
does have a solution at the corner (1, 4) , which is 11 in value.

This example shows that a linear programming problem may not have a finite solution
when the feasible region is unbounded. However, when a solution does exist, it may be
found by inspecting the corners in the normal way. In practice, linear programming
problems arise from realistic economic situations. We would therefore expect the
problem to possess a sensible (that is, finite) answer and so the difficulty of the non-
existence of a solution rarely occurs.

Applications of Linear Programming:

Here we start with information that given in words and try to express it using the more
precise language of mathematics. Once this has been done, it is a simple matter of
applying mathematical techniques to produce the solution.

Example 5.8

A small manufacturer produces two kinds of good, A and B, for which the demand
exceeds capacity. The production costs for A and B are $6 and $3, respectively for each
unit. The corresponding selling prices are $7 and $4. In addition, the transport costs are
20 cents and 30 cents for each good of type A and B, respectively. The conditions of a

Linear Algebra for Economists Lecture note, Wollega University, 96

Department of Economics
By:Kumadebis Tamiru
bank loan limit the manufacturer to maximum weekly production costs of $2700 and
maximum weekly transport costs of $120. How should the manufacturer arrange
production to maximize production?

As we have discussed before, there are three things constituting a linear programming
problem: a pair of unknown x and y, an objective function that needs maximizing and
minimizing, and some constraints. We consider each of these in turn. The manufacturer
has to decides how many of good A and good B to produce each week. These are
therefore the unknowns of this problem and we denote these unknowns by x and y; that
is, we let
x= number of goods of type A to be made each week
y=number of goods of type B to be made each week

The final sentence of the problem states that the manufacturer should choose these
quantities to maximize profit. Hence we need to find a formula for profit in terms of x
and y. For each type of good A the production costs are $6 and the transport costs are 20
cents. The total cost is therefore $6.20. If the selling price is $7, the profit made on the
single item of A is 80 cents. Consequently, when x goods of type A are made the total
profit is x times this amount, $0.8x. Notice that the question states that ‘demand exceeds
capacity’, so that all goods are guaranteed to be sold. Exactly the same reasoning can be
applied to B. The profit is 70 cents each, so when y goods of type B are made the total
profit is $0.7y.Hence, the profit resulting from the production of both A and B is
0.8x+0.7y
This is then the objective function that we want to maximize.
The next thing to do is that to specify the restrictions that imposed on the production
level. The total weekly production cost must not exceed $2700. The production costs are
$6 for A and $3 for B. Hence, if x goods made type of A and y goods made type of B the
total cost is
6x+3y ≤ 2700

Linear Algebra for Economists Lecture note, Wollega University, 97

Department of Economics
By:Kumadebis Tamiru
Similarly, the total weekly transport cost must not exceed $120. The transport costs are
20 cents for A and 30 cents for B. Hence, if x goods made type of A and y goods made
type of B the total transport cost is
0.2x+0.3y ≤ 120
Although not mentioned explicitly, it is obvious that it is impossible to manufacture a
negative number of goods.

Collecting all information together, the linear programming problem may be stated as
follows:
Maximize: 0.8x+0.7y
Subjected to:
6x+3y ≤ 2700
0.2x+0.3y ≤ 120

Thus, the problem can be solved using graphical approach as follows:

Step 1:
As usual non-negativity constraints indicate that we need only consider points in the
positive quadrant. The line 6x+3y=2700 passes through points (0, 900) and (450, 0). The
line 0.2x+0.3y=120 passes through points (0,400) and (600, 0). Using the origin as the
test point reveals that the region of interest lies below both lines. It is sketched in Figure
5.12.

Figure 5.12 Sketching the Feasible Region of Set of Inequalities in Example 5.

Y
900
800
700
600
500
Linear Algebra for Economists Lecture note, Wollega University, 98
Department of Economics
By:Kumadebis Tamiru
400
300
200
X 100
100 200 300 400 500 600 700

Step2:
The feasible region has four corners altogether, three of which have obvious coordinates
(0, 0), (0, 400), and (450, 0). Unfortunately, it is not at all easy to read off from the
diagram the exact coordinates of the remaining corner. This is formed by the intersection
of two lines
6x+3y ≤ 2700
0.2x+0.3y ≤ 120
The fourth corner therefore has coordinates (375, 150).

Step 3:
Corner Objective function value
(0, 0) 0
(0, 400) 280
(450, 0) 360
(375, 150) 405

The maximum weekly profit is $405, which occurs when 375 goods of A and 150 goods
of B are manufactured.
This example does suggest the following general strategy, which might be helpful:
i. Identify the unknowns and label them x and y
ii. Write down an expression for the objective function in terms of x and y, and
decide whether it needs maximizing and minimizing
iii. Write down all constraints on the variables x and y.

Activity 5.5

Linear Algebra for Economists Lecture note, Wollega University, 99

Department of Economics
By:Kumadebis Tamiru
A food producer uses two processing plant, P1 and P2 that operates 7days per week.
After processing, beef is graded in to high, medium and low quantity food stuffs. High-
quality beef is sold to butchers, medium-quantity beef is used in super-market ready-
meals and the low-quality beef used is dog food. The producer has contracted to provide
120kg of high, 80kg of medium, and 240kg of low-quality beef each week. It costs $4000
per day to run plant P1 and $3200 per day to run the plant P2. Each day P1 processes
60kg of high-quality beef, 20kg of medium-quantity beef and 40kg of low-quality beef.
The corresponding quantities for P2 are 20kg, 20kg and 120kg, respectively. How many
days each week should the plants be operated to fulfill the beef contact most
economically?

5.3. Simplex Method of Linear Programming Problem

When number of variables that is going to be considered is more than two, the graphical
method cannot be more useful because of the complexity of describing the problems on
n-dimensional space. George Dantzig developed an algorithm for solving such linear
programming problems. In this method, properties similar to those of rectilinear figure
having least numbers of vertices, called “Simplex” are exploited and this is the reason
that this method is called the simplex method.

The simplex method is an iterative algorithm (or matrix method) which is a systematic
repetitive solution procedure to obtain the best possible solution under the given
constraints. In every step, the method indicates whether the solution obtained is optimal,
if not, it specifies how to find an improved solution. This method is a little bit long, so we
shall discuss the whole procedure step by step for solving linear programming problems.

The basic requirements for solving linear programming problems are:

i. All constraints must be stated as equations.
ii. The right side of a constraint cannot be negative.

Linear Algebra for Economists Lecture note, Wollega University, 100

Department of Economics
By:Kumadebis Tamiru
 iii. All variables are restricted to non-negative values.

Slack and Surplus Variables:

In a linear programming problem we add some other non-negative variables to the left
hand sides of constraint function in order to convers constraints appear with the signless
than or equal to, “<”, to equality equations. These variables are called slack variables.

For instance, in the constraint

we add a non-negative variable “” on the left hand side to convert it into the equality
equation,

Where “s” denotes slack variable and “” with the subscript 1 implies slack in the first in
equation and “” is slack in the second equation. Thus, in this case slack variable can be
considered as the difference between the requirement and the amount produced.

Similarly, the inequality constraint of greater than or equal to, “≥”, in a linear
programming problem can be changed into equations by subtracting some negative
variables form left hand side of the constraint function. These variables are called surplus
variables.

For instance, in the constraint

we subtract a non-negative variable “” on the left hand side to convert it into the equality
equation,

Where “s” denotes slack variable and “” with the subscript 1 implies slack in the first in
equation.

Linear Algebra for Economists Lecture note, Wollega University, 101

Department of Economics
By:Kumadebis Tamiru
The right side of each constraint must be made positive i.e. if is negative, we multiply
both sides of the corresponding constraint by -1 and changing the sign.

For instance, if the constraint function is then multiplying both sides by -1 we get +in
which =16 is positive. By similarly, can be written as.

Here it is possible to notice that any linear constraint can be converted to linear equation
with the introduction of a slack and surplus variables and making allpositive. A linear
programming problem in which all constraints are in the form of equations is called a
Standard Linear Programming Problem.

Solving Maximization Problem Using the Simplex Method:

To apply simplex method in solving linear programming problems, let us first consider
maximization problems in which all constraints appear with the less than or equal to sign
“<”. In such a problem we shall write all constraints as equations by using slack
variables. We discuss the whole procedure in the following worked out problems.

NB: Row operations play a significant role in applying simplex method, so please recall
back the techniques of row operations first.

Example 5.9

Solve the following maximization problem using the simplex method

Maximize:
Subjected to:

Follow the following necessary steps in solving linear programming problems of

maximization using simplex method.

Step 1:

Linear Algebra for Economists Lecture note, Wollega University, 102

Department of Economics
By:Kumadebis Tamiru
First convert the two linear constraints (inequalities) into equations by adding two slack
variables (one for each constraint) S1 and S2 in the left sides of the in equations and
rewrite the given linear programming problem as follows:
Maximize:
Subjected to:

Better to rewrite the objective function as

Note that we add 0s1+0s2=0 to the objective function and s1+0s2= s1 to constraint one
and 0s1+s2 = s2 to constraint two. In other word, s1 and s2 should appear in the objective
function with coefficient zero, s 1 with coefficient 1 and s2 with zero coefficients in the
first constraint, ands1 with zero coefficient and s2 with coefficient 1 in the second
constraints.

Step 2:
We write the initial augmented matrix ignoring the non-negativity constraints and the
symbol Z1 but considering remaining number coefficients of the equations.

x1 x2 x3 s1 s2 bi ratio

2 1 3 1 0 36 36/3=12
1 2 1 0 1 42 42/1=42
-1 -4 -5 0 0 0 0

Step 3:
The next step involves the activity of determining pivot column, pivot row and pivot
entry. We see that the negative number with the largest absolute value in the objective
row is -5. The column containing the largest negative in absolute value is termed as the
pivot column.

Linear Algebra for Economists Lecture note, Wollega University, 103

Department of Economics
By:Kumadebis Tamiru
The smallest ratio ofthe last column, , to the corresponding positive entries in the pivot
column help us to determine the pivot row. The least or the smallest positive ratio, which
is 36/3 =12 and therefore the first row is the pivot row.

The intersection of the pivot row and pivot column determines the pivot element of the
augmented matrix. Thus, in the above matrix the pivot element is 3.

Step 4:
The next step in the general procedure of simplex method is termed as pivoting.This
consists of several row operations. Divide the pivotal row by the pivot (key entry) and
then add suitable multiples of this row to each of the other rows so that each entry except
the pivot (1) in the pivotal row becomes zero.

The resulting matrix after pivoting becomes:

x1 x2 x3 s1 s2 bi Ratio ratio
operation
2/3 1/3 1 1/3 0 12 r=1/3 r1 12*3=36
1/3 5/3 0 -1/3 1 30 r=r2-r1 30*3/5=18
7/3 -7/3 0 5/3 0 60 r=r3+5r1 60*3/726

This denotes the matrix in the second step of the simplex approach. The solution 12, 30,
and 60 is not optimal solution since the objective function row has the negative element,
which is -7/3.

Step 5:
Thus, procedurally we need to repeat step 3 and 4 until the objective row contains non
negative elements. Since -7/3 is the only negative element in the last row, x2, column is
the pivotal column. Again, since the ratio30*3/5=18 is least, the second row is the pivotal
row and 5/3, which is the intersection of pivotal column and row is the pivot element.

Pivoting the above matrix we obtain the simplex matrix of:

Linear Algebra for Economists Lecture note, Wollega University, 104

Department of Economics
By:Kumadebis Tamiru
 x1 x2 x3 s1 s2 bi Ratiooperation
3/5 0 1 2/5 -1/5 6r =r1-1/3 r2
1/5 1 0 -1/5 3/5 18r =3/5r2
14/5 0 0 6/5 7/5 102 r =r3+7/3r2

We see that the objective row (last row) contains no negative entries and hence this is the
final optimal simplex matrix.

Step 6:
Now to find the value of each variable for which z is maxima we proceed as follows. If
in a column of the final simplex matrix one element is 1 and the other elements are all
zeros, then the last entry in the row containing 1 is the value of the corresponding
variable. Otherwise the value of the variable is zero. Thus, x1 =0, x2=18, x3=6, s1=0, s2 =
0, and the Maximum value of z is 102.

Example 5.10

Solve the following maximization problem using the simplex method

Maximize:
Subjected to:

Solution:
Introducing slack variables S1, S2 and S3 we convert the given inequalities into equations
and rewrite the constraints and the objective functions as follows:

The initial augmented simplex matrix of the above linear programming problem is
x1 x2 s1 s2 s3 bi ratio

Linear Algebra for Economists Lecture note, Wollega University, 105

Department of Economics
By:Kumadebis Tamiru
 1 4 1 0 0 8 8/1=8
4 2 0 1 0 12 12/4=3
2 3 0 0 1 14 14/2=7
-5 -2 0 0 0 0

Pivoting this matrix we get the following simplex matrix

x1 x2 s1 s2 s3 bi Ratio
operation
0 7/2 1 -1/4 0 5 r=r1-r2
1 1/2 0 1/4 0 3 r=1/4r2
0 2 0 -1/2 1 8 r=r3-2r2
0 1/2 0 5/4 0 15 r=r4+5r2

We see that the objective row contains no negative entries. Hence, this is the final
optimal simplex for the above linear programming problem given and the maximum
value of the objective function is z=15. This maximum value is obtained for x1=3, x2=0,
s1=5, s2=0, and s3=8.
i. x1 column (1st column) consists all zero entries except the 2 nd row which is 1, then
value of x1 is obtained in the constant column (bi) corresponding to 2nd row, i.e., x1=3
ii. x2=0, since the entries of the 2nd column doesn’t satisfy the condition
iii. s1 = 5, since all entries except the first row (1) are zeros’
iv. s2=0, because it does not satisfy the condition
v. s3 =8, since all entries except the third row (1) are zeros. Z=5(3)+2(0)

Example 5.11

Solve the following maximization problem using simplex method

Maximize:
Subjected to:

Solution:

Linear Algebra for Economists Lecture note, Wollega University, 106

Department of Economics
By:Kumadebis Tamiru
Introducing slack variables s1 and s2 we can rewrite the given constraints and the
objective function as follows:

The initial simplex matrix of the given linear programming problem is

x1 x2 x3 s1 s2 bi ratio

3 4 1 1 0 2 2/4=1/2
1 3 2 0 1 1 1/3
-3 -6 -2 0 0 0 0

The last row is the objective row clearly 2 nd column is the pivotal column and the entry 3
corresponding to the least ratio is a pivot entry.
Pivoting this matrix we get the following simplex matrix
x1 x2 x3 s1 s2 bi Ratio ratio
operation
5/3 0 -5/3 1 -4/3 2/3 r=r1-4r2 2/3*3/5=2/5
1/3 1 2/3 0 1/3 1/3 r=1/3r2 1/3*3/1=1
-1 0 2 0 2 2 r=r3+6r2 2/1=2

In the above simplex matrix, first column is the pivotal column and the entry 5/3
corresponding to the least ratio 2/5 is a pivot entry and since the last row contains a
negative entry, the procedure continues.

Pivoting again this matrix we obtain the following simplex matrix.

x1 x2 x3 s1 s2 bi Ratio
operation
1 0 -1 3/5 -4/5 2/5 r=3/5r1
0 1 1 -1/5 3/5 1/5 r=r2 -1/3r1
0 0 1 3/5 3 12/5 r=r3+6r2

We see that the last row (objective row) contains no negative entries. Thus, the matrix
reaches at its optimal stage. Accordingly, the maximum value of the objective function is
z =12/5 which is obtained for x1=2/5, x2=1/5, x3=0, s1=0,and s2=0.

Linear Algebra for Economists Lecture note, Wollega University, 107

Department of Economics
By:Kumadebis Tamiru
In general from what we have discussed using examples, we are now in the position to
state the procedure that required in simplex method for all constraints appear with less
than or equal to sign, “≤”, as follows:
i. Introduce slack variables and rewrite the given inequalities and the objective
function as equations.
ii. Write the initial simplex matrix by omitting the non-negativity constraints of the
variables and z
iii. Find the pivot considering the negative number with the largest absolute value in the
objective row and the least of ratios which are entries in the last column to the
positive entries in the pivotal column.
iv. Perform the pivoting i.e. divide the row containing the pivot by the pivot and add (or
subtract) suitable multiples of the new pivotal row (i.e. the row obtained by dividing
by the pivot) to each of the other rows so that each entry except the pivot 1 becomes
zero.
v. Repeat steps 3 and 4 until the objective row contains no negative entries and final
simplex matrix is obtained.
vi. The last entry in the objective row of the final simplex matrix is the maximum
possible value of the objective function.

Activity 5.6

Solve the following maximization problem using simplex method

Maximize:
Subjected to:

Solving Minimization Problem Using the Simplex Method:

Note that minimizing an objective function Z with constraints of sign “<”is the same as
maximizing (-Z) of the objective function.

Linear Algebra for Economists Lecture note, Wollega University, 108

Department of Economics
By:Kumadebis Tamiru
Example 5.12

Solve the following minimization problem using simplex method:

Minimize:
Subjected to:

Solution:
Clearly to minimize z = x1-2x2+8 is to means maximize –z= -x1+2x2-8(Multiplying both
sides by -1). Introducing slack variables s1, s2 and s3 we can rewrite the constraints as
follows:

The initial simplex matrix of the given linear programming problem is

x1 x2 s1 s2 s3 bi ratio

1 1 1 0 0 7 7/1=7
-1 1 0 1 0 1 1/1=1
1 0 0 0 1 5
1 -2 0 0 0 -8

Clearly the second column is the pivotal column and the entry 1 is the pivot entry.

Pivoting this matrix as usual we get the following simplex matrix:

x1 x2 s1 s2 s3 bi Row ratio
operation
2 0 1 -1 0 6 r=r2 -r1 6/2=3
-1 1 0 1 0 1 r=r2
1 0 0 0 1 5 r=r3 5/1=5
-1 0 0 0 0 -6 r=r4 +2r2

Linear Algebra for Economists Lecture note, Wollega University, 109

Department of Economics
By:Kumadebis Tamiru
In the above matrix, first column is the pivotal column and the entry 2 is the pivot
element corresponding to the least ratio.

Pivoting this matrix we obtain the following simplex matrix.

x1 x2 s1 s2 s3 bi Row
operation
1 0 1/2 -1/2 0 3 r=1/2r1
0 1 0 1/2 0 4 r=r2+r1
0 0 -1/2 1/2 1 2 r=r3-r1
0 0 1/2 1/2 0 -3 r=r4 +r2

We see that the objective row except the last entry contains no negative entries. Hence,
this is the final simplex matrix and the maximum value of the objective function z=-3
which is obtained for x1=3, x2=4, s1=0, s2=0, s3=2. Hence, the minimum value of z =
Max. Value of –z = 3

Activity 5.7

Solve the following minimization problem using simplex method:

Minimize:
Subjected to:

5.4 Formulation of the Dual of Linear Programming Problems

Let consider a given linear programming problem of minimization with greater than or
equal to sign “≥”

Linear Algebra for Economists Lecture note, Wollega University, 110

Department of Economics
By:Kumadebis Tamiru
 Where i= 1, 2, 3. . . n

The method discussed in the earlier section cannot be used here. In such a situation we
transform the given linear programing problem to a new linear programming problem as
under:

Where i= 1, 2, 3. . . n
This new linear programming problem is called dual to the first linear programming
problem. The former is called primal to the latter. The dual can now be solved by means
of the simplex method described earlier.

Theorem: Primal has optimal solution if and only if dual has optimal solution and further
minimum z = maximum. Using the duality principle, we can, by means of simplex
method, solve linear programming problems of minimization with “>” constraints.

Example 5.13

Solve the following linear programming problem

Dual linear programming problem the above problem will be

using the simplex method we can solve the dual linear programming problem and use the
original variables x1 and x2 as slack variables of the dual problem.

Linear Algebra for Economists Lecture note, Wollega University, 111

Department of Economics
By:Kumadebis Tamiru
The initial simplex matrix of the given dual linear programming problem is
y1 y2 y3 x1 x2 bi ratio

2 1 1 1 0 2 2/1=2
1 3 3 0 1 4 4/31.34
-14 -12 -18 0 0 0

Pivoting this matrix as usual we obtain

y1 y2 y3 x1 x2 bi Row operation ratio

5/3 0 0 1 -1/3 2/3 r=r1-r2 2/3*3/5=2/5

1/3 1 1 0 1/3 4/3 r=1/3r2 4/3*3/1=4
-8 6 0 0 6 0 r=r3+18r1

Again pivoting this matrix we obtain

y1 y2 y3 x1 x2 bi Row operation
1 0 0 3/5 -1/5 2/5 r=3/5r1
0 1 1 -1/5 2/5 6/5 r=1/3r2-1/3 r1
0 6 0 24/5 22/5 136/5 r=r3+8r1

Hence, the optimal value for the dual linear programming problem is Max  = 136/5
which obtained at y1=2/5, y2 = 0 and y3 =6/5 or it is obtained at x1=24/5 and x2=22/5.

Example 5.14

Solve the following linear programming problem

Solution:
Dual of the above linear programming problem will be

Linear Algebra for Economists Lecture note, Wollega University, 112

Department of Economics
By:Kumadebis Tamiru
Introducing slack variables we can rewrite it as follows:

Accordingly, the initial simplex matrix of the given dual linear programming problem is
y1 y2 x1 x2 x3 bi ratio

1 3 1 0 0 2 2/1=2
4 1 0 1 0 9 9/42.25
2 2 0 0 1 1 ½=0.5
-5 -4 0 0 0 0

Pivoting this matrix as usual we obtain,

y1 y2 x1 x2 x3 bi Row
operation
0 2 1 0 -1/2 3/2 r=r1- r3
0 -3 0 1 -2 7 r=r2-4 r3
1 1 0 0 1/2 1/2 r=1/2r3
0 1 0 0 5/2 5/2 r=r4+5r3

We can see that this is the final simplex matrix. Hence, the optimal value for the dual
linear programming problem is Max  = 5/2 which obtained at y1=1/2 and y2 = 0 or it is
obtained at x1=3/2, x2=7 and x3=0.

Review Exercise

1. Use graphical method to solve the following linear programming problems

a. Maximize: Z =20x+30y b. Maximize Z = 4x+6y

Subject to: Subject to:
x+y≥12 4x+3y ≥36
2x+3y≥30 2x+y≥14
y≥4 4x+2y>16
x,y≥0 x,y≥0

Linear Algebra for Economists Lecture note, Wollega University, 113

Department of Economics
By:Kumadebis Tamiru
 c. Minimize Z = 4x+5y d. Minimize Z = 3x+4y
Subject to: Subject to:
2x+y≥14 4x+3y>36
2x+3y>30 2x+y>16
y≥4 4x+2y ≥ 16
x,y≥0 x,y≥0

2. Use simplex method to solve the following linear programming problems

a. Maximize: Z =4x1+10x2 b. Maximize Z = x1+x2+3x3

Subject to: Subject to:
2x1+x2<50 3x1+2x2+x3<3
2x1+5x2<100 2x1+x2+2x3<2
x1, x2>0 x1,x2,x3>0

c. Maximize Z = -x1+3x2-2x3 d. Minimize Z =6x1+8x2+x3

Subject to: Subject to:
3x1-x2+2x3<7 3x1+5x2+3x3>20
-x1+2x2<6 x1+3x3+2x3>9
4x1-3x2-8x3>-10 6x1+2x2+5x3>30
x1,x2,x3>0 x1,x2,x3>0

e. Minimize Z = 6x1+3x2 f. Minimize Z = x1+2x2+x3

Subject to: Subject to:
x1+x2>4 x1-3x2+4x3 >12
3x1+4x2>12 3x1+x2 +20x3>10
x1,x2>0 x1-x2 -x3< -8
x1-x2 -x3> 0
3. Application Questions

Linear Algebra for Economists Lecture note, Wollega University, 114

Department of Economics
By:Kumadebis Tamiru
 a) A dealer wishes to purchase a number of fans and sewing machines. He has
only birr 5,760 to invest and has space at most for 20 items. A fan costs him
birr 360 and a sewing machine birr 240. His expectation is that he can sell a
fan at a profit of birr 22 and a sewing machine at a profit of birr 18.
Assuming that he can sell all the items that he can buy, how should he invest
his money in order to maximize his profit? Formulate this problem as a linear
programming problem and then use graphical method to solve it.

b) A factory uses three different resources for the manufactures of two different
products. 20 units of the resource A,12 units of B and 16 unit of C being
available 1 unit of the first product requires 2,2 and 4 units of the respective
resources and 1 unit of the second product requires 4,2 and 0 units of the
respective resources. It is known that the first product gives a profit of $2 per
unit and the second $3. Formulate the linear programming problem. How
many units of each product should be manufactured for maximizing the
profit? Solve it graphically.

c) A farmer has 100 acres of land on which he can grow corn, wheat or bean.
Each acre of corn costs birr 100 for preparation, requires 7 man days of work
and yields a profit of birr 30. An acre of wheat costs birr 120 to prepare,
requires 10 man days of work and yields a profit of birr 40. An acre of bean
costs birr 70 to prepare, requires 8 man day of work and yields profit of birr
20. If the farmer has birr 100,000 for preparation and can count on 8000 man-
days work, how many acres should be allocated to each crop to maximize the
total profit?(Answer: corn= 250 acres, wheat 625 acres, bean 0 acres and
profit = 32500).

d) A television company has three major departments for its manufacture of two
models A and B.
Per unit time requirements (hours)
Model A Model B Model C
Department I 4.1 2.0 1,600
Linear Algebra for Economists Lecture note, Wollega University, 115
Department of Economics
By:Kumadebis Tamiru
 Department II 2.5 1.0 1,200
Department III 4.5 1.5 1,600

The Marginal profit of model A is birr 400 each and that of model B is birr
100 each. Assuming that the company can sell any quantity of either product
due to favorable market conditions, determine the optimum output for both
models and highest possible profit for this month. (Answer: 800/3 units of
both A and B. Maximum profit is birr 1333.3)

e) Food x contains 6 units of vitamin A per gram and 7 units of vitamin B per
gram and cost 12 birr per gram. Food y contains 8 units of vitamin A per
gram and 12 units of vitamin B per gram and cost 20 birr per gram. The daily
minimum requirements of vitamin A and vitamin B are 100 units and 120
units respectively. Find the dual problem and the minimum cost of the
product mix. (Answer: Food x = 15 units, Food Y= 5/4 units and minimum
cost = birr 2.05)

f) A small metal part manufacturing company most of whose business comes

from special orders, has decided to make use of the idle times of the
equipment. The company has determined that two particular parts A and B
could readily be sold at per unit profits of birr 30 and birr 20 respectively, in
quantities up to 144 units of each part per day. In order to make each unit of
part A, 5 minutes are required on a lathe, 7 minutes on a mill and 4 minute on
a grinder. To make each unit of part B, 3,9, 7 minutes are required
respectively on a lathe, mill and grinder. The company has one machine of
each type available. Lathe is currently idle for 85 minutes per day. How many
parts of type A and B should the company produce in order to make maximum
use of the idle time of its equipment?

Reference

Linear Algebra for Economists Lecture note, Wollega University, 116

Department of Economics
By:Kumadebis Tamiru
 Chiang, A.C., (1984), Fundamental Methods Of Mathematical Economics,
Mcgraw-Hill, Inc.

Grossman, S.I., (1993), Finite Mathematics, 2nd Ed., Wm.C.Brown Publishers.

Henderson, J.M. &Quandt,R.E., (1980), Microeconomic Theory: A Mathematical

Approach, 3rd Ed., Mcgraw Hill.

Yamane, T., (1978), Mathematics For Economists: An Elementary Survey, 2nd

Ed., Prentice-Hall.

Monga, G.S., (1972), Mathematics And Statistics For Business And Economics,
Vicas Publishing House.

Linear Algebra for Economists Lecture note, Wollega University, 117

Department of Economics
By:Kumadebis Tamiru