Mathematics for Economics and Business Section 2: Systems of linear equations.

Lecture
2.1. Rank of a matrix
Let be given a matrix 𝐴 ∈ 𝑀𝑚×𝑛 (ℝ).
The rank of a matrix is a natural number or zero uniquely associated with each matrix.
We will denote the rank of a matrix 𝐴: 𝑟𝑎𝑛𝑘(𝐴) or 𝑟(𝐴) or 𝑅(𝐴), 𝑟(𝐴) ∈ ℕ ∪ {0}.
Definition 2.1
1) If 𝐴 = 𝟎, then 𝑟(𝐴) = 0.
2) If 𝐴 ≠ 𝟎, then the rank 𝑟(𝐴) of the matrix 𝐴 is the order of the largest square submatrix of 𝐴 with a nonzero
determinant.
Remarks
 The rank of a matrix 𝐴 is less or equal to its dimensions: 𝑟(𝐴) ≤ min{𝑚, 𝑛}.
 For a square matrix 𝐴 ∈ 𝑀𝑛 (ℝ) ∶
a) if 𝑑𝑒𝑡𝐴 ≠ 0 then 𝑟(𝐴) = 𝑛,
b) if 𝑑𝑒𝑡𝐴 = 0 then 𝑟(𝐴) < 𝑛.
Example 2.1. Find the rank of the given matrix:
1 3
𝐴=[ ], 𝑟(𝐴) = ⋯ …;
2 4

1 2
𝐵=[ ], 𝑟(𝐵) = ⋯ ..;
2 4

3 1 2
𝐶=[ ], 𝑟(𝐶) = ⋯ ...
0 4 1

2.2. General system of linear equations and its matrix form

Definition 2.2 A general system of 𝒎 equations and 𝒏 variables can be written in a form:
𝑎11 𝑥1 + 𝑎12 𝑥2 + ⋯ + 𝑎1𝑛 𝑥𝑛 = 𝑏1
𝑎 𝑥 + 𝑎22 𝑥2 + ⋯ + 𝑎2𝑛 𝑥𝑛 = 𝑏2
(∗) { 21 1 where 𝑎𝑖𝑗 , 𝑏𝑗 ∈ ℝ (are given real numbers).
⋮
𝑎𝑚1 𝑥1 + 𝑎𝑚2 𝑥2 + ⋯ + 𝑎𝑚𝑛 𝑥𝑛 = 𝑏𝑚

Elements 𝑎𝑖𝑗 (𝑖 = 1,2, … , 𝑛, 𝑗 = 1,2, … 𝑚) are called coefficients of the system.

Components 𝑏𝑗 (𝑗 = 1,2, … 𝑚) are called constants (or free coefficient).
Components 𝑥𝑖 (𝑖 = 1,2, … , 𝑛) are unknowns.

 Beata Ciałowicz ~ 13 ~
 Mathematics for Economics and Business Section 2: Systems of linear equations. Lecture
From this system we can create few matrices:
𝑎11 𝑎12 … 𝑎1𝑛
𝑎21 𝑎22 … 𝑎2𝑛
 The matrix formed by the coefficients of the system is: 𝐴 = [ ⋮ ⋮ ⋱ ⋮ ]
𝑎𝑚1 𝑎𝑚2 … 𝑎𝑚𝑛 𝑚×𝑛
a matrix of coefficients
𝑥1
𝑥2
 The matrix formed from the unknowns: 𝑋=[ ⋮ ] a matrix of unknowns
𝑥𝑛 𝑛×1
𝑏1
𝑏
 The matrix formed from the constant terms is: 𝐵 = [ 2 ] a matrix of constants
⋮
𝑏𝑚 𝑚×1
0
If the matrix of constants is the zero vector 𝐵 = 𝟎 = [0] , then the system of equations is called
⋮
0 𝑚×1
a homogeneous system, otherwise (𝐵 ≠ 𝟎) it is called an inhomogeneous system of equations.
 Using coefficients and constants, we can write a matrix called an augmented matrix (matrix of coefficients
is completed by a column of constants):
𝑎11 𝑎12 … 𝑎1𝑛 𝑏1
𝑎 21 𝑎22 … 𝑎2𝑛 𝑏2
𝐴𝐵 = [𝐴|𝐵] = [ ]
⋮ ⋮ ⋱ ⋮ ⋮
𝑎𝑚1 𝑎𝑚2 … 𝑎𝑚𝑛 𝑏𝑚 𝑚×(𝑛+1)

A system of equations (∗) can be expressed as the matrix equation (can be written in matrix notation):

(∗) ⟺ 𝐴 ∙ 𝑋 = 𝐵 matrix form of a system

𝑎11 𝑎12 … 𝑎1𝑛 𝑥1 𝑏1

𝑎21 𝑎22 … 𝑎2𝑛 𝑥2 𝑏2
[ ⋮ ⋮ ⋱ ⋮ ] ∙ [⋮] = [ ⋮ ]
𝑎𝑚1 𝑎𝑚2 … 𝑎𝑚𝑛 𝑥𝑛 𝑏𝑚
↑ ↑ ↑
matrix of coefficients vector of unknowns constant matrix
Example 2.2. Construct all matrices characterized the given system of linear equations:
𝑥 − 2𝑦 + 3𝑧 + 𝑡 = 3
{ 3𝑥 − 𝑦 + 2𝑧 − 𝑡 = −2
−5𝑥 + 𝑦 + 𝑧 + 5𝑡 = 0

 Beata Ciałowicz ~ 14 ~
 Mathematics for Economics and Business Section 2: Systems of linear equations. Lecture
Solvability of a linear system of equations
A system of equations (∗) is called solvable or consistent if it has a solution, i.e. there exists at least one vector
(𝑥10 , 𝑥20 , … , 𝑥𝑛0 ) (a sequence of 𝑛 numbers) such that all 𝑚 equations in (∗) are satisfied.
Otherwise, it is called inconsistent.
A solution to (∗) needs not exist, and if a solution exists, it need not to be unique.
The solvability of a system of linear equations will depend on the ranks of matrices 𝐴 and 𝐴𝐵 .
There are three possibilities:
1. system has no solution (inconsistent system - the system has contradictory equations).
2. system has an infinite number of solutions (indeterminate system).
3. system has an unique solution (determinate system).

2.3. Cramer’s system

We have a very important special case of consistent systems of linear equations when the number of equations is
equal to the number of unknowns.
Definition 2.3 Let 𝐴 = [𝑎𝑖𝑗 ]𝑛×𝑛 and 𝑑𝑒𝑡𝐴 ≠ 0.
𝑎11 𝑥1 + 𝑎12 𝑥2 + ⋯ + 𝑎1𝑛 𝑥𝑛 = 𝑏1
𝑎 𝑥 + 𝑎22 𝑥2 + ⋯ + 𝑎2𝑛 𝑥𝑛 = 𝑏2
Then the system (∗∗) { 21 1 is called a Cramer’s system.
⋮
𝑎𝑛1 𝑥1 + 𝑎𝑛2 𝑥2 + ⋯ + 𝑎𝑛𝑛 𝑥𝑛 = 𝑏𝑛

Remarks
 A Cramer’s system has always a unique solution.
 The determinant of a matrix of coefficients 𝐷 = 𝑑𝑒𝑡𝐴 is called the coefficient determinant ……………
(or main determinant).

Theorem 2.1 (Cramer’s theorem)

A Cramer system has a unique solution in the form: 𝑋 = 𝐴−1 ∙ 𝐵.

2𝑥 + 𝑦 = 4
Example 2.3. Let us consider { the system of equations in two variables. Find the solution
4𝑥 + 3𝑦 = 6
by using Cramer’s theorem.

 Beata Ciałowicz ~ 15 ~
 Mathematics for Economics and Business Section 2: Systems of linear equations. Lecture
Theorem 2.2. (Cramer’s rule) The solution of a Cramer’s system is given by (𝑥1 , 𝑥2 , … , 𝑥𝑛 ),
𝑎11 𝑎12 … 𝑏1 … 𝑎1𝑛
𝐷𝑖 𝑎21 𝑎22 … 𝑏2 … 𝑎2𝑛
where: 𝑥𝑖 = , 𝐷𝑖 = | |
𝐷 ⋮ ⋮ ⋱ ⋮ ⋱ ⋮
𝑎𝑛1 𝑎𝑛2 … 𝑏𝑛 … 𝑎𝑛𝑛
↑
𝑖th 𝑐𝑜𝑙𝑢𝑚𝑛
is the determinant obtained from 𝐷 by replacing the 𝑖th column by the column of constants 𝐵.
2𝑥 + 𝑦 = 4
Example 2.4 Consider the same system of equations: { where 𝐷 = det𝐴 = 2. Find the solution
4𝑥 + 3𝑦 = 6
by using Cramer’s rule.

2.4. General rules of linear equation systems

Theorem 2.3 (Kronecker – Cappelli) (necessary and sufficient condition)
A system of equations (∗) has at least one solution (is consistent) if and only if 𝑟(𝐴) = 𝑟(𝐴𝐵 ).

Remarks
 If 𝑟(𝐴) ≠ 𝑟(𝐴𝐵 ) the system has no solution.
 If 𝑟 denotes the rank of equations, i.e. 𝑟 = 𝑟(𝐴) = 𝑟(𝐴𝐵 ), then:
a) for 𝑟 = 𝑛 the system has a unique solution,
b) for 𝑟 < 𝑛 the system has a infinitely many solutions.
 The homogeneous system of equations (𝐵 = 𝟎) always has at least one solution, the so-called trivial
solution: 𝑥1 = 𝑥2 = ⋯ = 𝑥𝑛 = 0.

Conditions for solution of 𝒏 simultaneous equations

Type of Equations Conditions of Rank Nature of Solution
Consistent equations 𝑟(𝐴) = 𝑟(𝐴𝐵 ) = 𝑟 = 𝑛 Unique solution
Infinite number of solutions
𝑟(𝐴) = 𝑟(𝐴𝐵 ) = 𝑟 < 𝑛
with 𝑛 − 𝑟 arbitrary constants
Inconsistent equations 𝑟(𝐴) < 𝑟(𝐴𝐵 ) No solution
Homogeneous equations Unique solution (the trivial solution)
𝑟(𝐴) = 𝑟(𝐴𝐵 ) = 𝑛
(𝐵 = 𝟎) 𝑥𝑖 = 𝟎
Infinite number of solutions with 𝑛 − 𝑟
𝑟(𝐴) = 𝑟(𝐴𝐵 ) = 𝑟 < 𝑛
arbitrary constants (parameters)

 Beata Ciałowicz ~ 16 ~
 Mathematics for Economics and Business Section 2: Systems of linear equations. Lecture
The basic idea underlying the method of solution described below is that certain transformations of a system of
linear equations do not affect the solution of the system, such that the solution of the transformed system is easy
to find. Generally we want to change the system (∗) into the equivalent Cramer’s system.
We say that two systems of equations are equivalent if they have the same solution.
Example 2.5 Solve the given system of equations:
−2𝑥 + 3𝑦 = 2
−2𝑥 + 3𝑦 = 2 −2𝑥 + 3𝑦 − 𝑧 = 2
1) { 2) { 3){ 𝑥 − 𝑦 = 1
4𝑥 − 6𝑦 = 1 𝑥 − 6𝑦 + 3𝑧 = −4
3𝑥 − 5𝑦 = −5

2.5. Gauss-Jordan elimination method

Elementary row or column operations (elementary transformations) are:
1. The interchanging of any two rows (columns) of a matrix (𝑟𝑖 ↔ 𝑟𝑗 ; ↔ means “change”).
2. Multiplying of each element of a row (column) by the same nonzero number (𝑘𝑟𝑖 ↦ 𝑟𝑖′ ; ↦ means “replaces”).
3. The replacement of any row (column) by the sum of that row (column) and a nonzero multiple of some other
row (column) (𝑟𝑖 + 𝑘𝑟𝑗 ↦ 𝑟𝑖′ ).

Applications:
1. Calculating determinants: we may transform a determinant with the help of elementary transformations
(no.3) into a form such that it contains as many zeros as possible.
2. Finding the inverse of a matrix.
3. Finding the rank of a matrix.
4. Solving general system of linear equations (the Gauss-Jordan elimination method).
Remark The row operations on the augmented matrix of a system produce the augmented matrix of
an equivalent system, i.e., a system with the same solution as the original one.
3 0 −2 3 1 4
Example 2.6. −1
Find a rank of the matrix: 𝐷 = [ 2 −22 4 0 −4 1]
2 3 6 4
4 0 4 6 3 9

 Beata Ciałowicz ~ 17 ~
 Mathematics for Economics and Business Section 2: Systems of linear equations. Lecture
The Gauss-Jordan elimination method to solve a system of linear equations is described in the
following steps:
Step 1. Write the augmented matrix of the system.
Step 2. Use row operations to transform the augmented matrix in the form described below, which is called the
reduced echelon form form (REF).
REDUCED ECHELON FORM (REF)
A matrix 𝐴 ∈ 𝑀𝑚,𝑛 (ℝ) is in reduced echelon form if it has leading ones on the main diagonal and zeros above
𝐼 𝑅
and below the leading ones: [ 𝑘 ], where 𝐼𝑘 is an identity matrix of dimension 𝑘; 𝑅 is a (nonzero) matrix
𝟎 𝟎
of rests, 𝟎 is a zero matrix.
Step 3. Stop process in step 2 if you obtain a row whose elements are all zeros except the last one on the right.
In that case, the system is inconsistent and has no solution. Otherwise, finish step 2 and read the solutions of the
system from the final matrix.
Example 2.7. Solve the system of equations by using the Gauss-Jordan elimination method:
−2𝑥 + 3𝑦 = 2 𝑥 − 2𝑦 = 3 2𝑥 + 𝑦 − 3𝑧 = −2
−3𝑥 + 2𝑦 − 𝑧 + 𝑡 = 4
𝟏) { 𝑥 + 4𝑦 = 1 2) { 3){3𝑥 + 5𝑦 = −1 4){ 3𝑥 − 4𝑦 + 2𝑧 = 1
6𝑥 − 4𝑦 + 2𝑧 − 3𝑡 = 1
3𝑥 + 𝑦 = 3 𝑥 + 9𝑦 = −7 𝑥 + 6𝑦 − 8𝑧 = −5

Remark. During the tests and the final exam substitution method is FORBIDDEN (to use the
substitution method we solve one equation for one of the variables, then we substitute in the other
equation and solve).
 Beata Ciałowicz ~ 18 ~
 Section 2: Systems of linear equations. Lecture
2.6. Input-output model
Wassily Leontief (1905-1999); 1973 - the Nobel Memorial Prize in Economic Sciences. The input-output
model uses a matrix representation of a nation's (or a region's) economy to predict the effect of changes in one
industry on others and by consumers, government, and foreign suppliers on the economy.
 Basic notation
Assumptions:
 Let be given an economy as a system of 𝑛 independent industries.
 Suppose that in this economy each industry produces 𝑌𝑗 units.
 Each industry consumes some of the goods produced by the other industries, including itself.
 Each industry produces total product, sells part of them to other industries and consumes.
Notation:
𝑌1
𝑌2
 𝑌𝑗 the total production (output) of j-th industry (in physical or monetary units), 𝑌 = [ ]
⋮
𝑌𝑛 𝑛×1
a vector of total production
𝑦11 𝑦12 … 𝑦1𝑛
𝑦21 𝑦22 … 𝑦2𝑛
 𝑦𝑗𝑖 the flow from j-th industry to i-th industry (intersector flow), 𝑀 = [𝑦𝑗𝑖 ] = [ ⋮ ⋮ ⋱ ⋮ ]
𝑦𝑛1 𝑦𝑛2 … 𝑦𝑛𝑛 𝑛×𝑛
a matrix of flows (input-output matrix)
𝑦1
𝑦2
 𝑦𝑗 the final production (output) of j-th industry 𝑦=[⋮] a vector of final production
𝑦𝑛 𝑛×1
 The market equilibrium condition
Since the economy is closed: 𝑦𝑗 = 𝑌𝑗 − (𝑦𝑗1 + 𝑦𝑗2 + ⋯ + 𝑦𝑗𝑛 ) ⟺ 𝑌𝑗 = 𝑦𝑗 + (𝑦𝑗1 + 𝑦𝑗2 + ⋯ + 𝑦𝑗𝑛 )
𝑦1 = 𝑌1 − (𝑦11 + 𝑦12 + ⋯ + 𝑦1𝑛 )
𝑦 = 𝑌2 − (𝑦21 + 𝑦22 + ⋯ + 𝑦2𝑛 )
Thus the market equilibrium condition: { 2
⋮
𝑦𝑛 = 𝑌𝑛 − (𝑦𝑛1 + 𝑦𝑛2 + ⋯ + 𝑦𝑛𝑛 )
 An input-output table (a transactions tableau)
All information about our model mentioned before (flows and final and total production) we may write in a form
of input-output table:
𝑦𝑗𝑖
𝑦𝑗 𝑌𝑗
𝑖/𝑗 1 2 … 𝑛
1 𝑦11 𝑦12 … 𝑦1𝑛 𝑦1 𝑌1
2 𝑦21 𝑦22 … 𝑦2𝑛 𝑦2 𝑌2
⋮
𝑛 𝑦𝑛1 𝑦𝑛2 … 𝑦𝑛𝑛 𝑦𝑛 𝑌𝑛
Example 2.8. Find unknown elements in input-output tables
𝑦𝑗𝑖 𝑗/𝑖 1 2 3 4 𝑦𝑗 𝑌𝑗
𝑦𝑗 𝑌𝑗
𝑗/𝑖 1 2 3 1 10 30 50 40 170
1 10 15 40 5 2 30 20 5 10 80
2 18 14 20 7
3 40 10 15 35 100
𝑛 10 5 30 8
4 30 50 100 40 20
 Beata Ciałowicz ~ 11 ~
 Section 2: Systems of linear equations. Lecture
 Cost coefficient
To describe a technology profitability of industries in the economy a special cost coefficient is used. General
assumption is that total product is proportional to inputs. This proportion is constant in time and denoted 𝑎𝑗𝑖 .
𝑎𝑗𝑖 – is fixed cost (production coefficient; input-output coefficient); the output (monetary value of ) from 𝑖th
industry which is needed to produce one unit (of monetary value) of output of 𝑗th industry
𝒚𝒋𝒊
𝒚𝒋𝒊 = 𝒂𝒋𝒊 ∙ 𝒀𝒊 ⟺ 𝒂𝒋𝒊 = (0 ≤ 𝑎𝑗𝑖 ≤ 1)
𝒀𝒊

𝐴 = [𝑎𝑗𝑖 ]𝑛×𝑛 - a matrix of cost coefficients (a technology matrix)

industry is profitable  if each industry makes a positive profit  the sum of entries in each column of a
matrix A is less than one

 A matrix form of input-output model

By substitution 𝑦𝑗𝑖 = 𝑎𝑗𝑖 ∙ 𝑌𝑖 in the market equilibrium condition we will obtain:

𝑦1 = 𝑌1 − (𝑎11 𝑌1 + 𝑎12 𝑌2 + ⋯ + 𝑎1𝑛 𝑌𝑛 ) 𝑦1 𝑌1 𝑎11 𝑎12 … 𝑎1𝑛 𝑌1

𝑦 = 𝑌2 − (𝑎21 𝑌1 + 𝑎22 𝑌2 + ⋯ + 𝑎2𝑛 𝑌𝑛 ) 𝑦 2 𝑌 𝑎 21 𝑎22 … 𝑎2𝑛 𝑌2
{ 2 ⟺ [ ⋮ ] = [ 2] − [ ⋮ ⋮ ⋱ ⋮ ]∙[ ⋮ ]
⋮ ⋮
𝑦𝑛 = 𝑌𝑛 − (𝑎𝑛1 𝑌1 + 𝑎𝑛2 𝑌2 + ⋯ + 𝑎𝑛𝑛 𝑌𝑛 ) 𝑦𝑛 𝑌𝑛 𝑎𝑛1 𝑎𝑛2 … 𝑎𝑛𝑛 𝑌𝑛

𝒚=𝒀−𝑨∙𝒀 ⟺ 𝒚 = (𝑰 − 𝑨) ∙ 𝒀 ⟺ 𝒚=𝑳∙𝒀

1 0 … 0
where 𝐼 = [0 1 … 0] is an identity matrix; 𝐿 = 𝐼 − 𝐴 is called Leontieff matrix.
⋮ ⋮ ⋱ ⋮
0 0 … 1 𝑛×𝑛

If det(𝐼 − 𝐴) = det 𝐿 ≠ 0, then: 𝑌 = (𝐼 − 𝐴)−1 ∙ 𝑦.

Example 2.9 In the input-output model we have a matrix of flows 𝑀 = [20 30] and a vector of final
40 50
production 𝑦 = [ 50 ]. Find: 1) a vector of total production, 2) a matrix of cost coefficients,
110
3) a vector of final production with a vector of total production 𝑌 = [ 50 ]
150
20
4) change in vector of total production with change in vector of final production △ 𝑦 = [ ]
−10

0.1 0.5
Example 2.10 In the input-output model we have a matrix of cost coefficients 𝐴 = [ ] and a vector of
0.3 0.5
150
final production 𝑦 = [ ]. Find a matrix of flows.
90
Example 2.11 In the input-output model we have a matrix of flows 𝑀 and a vector of final production 𝑦:
30 20 50 55 120 25
a) 𝑀 = [ ] , 𝑦 = [ ], b) 𝑀 = [ ] , 𝑦 = [ ].
30 40 30 60 30 70
Find a vector of total production and a matrix of cost coefficients.

 Beata Ciałowicz ~ 12 ~