Solving linear Systems

Z. ABBAS

LAU

February 1, 2024

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 1 / 36

Contents

1 2.3: Inverse of a matrix

Definition of inverse of a matrix
Cancelation Property
Finding the inverse of (n × n) matrix.

2 Solving a system of Linear Equation using Inverse of Matrices

3 LU factorization

4 Trace of a matrix

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 2 / 36

Question I
Let A, M and N be three (n × n) matrices. Show that if AM = In and
NA = In then M = N.

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 3 / 36

Question I
Let A, M and N be three (n × n) matrices. Show that if AM = In and
NA = In then M = N.

Answer: we have NA = In . Multiplying by M, we get:

NAM = (NA)M = In M = M and also NAM = N(AM) = NIn = N. Hence
the equality holds.

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 3 / 36

Inverse of a square matrix

Definition
A square matrix A of order n is said to be invertible/non-singular if there is
a matrix B of order n such that AB = BA = In .

Remark: It’s enough to show that AB = In or BA = In to conclude that A

is invertible and it’s inverse is B.

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 4 / 36

Question II

Let A be square matrix of order n . Suppose that A is invertible, show that:

1 A−1 invertible.
2 AT also written as tA is invertible.
3 If B also is an invertible square matrix of order n, then AB is
invertible.

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 5 / 36

Question II-Answer

1 since A is invertible, then there exists a square matrix of order n

denoted A−1 such that AA−1 = In , thus evidently A−1 is invertible
and (A−1 )−1 = A .
2 since A is invertible, we have AA−1 = In .
Then (AA−1 )T = (In )T = In .
Thus (A−1 )T AT = In , so AT is invertible and (AT )−1 = (A−1 )T .
3 B −1 A−1 (AB) = B −1 (A−1 A)B = B −1 (In B) = B −1 B = In . so Ab is
invertible and (AB)−1 = B −1 A−1

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 6 / 36

Properties of inverses

If A is an invertible matrix, k is a positive integer and c is a scalar not

equal to zero. Then A−1 , Ak , cA and AT are invertible and the following
statements are true:
a) (A−1 )−1 = A.
b) (AT )−1 = (A−1 )T .
c) (Ak )−1 = (A−1 )k .
d) (cA)−1 = c1 A−1 .

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 7 / 36

Question III

   
a b d −b
1 Calculate the product AB where A = and B =
c d −c a
2 What do you deduce?.

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 8 / 36

Answer

    
a b d −b ad − bc 0
1 AB = =
c d −c a 0 ad − bc
2 From part 1, we have AB = (ad − bc)In . If (ad − bc) ̸= 0, then
1 1 1
A( (ad−bc) B) = (ad−bc) AB = (ad−bc) (ad − bc)In = In . Thus, if
(ad-bc)̸= 0 , then A is invertible and
 −1  
a b 1 d −b
= (ad−bc)
c d −c a

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 9 / 36

Inverse of a (2 × 2) matrix

Strategy
 
a b
If A = is an 2 × 2 matrix
c d

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 10 / 36

Inverse of a (2 × 2) matrix

Strategy
 
a b
If A = is an 2 × 2 matrix such that ad − bc ̸= 0,
c d

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 10 / 36

Inverse of a (2 × 2) matrix

Strategy
 
a b
If A = is an 2 × 2 matrix such that ad − bc ̸= 0,then A is
c d
invertible and its inverse is given by
 
−1 1 d −b
A = .
ad − bc −c a

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 10 / 36

Example
 
2 4
a) A = .
1 −3

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 11 / 36

Example
 
2 4
a) A = . ad − bc = −10 ̸= 0
1 −3

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 11 / 36

Example
 
2 4
a) A = . ad − bc = −10 ̸= 0
1 −3

=⇒ A−1 =

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 11 / 36

Example
 
2 4
a) A = . ad − bc = −10 ̸= 0
1 −3
 
−1 1 −3 −4
=⇒ A = − =
10 −1 2

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 11 / 36

Example
 
2 4
a) A = . ad − bc = −10 ̸= 0
1 −3
  3 2

−1 1 −3 −4
=⇒ A = − = 10
1
5
1 .
10 −1 2 10 −5

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 11 / 36

Example
 
2 4
a) A = . ad − bc = −10 ̸= 0
1 −3
  3 2

−1 1 −3 −4
=⇒ A = − = 10
1
5
1 .
10 −1 2 10 −5

 
1 2
b) B = .
2 4

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 11 / 36

Example
 
2 4
a) A = . ad − bc = −10 ̸= 0
1 −3
  3 2

−1 1 −3 −4
=⇒ A = − = 10
1
5
1 .
10 −1 2 10 −5

 
1 2
b) B = . ad − bc = 0
2 4

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 11 / 36

Example
 
2 4
a) A = . ad − bc = −10 ̸= 0
1 −3
  3 2

−1 1 −3 −4
=⇒ A = − = 10
1
5
1 .
10 −1 2 10 −5

 
1 2
b) B = . ad − bc = 0 =⇒ B is a singular matrix.
2 4

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 11 / 36

Question IV

 
1 1
Compute A−2 in two different ways for A = .
2 4

Solution:

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 12 / 36

Question IV

 
1 1
Compute A−2
in two different ways for A = .
2 4
 9
− 54

−2 2 −1
Solution: A = (A ) = (A ) = −1 2 2
3 .
− 52 4

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 12 / 36

Question V

Find the inverse of AB with

   
1 3 3 1 2 3
A = 1 4 3 and B = 1 3 3.
1 3 4 2 4 3

using two different methods.

Solution:

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 13 / 36

Question V

Find the inverse of AB with

   
1 3 3 1 2 3
A = 1 4 3 and B = 1 3 3.
1 3 4 2 4 3

using two different methods.

   
7 −3 −3 1 −2 1
Solution: A−1 = −1 1 0  and B −1 = −1 1 0 .
2
−1 0 1 3 0 − 13
−1
method 1: (AB) = B A −1 −1

method 2: Compute −1
 AB then findits inverse which is (AB) .
8 −5 −2
Hence (AB)−1 = −8 4 3 
5 −2 − 37

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 13 / 36

1 2.3: Inverse of a matrix
Definition of inverse of a matrix
Cancelation Property
Finding the inverse of (n × n) matrix.

2 Solving a system of Linear Equation using Inverse of Matrices

3 LU factorization

4 Trace of a matrix

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 14 / 36

Theorem:
If C is an invertible matrix, then the following statements hold:
a) IfAC = BC =⇒ A = B.
b) If CA = CB =⇒ A = B.

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 15 / 36

Example
Solve the following matrix equation AX + 2B = B T where
   
1 3 3 1 2 3
A= 1 4 3
  and B= 1 3 3
1 3 4 2 4 3

Solution
X = A−1 (B T − 2B)
so      
7 −3 −3 1 1 2 1 2 3 !
X = −1 1 0 2 3 4 − 2 1 3 3 = · · ·
−1 0 1 3 3 3 2 4 3

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 16 / 36

Elementary matrices

definition
We call elementary matrix each of the following matrices:
1 In (Rt ↔ Rs ).
2 In (Rt → Rt + αRs ).
3 In (Rt → αRt ), where α ̸= 0.

Let A be a (n × m) matrix. The following hold:

1 A(Rt ↔ Rs ) = In (Rt ↔ Rs )A.
2 A(Rt → Rt + αRs ) = In (Rt → Rt + αRs )A.
3 A(Rt → αRt ) = In (Rt → αRt )A, where α ̸= 0.

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 17 / 36

Show that A(Rt ↔ Rs ) = In (Rt ↔ Rs )A.

Let 1 ≤ i ≤ n. Then:
1 if i=t:
Ri (In (Rt ↔ Rs )A) = Rt (In (Rt ↔ Rs )A) = Rt (In (Rt ↔ Rs ))A =
Rs (In )A = Rs (A).
2 if i=s:
Ri (A(Rt ↔ Rs )) = Rs (A(Rt ↔ Rs )) = Rs (In (Rt ↔ Rs )A) =
Rs (In (Rt ↔ Rs ))A = Rt (In )A = Rt (A).
3 if i ̸= s and i ̸= t:
Ri (A(Rt ↔ Rs )) = Ri (In (Rt ↔ Rs )A) = Ri (In (Rt ↔ Rs ))A =
Ri (In )A = Ri (A).
Hence In (Rt ↔ Rs )A = A(Rt ↔ Rs )

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 18 / 36

Inverses of elementary matrices

The inverse of
1 In (Rt ↔ Rs ) is In (Rt ↔ Rs ) .
2 for s ̸= t, In (Rt → Rt + αRs ) is In (Rt → Rt − αRs ).
3 In (Rt → αRt ), where α ̸= 0 is In (Rt → α−1 Rt ) .

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 19 / 36

Theorem
If A is a (n × n) invertible matrix over K, then A can be changed to In by
a finite sequence of elementary row operations.

Proof: By induction on n.
Theorem
Let A be a (n × n) matrix over K. Then A is invertible if and only if A can
be changed to In by a finite sequence of elementary row operations.

Proof: The N.C is the previous theorem. Here is the sufficient condition.

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 20 / 36

Question

Check if the following matrix is invertible.

Then A is not invertible.

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 21 / 36

Corollary
If A is a (n × n) invertible matrix over K, then A−1 is obtained from In by
carrying the same row operations after which A is changed to In .

Proof: Let E1 , E2 , · · · , En be the elementary matrices corresponding to the

elementary operations after which A is changed to In . Then

En En−1 · · · E1 A = In

. Then A−1 = En En−1 · · · E1 , thus A−1 is obtained from In by carrying the

same elementary operations after which A is changed to In .

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 22 / 36

J.G.R to find A−1
Let A be an n × n matrix.
To determine its inverse, we follow these steps:

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 23 / 36

J.G.R to find A−1
Let A be an n × n matrix.
To determine its inverse, we follow these steps:
1) Write the n × 2n matrix [A|In ].
2) Apply, to the whole matrix [A|In ], the row operations used in J.G.R
method to find the r.r.e.f of A.

Theorem
A (n × n) matrix A is invertible iff it’s row equivelant to In .

If the previous n × 2n matrix is transformed to [In |B]. The matrix A is

then invertible and B is the inverse of A, equivalently, B = A−1 .
remark. If A can’t be tranformed into In through elementary row
operations, then it’s not invertible.

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 23 / 36

Example (Find the inverse of the following matrices)
 
1 4 3
a) A = −1 −2 0.
2 2 3
Solution:

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 24 / 36

Example (Find the inverse of the following matrices)
 
1 4 3
a) A = −1 −2 0.
2 2 3
Solution:

[A|I3 ] =

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 24 / 36

Example (Find the inverse of the following matrices)
 
1 4 3
a) A = −1 −2 0.
2 2 3
Solution:
 
1 4 3 1 0 0
[A|I3 ] = −1 −2 0 0 1 0
2 2 3 0 0 1

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 24 / 36

Example (Find the inverse of the following matrices)
 
1 4 3
a) A = −1 −2 0.
2 2 3
Solution:
 
1 4 3 1 0 0
R + R1 −→ R2
[A|I3 ] = −1 −2 0 0 1 0 2 =⇒
R3 − 2R1 −→ R3
2 2 3 0 0 1
 
1 4 3 1 0 0
0 2 3 1 1 0 R3 + 3R2 −→ R3 =⇒
0 −6 −3 −2 0 1
 
1 4 3 1 0 0
R − 1 R −→ R2
0 2 3 1 1 0 2 21 3 =⇒
R1 − 2 R3 −→ R1
0 0 6 1 3 1

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 24 / 36

Example
1
− 23 − 12
 
1 4 0 2
 
1
− 21 − 12 
 
0
 2 0 2  R1 − 2R2 −→ R1 =⇒
 
0 0 6 1 3 1
0 0 − 21 − 12 1
 
1 2
1
 
1 R2 × −→ R2
− 12 − 12 
 
0 2 0 2 =⇒
2 1


 R3
 × 6 −→ R3
0 0 6 1 3 1

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 25 / 36

Example
1 0 0 − 12 − 21 1
 
2
 
0 1 0 1 − 41 − 14 
 
 4  = [I3 |B].
 
0 0 1 16 1
2
1
6

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 26 / 36

Example
1 0 0 − 12 − 21 1 

2
 
1 1 1
 
0 1 0
 4 −4 −4  = [I3 |B].
 
1 1 1
0 0 1 6 2 6
 1 1 1 
  −2 −2 2  
1 4 3  1
 1 0 0
1 1

AB = −1 −2 0  4 − 4 − 4  = 0 1 0 = I3
 
2 2 3   0 0 1
1 1 1
6 2 6
and

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 26 / 36

Example

BA =

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 27 / 36

Example
 1
− 12 1 
−2 2    
  1 4 3 1 0 0
 1
− 41 − 14 

BA =   −1 −2 0 = 0 1 0 = I3 .
   
 4
  2 2 3 0 0 1
1 1 1
6 2 6

Thus, B = A−1 .

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 27 / 36

Example
 1
− 12 1 
−2 2    
  1 4 3 1 0 0
 1
− 41 − 14 

BA =   −1 −2 0 = 0 1 0 = I3 .
   
 4
  2 2 3 0 0 1
1 1 1
6 2 6

Thus, B = A−1 .
 
1 −1 0
b) A =  1 0 −1.
−6 2 3
Solution: Homework.

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 27 / 36

1 2.3: Inverse of a matrix
Definition of inverse of a matrix
Cancelation Property
Finding the inverse of (n × n) matrix.

2 Solving a system of Linear Equation using Inverse of Matrices

3 LU factorization

4 Trace of a matrix

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 28 / 36

Theorem
The system AX = b of n linear equations in n unknowns has a unique
solution if and only if A is Invertible. In this case, the solution X is given
by X = A−1 b.

Note that at this point we have shown that the following statements are
equivalent for an (n × n) matrix A:
1 A is nonsingular.
2 Ax = 0 has a unique solution-the trivial solution.
3 A is row (column) equivalent to In ( the r.r.e.f of A is In ).
4 The linear system Ax = b has a unique solution for every (n × n)
matrix b.
5 A is a product of elementary matrices.

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 29 / 36

Example (Solve the system using the inverse of Matrices)
a) 
 x1 +4x2 +3x3 = 12,
(S) −x1 −2x2 = −12,
2x1 +2x2 +3x3 = 8.


Solution:

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 30 / 36

Example (Solve the system using the inverse of Matrices)
a) 
 x1 +4x2 +3x3 = 12,
(S) −x1 −2x2 = −12,
2x1 +2x2 +3x3 = 8.


Solution:System (S) can be written as AX = b with

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 30 / 36

Example (Solve the system using the inverse of Matrices)
a) 
 x1 +4x2 +3x3 = 12,
(S) −x1 −2x2 = −12,
2x1 +2x2 +3x3 = 8.


Solution:System
 (S)
 can be written as AX = b with
1 4 3
A = −1 −2 0,
2 2 3

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 30 / 36

Example (Solve the system using the inverse of Matrices)
a) 
 x1 +4x2 +3x3 = 12,
(S) −x1 −2x2 = −12,
2x1 +2x2 +3x3 = 8.


Solution:System
 (S)
 can be written as AX = b with
1 4 3 x1
A = −1 −2 0,X = x2 
2 2 3 x3

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 30 / 36

Example (Solve the system using the inverse of Matrices)
a) 
 x1 +4x2 +3x3 = 12,
(S) −x1 −2x2 = −12,
2x1 +2x2 +3x3 = 8.


Solution:System
 (S)
 can be written as AX
 = b with
1 4 3 x1 12
A = −1 −2 0,X = x2  and b = −12.
2 2 3 x3 8

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 30 / 36

Example (Solve the system using the inverse of Matrices)
In the previous example, we have find that the inverse of A is given by
 1
− 2 − 21 1 
2
 
A−1 =  1 1 1
 
 4 − 4 − 4 .
 
1 1 1
6 2 6

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 31 / 36

Example (Solve the system using the inverse of Matrices)
In the previous example, we have find that the inverse of A is given by
 1
− 2 − 21 1 
2
 
A−1 =  1 1 1
 
 4 − 4 − 4 .
 
1 1 1
6 2 6

Then, system (S) has a unique solution given by

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 31 / 36

Example (Solve the system using the inverse of Matrices)
In the previous example, we have find that the inverse of A is given by
 1
− 2 − 21 1 
2
 
A−1 =  1 1 1
 
 4 − 4 − 4 .
 
1 1 1
6 2 6

Then, system (S) has a unique solution given by

X =

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 31 / 36

Example (Solve the system using the inverse of Matrices)
In the previous example, we have find that the inverse of A is given by
 1
− 2 − 21 1 
2
 
A−1 =  1 1 1
 
 4 − 4 − 4 .
 
1 1 1
6 2 6

Then, system (S) has a unique solution given by

 1
− 2 − 21 1 
2    
  12 4
X = A−1 b =  1 1 1 
 
 4 − 4 − 4 −12 =  4 .
  8 − 38
1 1 1
6 2 6

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 31 / 36

Example (Solve the system using the inverse of Matrices)
Hence, x1 = 4, x2 = 4 and x3 = − 83 is the solution of system (S).

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 32 / 36

Introduction

Our aim is to write a matrix A = [aij ] ∈ Mn (C) as follows:

A = LU,

where L = [lij ] ∈ Mn (C) is a Lower Triangular matrix and

U = [uij ] ∈ Mn (C) is an Upper Triangular one.
Equivalently lij = 0 for all i < j and uij = 0 for all i > j. We will assume in
addition, that lii = 1.
Thus  
1 0 0
L = l21 1 0 ,
l31 l32 1

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 33 / 36

Introduction

Our aim is to write a matrix A = [aij ] ∈ Mn (C) as follows:

A = LU,

where L = [lij ] ∈ Mn (C) is a Lower Triangular matrix and

U = [uij ] ∈ Mn (C) is an Upper Triangular one.
Equivalently lij = 0 for all i < j and uij = 0 for all i > j. We will assume in
addition, that lii = 1.
Thus    
1 0 0 u11 u12 u13
L = l21 1 0 , U =  0 u22 u23  .
l31 l32 1 0 0 u33

By applying the matrix multiplication between L and U and by comparing

with A, we can find the entries of the matrices L and U.

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 33 / 36

Solve using the LU decomposition the following statement

 x1 +4x2 +3x3 = 12,
(S) −x1 −2x2 = −12,
2x1 +2x2 +3x3 = 8.


Z. ABBAS (LAU) Solving linear Systems February 1, 2024 34 / 36

Exs

Example
Find the LU factorization of the following matrices:
 
1 2 4
1- A = 3 8 14.
2 6 13
 
3 1 6
2- B = −6 0 −16.
0 8 −17

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 35 / 36

Definition
Let A = [aij ] ∈ Mn (C). The trace function tr defined by
tr : Mn (C) −→ C, is given by
n
X
tr (A) = aii .
i=1

Let A, B ∈ Mn (C) and let α and β be two scalars. Then:

1- tr (αA + βB) = αtr (A) + βtr (B).

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 36 / 36

Definition
Let A = [aij ] ∈ Mn (C). The trace function tr defined by
tr : Mn (C) −→ C, is given by
n
X
tr (A) = aii .
i=1

Let A, B ∈ Mn (C) and let α and β be two scalars. Then:

1- tr (αA + βB) = αtr (A) + βtr (B).
2- tr (AT ) = tr (A).
3- tr (AB) = tr (BA).

Z. ABBAS (LAU) Solving linear Systems February 1, 2024 36 / 36