1

Chapter I
Systems of Linear Equations and Matrices

Section 1.1:Introduction to Systems of Linear Equations

A linear equation in the 𝑛 unknowns (or variables) 𝑥1 , 𝑥2 , … , 𝑥𝑛 is an equation
of the form𝑎1 𝑥1 + 𝑎2 𝑥2 + ⋯ + 𝑎𝑛 𝑥𝑛 = 𝑏,
where 𝑎1 , 𝑎2 , … , 𝑎𝑛 and 𝑏 are real numbers (constants).
A solution of a linear equation 𝑎1 𝑥1 + 𝑎2 𝑥2 + ⋯ + 𝑎𝑛 𝑥𝑛 = 𝑏 is a sequence of
real numbers 𝑠1 , 𝑠2 , … , 𝑠𝑛 such that the equation is satisfied when we substitute
𝑥1 = 𝑠1 , 𝑥2 = 𝑠2 , … , 𝑥𝑛 = 𝑠𝑛 . The set of all solutions of the equation is
calledits solution set (or the general solution of the equation).

Example 1: Find the solution set of each of the following equations.

(i) 2𝑥 − 3𝑦 = 4 (ii) 𝑥1 − 3𝑥2 + 5𝑥3 = 2
Solution:
2𝑡−4
(i) Let 𝑥 = 𝑡 be an arbitrary real number. Then 𝑦 = 3 .
2 4
Thus, 𝑥 = 𝑡, 𝑦 = 3 𝑡 − 3 (𝑡 ∈ ℝ) gives the solution set of (i).
2
In particular, if 𝑡 = 1, then 𝑥 = 1, 𝑦 = − is a solution of (i).
3
3𝑡+4 3𝑡+4
Similarly, if we let 𝑦 = 𝑡, then 𝑥 = 2 . Thus 𝑥 = 2 , 𝑦 = 𝑡(𝑡 ∈ ℝ)also
gives the solution set of (i).
In particular, if 𝑡 = 2, then 𝑥 = 5, 𝑦 = 2 is a solution of (i).
(ii) Assume that 𝑥2 = 𝑟, 𝑥3 = 𝑠 are arbitrary real numbers. Then 𝑥1 = 2 + 3𝑟 − 5𝑠.
Thus the solution set is 𝑥1 = 2 + 3𝑟 − 5𝑠, 𝑥2 = 𝑟, 𝑥3 = 𝑠, where 𝑟, 𝑠 ∈ ℝ are
arbitrary.

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

𝑎11 𝑥1 + 𝑎12 𝑥2 + ⋯ + 𝑎1𝑛 𝑥𝑛 = 𝑏1
𝑎21 𝑥1 + 𝑎22 𝑥2 + ⋯ + 𝑎2𝑛 𝑥2 = 𝑏2
]……….(*)
⋮ ⋮ ⋮ ⋮
𝑎𝑚1 𝑥1 + 𝑎𝑚2 𝑥2 + ⋯ + 𝑎𝑚𝑛 𝑥𝑛 = 𝑏𝑚
where 𝑥1 , 𝑥2 , … , 𝑥𝑛 are the unknowns and the subscripted 𝑎′𝑠 and 𝑏 ′ 𝑠 denote
constants.
 2

Aline in the 𝑥𝑦-plane can be algebraically represented by an equation:

𝑎1 𝑥 + 𝑎2 𝑦 = 𝑏 (𝑎1 , 𝑎2 not both zero)

ℓ1 ℓ2 ℓ1
ℓ1 ℓ2
ℓ2

No solution One solution Infinitely many

ℓ1 and ℓ2 are parallel ℓ1 and ℓ2 intersect at solutions
one point only ℓ1 and ℓ2 coincide
A sequence of real 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 (*).
A system is inconsistent if it has no solutions. If it has at least one solution, it is
called consistent.

𝑥+𝑦=2 𝑥+𝑦=2 𝑥+𝑦=2

(2,0) 𝑥+𝑦=1
𝑥−𝑦=2

The system: The system: The system:

𝑥+𝑦 =2 𝑥+𝑦 =1 𝑥+𝑦 =2
𝑥−𝑦 =2 𝑥+𝑦 =2 −𝑥 − 𝑦 = −2
has exactly one solution has no solution has infinitely many
(unique solution: namely (inconsistent) solutions
𝑥 = 2, 𝑦 = 0)
𝑎11 𝑎12 ⋯ 𝑎1𝑛 𝑏1
𝑎 𝑎22 ⋯ 𝑎2𝑛 𝑏2
The matrix [ 21 ] is called the augmented matrix for the
⋮ ⋮ ⋮ ⋮ ⋮
𝑎𝑚1 𝑎𝑚2 ⋯ 𝑎𝑚𝑛 𝑏𝑚
system

𝑎11 𝑥1 + 𝑎12 𝑥2 + ⋯+ 𝑎1𝑛 𝑥𝑛 = 𝑏1

𝑎21 𝑥1 + 𝑎22 𝑥2 + ⋯+ 𝑎2𝑛 𝑥2 = 𝑏2
]……….(*)
⋮ ⋮ ⋮ ⋮
𝑎𝑚1 𝑥1 + 𝑎𝑚2 𝑥2 + ⋯+ 𝑎𝑚𝑛 𝑥𝑛 = 𝑏𝑚

❖ Thus, the augmented matrix for the system

 3

𝑥1 − 𝑥2 + 2𝑥3 = 1 1 −1 2 1
−𝑥1 + 4𝑥2 − 3𝑥3 = 0 is [−1 4 −3 0]
2𝑥1 + 7𝑥2 − 𝑥3 = −3 2 7 −1 −3
❖ The augmented matrix for the system
𝑥 − 2𝑦 − 2𝑧 = 1 1 −2 −2 1
2𝑥 + 3𝑧 = 0 is [ 2 0 3 0]
−𝑥 + 𝑦 − 2𝑧 = −1 −1 1 −2 −1
In solving a system of linear equations we usually transform the given system
into a new (equivalent) system that has the same solution set but it is easier to
solve. This can be done in a series of steps (elimination of unknowns) as follows.
1. Multiply any equation through by a nonzero constant
2. Interchange any two equations
3. Add a multiple of one equation to another equation
In the associated augmented matrix these operations (called elementary row
operations) are as follows.
1. Multiply any row through by a nonzero constant
2. Interchange any two rows
3. Add a multiple of one row to another row
Example 2: Consider the system of linear equations
𝑥 + 𝑦 + 2𝑧 = 9
2𝑥 + 4𝑦 − 3𝑧 = 1
3𝑥 + 6𝑦 − 5𝑧 = 0
1 1 2 9
Solution:The augmented matrix is [2 4 −3 1]:
3 6 −5 0
① 1 2 9 −2R1 + R2 ① 1 2 9 1R ① 1 2 9
2 7 17
[2 4 −3 1] −3R1 + R3[ 0 2 −7 −17] 2 [ 0 ① − −2]
−27 ⃗⃗ 0
2
3 6 −5 0 ⃗⃗ 0 3 −11 3 −11 −27
11 35 11 35
−R 2 + R1
① 0 2 2 ① 0 2 2
17 −2R 3
−3R 2 + R 3 0 ① − 72 − 2 ⃗⃗⃗ 0 ① − 72 − 17
⃗⃗⃗ 2
[ 0 0 − 12 − 32 ]
0 ① [ 0
3]
11
− 2 R 3 + R1 ① 0 0 1 𝑥 = 1
7
R + R2
2 3
0 ① 0 2 with equivalent system 𝑦 = 2
⃗⃗⃗ [ 0 0 ① 3] 𝑧 = 3

So the system has the unique solution 𝑥 = 1, 𝑦 = 2, 𝑧 = 3.

Section 1.2: Gaussian Elimination
 4

A matrix is said to be in reduced row-echelon form (r.r.e.f) if:

1. The first nonzero entry in each nonzero row is 1, called the leading 1 of the
row.
2. The zero rows (if any) are grouped together at the bottom of the matrix.
3. In any two successive nonzero rows, the leading 1 in the lower row occurs
farther to the right than the leading 1 in the higher row.
4. Each column that contains a leading 1 has zeros everywhere else.

If only properties 1, 2, 3 are satisfied, then the matrix is said to be in row-echelon

form (r.e.f).

Example 1:
0 1 −4 0 3
1 0 0 3 1 0 0
0 0 0 1 1 0 0
(i) The matrices:[0 1 0 −1] , [0 1 0] , [ ],[ ]
0 0 0 0 0 0 0
0 0 1 −2 0 0 1
0 0 0 0 0
are in r.r.e.f.
(ii) The matrices
1 2 4 5 1 1 0 0 1 6 3 0
[0 1 −1 2] , [0 1 0] , [0 0 1 −1 0] are inr.e.f.
0 0 1 3 0 0 0 0 0 0 0 1

Example 2: Suppose that the augmented matrix for a system of linear equations
has been reduced by row operations to the given r.r.e.f. Solve the system.
1 0 0 2 1 0 0 4 −1
(a) [0 1 0 −3] (b) [0 1 0 3 7]
0 0 1 4 0 0 1 2 5
1 6 0 0 1 −2
1 0 0 0
0 0 1 0 2 1
(c) [ ] (d) [0 1 −1 0]
0 0 0 1 3 4
0 0 0 1
0 0 0 0 0 0

Solution:
(a) The corresponding system of equations is
𝑥1 = 2
𝑥2 = −3
𝑥3 = 4
Thus, the system has only one (unique) solution 𝑥1 = 2, 𝑥2 = −3, 𝑥3 = 4.

(b) The corresponding system of equations is

 5

𝑥1 + 4𝑥4 = −1
𝑥2 + 3𝑥4 = 7
𝑥3 + 2𝑥4 = 5

𝑥1 , 𝑥2 , 𝑥3 are leading variables.

𝑥4 is a nonleading variable (a free variable).
Let 𝑥4 = 𝑡 be arbitrary. Then the system has infinitely many solutions of the
form
𝑥1 = −1 − 4𝑡, 𝑥2 = 7 − 3𝑡, 𝑥3 = 5 − 2𝑡, 𝑥4 = 𝑡 (𝑡 ∈ ℝ).
(c) The corresponding system is
𝑥1 + 6𝑥2 + 5𝑥5 = −2
𝑥3 + 2𝑥5 = 1
𝑥4 + 3𝑥5 = 4
The leading variables are 𝑥1 , 𝑥3 , 𝑥4
The free variables are 𝑥2 , 𝑥5 .
Let 𝑥2 = 𝑟, 𝑥5 = 𝑠 be arbitrary.
Then 𝑥1 = −2 − 6𝑟 − 5𝑠, 𝑥2 = 𝑟, 𝑥3 = 1 − 2𝑠, 𝑥4 = 4 − 3𝑠, 𝑥5 = 𝑠 is the
general solution of the system.

(d) The last equation of the corresponding system is 0𝑥1 + 0𝑥2 + 0𝑥3 = 1.
This equation cannot be satisfied(no solution). So, the system has no solutions
(inconsistent).
1 −1 0 −2 3
2 −2 2 −2 4
Example 3:Reduce the matrix [ ] to a r.r.e.f.
0 0 1 1 −1
1 −1 2 0 1
Solution:
1 −1 0 −2 3 1 −1 0 −2 3
2 −2 2 −2 4 −2R1 + R2 0 0 2 2 −2
[ ] [ ]
0 0 1 1 −1 ⃗⃗⃗ 0 0 1 1 −1
1 −1 2 0 1 1 −1 2 0 1
1 −1 0 −2 3 1 −1 0 −2 3
−R1 + R 4 0 0 2 2 −2 −R 2 + R 4 0 0 2 2 −2
[ ] [ ]
⃗⃗⃗ 0 0 1 1 −1 ⃗⃗⃗ 0 0 1 1 −1
0 0 2 2 −2 0 0 0 0 0
1 1 −1 0 −2 3 1 −1 0 −2 3
R
2 2 [0 0 1 1 −1] −R 2 + R 3 [0 0 1 1 −1], which is in r.r.e.f.
⃗⃗⃗ 0 0 1 1 −1 ⃗⃗⃗ 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0
The involved matrices are row equivalent.
 6

The procedure of reducing a matrix (a system) to r.r.e.f. (its corresponding

system) is called Gauss-Jordan elimination. It is called Gaussian elimination when
the matrix is it transformed into a r.e.f.
Example 4: Solve by Gauss-Jordan elimination.
𝑥1 + 3𝑥2 − 2𝑥3 + 2𝑥5 = 0
2𝑥1 + 6𝑥2 − 5𝑥3 − 2𝑥4 + 4𝑥5 − 3𝑥6 = −1
5𝑥3 + 10𝑥4 + 15𝑥6 = 5
2𝑥1 + 6𝑥2 + 8𝑥4 + 4𝑥5 + 18𝑥6 = 6
Solution: The augmented matrix is
1 3 −2 0 2 0 0 −2R + R 1 3 −2 0 2 0 0
1 2
2 6 −5 −2 4 −3 −1 −2R + R 0 0 −1 −2 0 −3 −1
[ ] 1 4[ ]
0 0 5 10 0 15 5 0 0 5 10 0 15 5
⃗⃗
2 6 0 8 4 18 6 0 0 4 8 0 18 6
5R 2 + R 3 1 3 −2 0 2 0 0
4R 2 + R 4 [0 0 −1 −2 0 −3 −1]
0 0 0 0 0 0 0
⃗⃗⃗
0 0 0 0 0 6 2
1 3 −2 0 2 0 0
1 3 −2 0 2 0 0 1
−R 2 0 0 1 2 0 3 1
0 0 1 2 0 3 1 R3 1
[ ] 6
⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗
R3 ↔ R4 0 0 0 0 0 6 2
⃗⃗⃗ 0 0 0 0 0 1
3
0 0 0 0 0 0 0 [0 0 0 0 0 0 0]
1 3 0 4 2 6 2 1 3 0 4 2 0 0
2R 2 + R1 0 0 1 2 0 3 1 −3R 3 + R 2 0 0 1 2 0 0 0
1 −6R 3 + R1 1.
⃗⃗⃗ 0 0 0 0 0 1 ⃗⃗⃗ 0 0 0 0 0 1
3 3
[0 0 0 0 0 0 0] [0 0 0 0 0 0 0]
The corresponding system is
𝑥1 + 3𝑥2 + 4𝑥4 + 2𝑥5 = 0
𝑥3 + 2𝑥4 = 0
1
𝑥6 =
3
Thus,
𝑥1
= −3𝑥2 − 4𝑥4 − 2𝑥5
𝑥3
= −2𝑥4
1
𝑥6 =
3
Let 𝑥2 = 𝑟, 𝑥4 = 𝑠, 𝑥5 = 𝑡 be arbitrary. Then the solution is given by:
1
𝑥1 = −3𝑟 − 4𝑠 − 2𝑡, 𝑥2 = 𝑟, 𝑥3 = −2𝑠, 𝑥4 = 𝑠, 𝑥5 = 𝑡, 𝑥6 = .
3
 7

Example 5: Solve the system

𝑥 + 𝑦 + 2𝑧 = 9
2𝑥 + 4𝑦 − 3𝑧 = 1
3𝑥 + 6𝑦 − 5𝑧 = 0
by Gaussian elimination and back-substitution.
1 1 2 9
Solution: The augmented matrix [2 4 −3 1]can be reduced to the r.e.f.
3 6 −5 0
1 1 2 9 𝑥 + 𝑦 + 2𝑧 = 9
7 17 7 17
[0 1 − 2 − 2 ]. This matrix has the system 𝑦 − 2𝑧 = − 2
0 0 1 3 𝑧 = 3
Solving the system for the leading variables:
𝑥 = 9 − 𝑦 − 2𝑧
17 7
𝑦=− + 𝑧
2 2
𝑧=3
Substituting 𝑧 = 3 in the above equations we get:
𝑥 = 3−𝑦 𝑥=1
𝑦=2 Substituting the 2 in the 1 we have𝑦 = 2
nd st

𝑧=3 𝑧=3
as we have obtained before.
Example 6: Solve the system by Gauss-Jordan elimination.
𝑥 − 𝑦 − 2𝑤 = 3
2𝑥 − 2𝑦 + 2𝑧 − 2𝑤 = 4
𝑧 + 𝑤 = −1
𝑥 − 𝑦 + 2𝑧 = 1
Solution: The augmented matrix
1 −1 0 −2 3 1 −1 0 −2 3
2 −2 2 −2 4 0 0 1 1 −1
[ ] has the r.r.e.f. [ ] which has the system:
0 0 1 1 −1 0 0 0 0 0
1 −1 2 0 1 0 0 0 0 0
𝑥 − 𝑦 − 2𝑤 = 3 𝑥 = 3 + 𝑦 + 2𝑤
or
𝑧 + 𝑤 = −1 𝑧 = −1 − 𝑤

Let 𝑦 = 𝑟, 𝑤 = 𝑠 be arbitrary. The solution is given by:

𝑥 = 3 + 𝑟 + 2𝑠, 𝑦 = 𝑟, 𝑧 = −1 − 𝑠, 𝑤 = 𝑠.
 8

Homogeneous Systems of Linear Equations

A system of linear equations is called homogeneous if it is of the form

𝑎11 𝑥1 + 𝑎12 𝑥2 + ⋯ + 𝑎1𝑛 𝑥𝑛 = 0
𝑎21 𝑥1 + 𝑎22 𝑥2 + ⋯ + 𝑎2𝑛 𝑥2 = 0
i.e., if𝑏1 = 𝑏2 = ⋯ = 𝑏𝑚 = 0.
⋮ ⋮ ⋮ ⋮
𝑎𝑚1 𝑥1 + 𝑎𝑚2 𝑥2 + ⋯ + 𝑎𝑚𝑛 𝑥𝑛 = 0

Note that every homogeneous system is consistent, since𝑥1 = 0, 𝑥2 = 0, … , 𝑥𝑛 = 0

is always a solution. This solution is called the trivial solution. If there are other
solutions, they are called nontrivial solutions.
Every homogeneous system has exactly one of the following cases:
1. The system has only the trivial solution.
2. The system has infinitely many nontrivial solutions in addition to the trivial
one.

Example 7:Solve the homogeneous system by Gauss-Jordan elimination.

2𝑥1 + 3𝑥2 − 3𝑥3 + 7𝑥4 + 𝑥5 + 4𝑥6 = 0
𝑥2 − 𝑥3 + 3𝑥4 − 𝑥5 + 2𝑥6 = 0
2𝑥1 + 2𝑥2 − 𝑥3 + 6𝑥4 + 2𝑥5 − 𝑥6 = 0
2𝑥1 + 𝑥2 − 𝑥3 + 𝑥4 + 3𝑥5 =0

Solution: The augmented matrix for the system is

2 3 −3 7 1 4 0
0 1 −1 3 −1 2 0
[ ] which has the r.r.e.f.
2 2 −1 6 2 −1 0
2 1 −1 1 1 0 0

1 0 0 −1 2 −1 0
0 1 0 5 −1 −1 0
[ ]. The corresponding system is
0 0 1 2 0 −3 0
0 0 0 0 0 0 0

𝑥1 − 𝑥4 +2𝑥5 − 𝑥6 =0
𝑥2 + 5𝑥4 − 𝑥5 − 𝑥6 =0
𝑥3 + 2𝑥4 − 3𝑥6 =0

Solving for the leading variables:

 9

𝑥1 = 𝑥4 − 2𝑥5 + 𝑥6
𝑥2 = −5𝑥4 + 𝑥5 + 𝑥6 }. Let 𝑥4 = 𝑟, 𝑥5 = 𝑠, 𝑥6 = 𝑡 be arbitrary. Then
𝑥3 = −2𝑥4 + 3𝑥6

𝑥1 = 𝑟 − 2𝑠 + 𝑡, 𝑥2 = −5𝑟 + 𝑠 + 𝑡, 𝑥3 = −2𝑟 + 3𝑡, 𝑥4 = 𝑟, 𝑥5 = 𝑠, 𝑥6 = 𝑡is the

general solution of the given system. This system has nontrivial solutions (the
trivial solution is obtained when 𝑟 = 0, 𝑠 = 0, 𝑡 = 0).

Theorem 8: A homogeneous system of linear equations with

moreunknowns than equations has infinitely many solutions.

Note 9: If the given homogeneous system has 𝑚 equations, 𝑛 unknowns and

𝑚 < 𝑛 and if 𝑟 is the number of nonzero rows in ther.r.e.f. of the augmented
matrix, then 𝑟 < 𝑛. It follows that the corresponding system is of the form:
⋯ 𝑥𝑘1 + ∑( ) =0
⋯ 𝑥𝑘2 + ∑( ) =0
⋱ =0
𝑥𝑘𝑟 + ∑ ( ) = 0

where 𝑥𝑘1 , 𝑥𝑘2 , … , 𝑥𝑘𝑟 are the leading variables and ∑ ( )denotes sums
involving the 𝑛 − 𝑟 remaining variables. Thus, the system has infinitely many
solutions in which 𝑛 − 𝑟 of the unknowns are arbitrary.

Example 10: Solve the homogeneous system

𝑥 + 𝑦 + 2𝑧 = 0
2𝑥 + 4𝑦 − 3𝑧 = 0
3𝑥 + 6𝑦 − 5𝑧 = 0
Solution: The augmented matrix:
1 1 2 0 1 0 0 0
[2 4 −3 0] [0 1 0 0]. Thus, the system has only the trivial
⃗⃗⃗
3 6 −5 0 0 0 1 0
solution𝑥 = 0, 𝑦 = 0, 𝑧 = 0.
 10

Example 11: For which value(s) of 𝜆 does the system

(𝜆 − 1)𝑥 + 𝑦 = 0
−4𝑥 + (𝜆 + 3)𝑦 = 0

have nontrivial solutions?

Solution:
1 𝜆−1 1 0 R ↔R 𝜆+3
𝜆−1 1 0 − R2 𝜆+3 1 2 1 − 0]
[ ] 4 [ ] [ 4
−4 𝜆 + 3 0 1 − 0 ⃗⃗
⃗⃗ 4 𝜆−1 1 0
𝜆+3 𝜆+3
−(𝜆 − 1)R1 + R 2 1 − 0 1 − 0
4 = 4
⃗⃗⃗ (𝜆 − 1)(𝜆 + 3) 4 + (𝜆 − 1)(𝜆 + 3)
0 1+ 0] [0 0]
[ 4 4
4+(𝜆−1)(𝜆+3)
Thus, the system has nontrivial solutions iff = 0 iff 𝜆2 + 2𝜆 + 1 = 0
4
iff (𝜆 + 1)2 = 0 iff 𝜆 = −1.
 11

Section 1.3:Matrices and Matrix Operations

Definition 1: A matrix is a rectangular array of numbers. The numbers in the array

are called the entries in the matrix.

2 −1 𝜋 √𝜋 𝑒 1
[0 4 ] , [1 1 −2 0 5], [0 −𝑒 1 ] , [2], and [5] are matrices.
5 −2 0 0 0.5

If 𝐴 is a matrix with 𝑚 rows (horizontal lines) and 𝑛 columns (vertical lines), then
the size of 𝐴 is 𝑚 × 𝑛.

❖ If 𝐴 is a matrix with 𝑚 rows and 𝑛 columns, then 𝐴 is said to be an

𝑚 × 𝑛 matrix. We will use 𝑎𝑖𝑗 to denote the entry that occurs in row 𝑖 and
column 𝑗 of 𝐴. A general 𝑚 × 𝑛 matrix can be written as
𝑎11 𝑎21 ⋯ 𝑎1𝑛
𝑎21 𝑎22 ⋯ 𝑎2𝑛
𝐴=[ ⋮ ⋮ ⋱ ⋮ ] or 𝐴 = [𝑎𝑖𝑗 ]𝑚×𝑛, 𝑎𝑖𝑗 is called the (𝑖, 𝑗) entry of 𝐴.
𝑎𝑚1 𝑎𝑚2 ⋯ 𝑎𝑚𝑛

❖ A matrix with 𝑛 rows and 𝑛 columns is called a square matrix of order

𝑛. The entries 𝑎11 , 𝑎22 , … , 𝑎𝑛𝑛 form the main diagonal of 𝐴.
❖ Two 𝑚 × 𝑛 matrices 𝐴 = [𝑎𝑖𝑗 ] and 𝐵 = [𝑏𝑖𝑗 ] are said to be equal (written
𝐴 = 𝐵) iff 𝑎𝑖𝑗 = 𝑏𝑖𝑗 for 1 ≤ 𝑖 ≤ 𝑚, 1 ≤ 𝑗 ≤ 𝑛.

Definition 2: Let𝐴 = [𝑎𝑖𝑗 ]𝑚×𝑛 , 𝐵 = [𝑏𝑖𝑗 ]𝑚×𝑛 and let 𝑐 be a scalar.

1. The sum𝐴 + 𝐵 is the matrix 𝐴 + 𝐵 = [𝑎𝑖𝑗 + 𝑏𝑖𝑗 ]𝑚×𝑛.
2. The product 𝑐𝐴 (the scalar multiple of 𝐴 by 𝑐) is given by 𝑐𝐴 = [𝑐𝑎𝑖𝑗 ]𝑚×𝑛 .
1 −2 3 4 3 −3 1 −1
1 2
Example 3: If 𝐴 = [0 −2 1 5] , 𝐵 = [1 2 0 1 ], and 𝐶 = [ ], then
1 3
1 3 7 2 2 −1 8 3
2 −5 4 3
𝐴 + 𝐵 = [1 0 1 6], 𝐴 + 𝐶, 𝐵 + 𝐶 are undefined.
3 2 15 5
 12

1 3 1 3 9 3
Example 4: If 𝐴 = [−1 5 2], then 3𝐴 = [−3 15 6] and
6 7 3 18 21 9
−1 −3 −1
(−1)𝐴 = [ 1 −5 −2].
−6 −7 −3

Note that if 𝐵 is a matrix, then – 𝐵denotes(−1)𝐵 . Note also that the

difference 𝐴 − 𝐵 is 𝐴 − 𝐵 = 𝐴 + (−1)𝐵.
2 3 1 3 1 0
Example 5: If 𝐴 = [−1 4], 𝐵 = [2 2 ], then 𝐴 − 𝐵 = [−3 2].
0 5 3 −1 −3 6

Definition 6: If 𝐴 = [𝑎𝑖𝑗 ]𝑚×𝑟 is an 𝑚 × 𝑟 matrix, 𝐵 = [𝑏𝑖𝑗 ]𝑟×𝑛 is an 𝑟 ×

𝑛 matrix, then the product 𝐴𝐵 is the 𝑚 × 𝑛 matrix given by𝐴𝐵 = [𝑐𝑖𝑗 ]𝑚×𝑛 , where
𝑐𝑖𝑗 = 𝑎𝑖1 𝑏1𝑗 + 𝑎𝑖2 𝑏2𝑗 + ⋯ + 𝑎𝑖𝑟 𝑏𝑟𝑗 .
𝑎11 𝑎12 ⋯ 𝑎1𝑟
𝑎21 𝑎22 ⋯ 𝑎2𝑟 𝑏11 𝑏12 ⋯ 𝑏1𝑗 ⋯ 𝑏1𝑛
⋮ ⋮ ⋮ ⋮ 𝑏21 𝑏22 ⋯ 𝑏2𝑗 ⋯ 𝑏2𝑛
𝐴𝐵 = 𝑎𝑖1 𝑎𝑖2 ⋯ 𝑎𝑖1 = [𝑐𝑖𝑗 ]𝑚×𝑛 .
⋮ ⋮ ⋮ ⋮ ⋮ ⋮
⋮ ⋮ ⋮ ⋮
[𝑏𝑟1 𝑏𝑟2 ⋯ 𝑏𝑟𝑗 ⋯ 𝑏𝑟𝑛 ]
𝑟×𝑛
[𝑎𝑚1 𝑎𝑚2 ⋯ 𝑎𝑚𝑟 ]𝑚×𝑟
(number of columns 𝐴 = number of rows of 𝐵)
1 −1
2 3 −1 1 −1
Example 7: Let 𝐴 = [ ] , 𝐵 = [−1 0] , 𝐶 = [ ] . Then
−1 0 2 2×3 −3 2 2×2
1 2 3×2
1 −1
2 3 −1 −2 −4
❖ 𝐴𝐵 = [ ] [−1 0] =[ ] .
−1 0 2 2×3 1 5 2×2
1 2 3×2
1 −1 4 −3
1 −1
❖ 𝐵𝐶 = [−1 0] [ ] = [−1 1] .
−3 2 2×2
1 2 3×2 −5 3 3×2
❖ Find 𝐵𝐴 and 𝐶𝐴.
❖ Note that 𝐶𝐵 is not defined.
 13

Note that the system

𝑎11 𝑥1 + 𝑎12 𝑥2 + ⋯+ 𝑎1𝑛 𝑥𝑛 = 𝑏1
𝑎21 𝑥1 + 𝑎22 𝑥2 + ⋯+ 𝑎2𝑛 𝑥2 = 𝑏2
⋮ ⋮ ⋮ ⋮
𝑎𝑚1 𝑥1 + 𝑎𝑚2 𝑥2 + ⋯+ 𝑎𝑚𝑛 𝑥𝑛 = 𝑏𝑚

can be replaced by the single matrix equation 𝐴𝑥 = 𝑏, where 𝐴 = [𝑎𝑖𝑗 ]𝑚×𝑛 (called
𝑥1 𝑏1
𝑥2 𝑏
the coefficient matrix), 𝑥 = [ ⋮ ] and 𝑏 = [ 2 ] .
⋮
𝑥𝑛 𝑛×1 𝑏𝑚 𝑚×1

Remark 8: The 𝑗 𝑡ℎ column of 𝐴𝐵 is 𝐴𝐵𝑗 , where 𝐵𝑗 is the 𝑗 𝑡ℎ column of 𝐵.

The 𝑖 𝑡ℎ row of 𝐴𝐵 is 𝐴𝑖 𝐵, where 𝐴𝑖 is the 𝑖 𝑡ℎ row of 𝐴.

2 0
1 1
Example 9: Let 𝐴 = [1 3] and 𝐵 = [ ] .
−1 −3 2×2
0 −1 3×2
2 0 2
nd 1
❖ The 2 column of 𝐴𝐵 is [1 3 ] [ ] = [−8].
−3
0 −1 3
1 1
❖ The 1st row of AB is [2 0] [ ] = [2 2].
−1 −3
Definition 10: If 𝐴 = [𝑎𝑖𝑗 ]𝑚×𝑛 , then the transpose of 𝐴, denoted by𝐴𝑡 is the 𝑛 × 𝑚
matrix 𝐴𝑡 = [𝑎𝑗𝑖 ]𝑛×𝑚 .

Example 11:
𝑎11 𝑎21 𝑎31
𝑎11 𝑎12 𝑎13 𝑎14
𝑎 𝑎22 𝑎32
(a) If 𝐴 = [𝑎21 𝑎22 𝑎23 𝑎24 ], then 𝐴𝑡 = [ 12
𝑎13 𝑎23 𝑎33 ].
𝑎31 𝑎32 𝑎33 𝑎34
𝑎14 𝑎24 𝑎34
1 2
1 −1 4
(b) If 𝐵 = [−1 0], then 𝐵 𝑡 = [ ].
2 0 3
4 3
1
𝑡
(c) If 𝐶 = [1 −1 0], then 𝐶 = [−1].
0
2 −1 3 2 −1 3
𝑡
(d) If 𝐷 = [−1 4 5 ], then 𝐷 = [−1 4 5]⇒𝐷 𝑡 = 𝐷
3 5 1 3 5 1
(e) If 𝐸 = [5], then 𝐸 𝑡 = [5] = 𝐸.
 14

Section 1.4: Inverses; Algebraic Properties of Matrices

If 𝐴, 𝐵 are 𝑚 × 𝑛 matrices, then 𝐴 + 𝐵 = 𝐵 + 𝐴 i.e., matrix addition is

commutative. However, matrix multiplication is not. If 𝐴 and 𝐵 are two matrices of
the certain sizes, then 𝐴𝐵 maybe defined while 𝐵𝐴 is not defined and it may be the
case that 𝐴𝐵 are 𝐵𝐴 both are defined but 𝐴𝐵 ≠ 𝐵𝐴. If 𝐴, 𝐵 are 𝑛 × 𝑛 matrices
such that 𝐴𝐵 = 𝐵𝐴, we say that 𝐴 and 𝐵 commute.

1 1 2 2 3 3
Example 1: Let 𝐴 = [ ],𝐵 = [ ]. Then 𝐴𝐵 = [ ],
−1 −1 1 1 −3 −3
0 0
while 𝐵𝐴 = [ ],i.e., 𝐴𝐵 ≠ 𝐵𝐴. Thus, 𝐴, 𝐵 do not commute.
0 0

Theorem 2: Assuming that the sizes of the matrices are such that the indicated
operations can be performed, the following are valid:
(a) 𝐴 + 𝐵 = 𝐵 + 𝐴 (Commutative law for addition)
(b) 𝐴 + (𝐵 + 𝐶) = (𝐴 + 𝐵) + 𝐶 (Associative law for matrix addition)
(c) 𝐴(𝐵 𝐶) = (𝐴 𝐵) 𝐶 (Associative law for matrix multiplication)
(d) 𝐴(𝐵 + 𝐶) = 𝐴 𝐵 + 𝐴 𝐶 (Left distributive law of multiplication)
(e) (𝐵 + 𝐶) 𝐴 = 𝐵 𝐴 + 𝐶 𝐴 (Right distributive law of multiplication)
(f) 𝐴(𝐵 − 𝐶) = 𝐴 𝐵 − 𝐴 𝐶
(g) (𝐵 − 𝐶) 𝐴 = 𝐵 𝐴 − 𝐶 𝐴
(h) 𝑎(𝐵 + 𝐶) = 𝑎 𝐵 + 𝑎 𝐶(a is a scalar)
(i) 𝑎(𝐵 − 𝐶) = 𝑎 𝐵 − 𝑎 𝐶
(j) (𝑎 + 𝑏) 𝐶 = 𝑎 𝐶 + 𝑏 𝐶 (a, b are scalars)
(k) (𝑎 − 𝑏) 𝐶 = 𝑎 𝐶 − 𝑏 𝐶
(l) 𝑎(𝑏 𝐶) = (𝑎 𝑏) 𝐶
(m) 𝑎(𝐵𝐶) = (𝑎𝐵)𝐶 = 𝐵(𝑎𝐶)
 15

1 −1
1 3 −2 0
Example 3: Let 𝐴 = [0 1] , 𝐵 = [ ],𝐶 = [ ].
1 4 1 5
2 3
Show that(𝐴𝐵)𝐶 = 𝐴(𝐵𝐶).
Solution:
1 −1 0 −1 −1 5
1 3 −2 0 −2 0
(𝐴𝐵)𝐶 = ([0 1] [ ]) [ ] = [1 4] [ ]=[ 2 20]
1 4 1 5 1 5
2 3 5 18 8 90
1 −1 1 −1 −1 5
1 3 −2 0 1 15
𝐴(𝐵𝐶) = [0 1 ] ([ ][ ]) = [0 1] [ ]=[ 2 20]
1 4 1 5 2 20
2 3 2 3 8 90
⇒(𝐴𝐵)𝐶 = 𝐴(𝐵𝐶).

Definition 4: A matrix, all of whose entries are zero is called a zero matrix,
denoted by 𝟎 = [0𝑖𝑗 ], where 0𝑖𝑗 = 0.

❖ Note that if 𝐴 is any matrix and0 is the zero matrix, then 𝐴 + 𝟎 = 𝟎 + 𝐴 = 𝐴.

(0 is the additive identity)

0 1 1 1 2 5 3 7
Example 5: Let 𝐴 = [ ],𝐵 = [ ],𝐶 = [ ],𝐷 = [ ]. Note that
0 2 3 4 3 4 0 0
3 4
𝐴𝐵 = 𝐴𝐶 = [ ], yet 𝐵 ≠ 𝐶, 𝐴 ≠ 𝟎 (the cancellation law does not hold).
6 8
Also, 𝐴𝐷 = 𝟎, yet 𝐴 ≠ 𝟎, 𝐷 ≠ 𝟎 (there are zero divisors in matrices).

Theorem 6: If 𝐴 it is a matrix and 0 is a zero matrix, then

(a) 𝐴 + 𝟎 = 𝟎 + 𝐴 = 𝐴
(b) 𝐴 − 𝐴 = 𝟎
(c) 𝟎 − 𝐴 = −𝐴
(d) 𝐴𝟎 = 𝟎; 𝟎𝐴 = 𝟎
whenever these operations are defined.
1 0 0 ⋯ 0
0 1 0 ⋯ 0
Definition 7: The 𝑛 × 𝑛 matrix 𝐼𝑛 defined by 𝐼𝑛 = 0 0 1 ⋯ 0 is said to
⋮ ⋮ ⋮ ⋱ ⋮
[0 0 0 ⋯ 1]
1 0 0
1 0
be the identity matrix of order n. Thus, 𝐼1 = [1], 𝐼2 = [ ] , 𝐼 = [0 1 0] etc.
0 1 3
0 0 1
 16

Remark 8:
(1) Sometimes we write 𝐼 instead of 𝐼𝑛 .
(2) If 𝐴 is 𝑚 × 𝑛, then 𝐴𝐼𝑛 = 𝐼𝑛 𝐴 = 𝐴 (verify).

Definition 9: If 𝐴 is a square matrix, and if there is a matrix 𝐵 be such that

𝐴𝐵 = 𝐵𝐴 = 𝐼, then 𝐴 is said to be invertible and 𝐵 is called an inverse of 𝐴.

3 2 −5 2
Example 10: If 𝐴 = [ ], then its inverse is 𝐵 = [ ].
8 5 8 −3
Not that 𝐴𝐵 = 𝐵𝐴 = 𝐼2 .

1 2
Example 11:[ ] is not invertible. Why?
2 4

Theorem12: If 𝐵 and 𝐶 are both inverses of the matrix 𝐴, then 𝐵 = 𝐶.

Proof: We have 𝐵𝐴 = 𝐼 and 𝐴𝐶 = 𝐼. So,

𝐵 = 𝐵𝐼 = 𝐵(𝐴𝐶) = (𝐵𝐴)𝐶 = 𝐼𝐶 = 𝐶.
Thus, the inverse (when it exists) is unique.

❖ If 𝐴 is invertible, then its inverse is denoted by 𝐴−1 . Thus, 𝐴𝐴−1 = 𝐴−1 𝐴 = 𝐼.

𝑎 𝑏
Example 13: If 𝐴 = [ ] and 𝑎𝑑 − 𝑏𝑐 ≠ 0, thenA is invertible and
𝑐 𝑑
𝑑 −𝑏
1 𝑑 −𝑏
𝐴−1 = [ ] = [𝑎𝑑−𝑐
− 𝑏𝑐 𝑎𝑑 − 𝑏𝑐 ].
𝑎
𝑎𝑑 − 𝑏𝑐 −𝑐 𝑎
𝑎𝑑 − 𝑏𝑐 𝑎𝑑 − 𝑏𝑐
Theorem 14: If 𝐴 and 𝐵 are invertible matrices of the same size, then
(a) 𝐴𝐵 is invertible
(b) (𝐴𝐵)−1 = 𝐵 −1 𝐴−1

Proof:This follows from

(𝐴𝐵)(𝐵 −1 𝐴−1 ) = 𝐴(𝐵𝐵 −1 )𝐴−1 = 𝐴𝐼𝐴−1 = 𝐴𝐴−1 = 𝐼
and the similarly established fact that (𝐵 −1 𝐴−1 )(𝐴𝐵) = 𝐼.

Remark 15: If 𝐴1 , 𝐴2 , … , 𝐴𝑘 are invertible matrices of the same size, then so is

𝐴1 𝐴2 ⋯ 𝐴𝑘 and (𝐴1 𝐴2 ⋯ 𝐴𝑘 )−1 = 𝐴−1 −1 −1 −1
𝑘 𝐴𝑘 ⋯ 𝐴2 𝐴1 .
 17

Definition 16: If 𝐴 is a square matrix, then 𝐴0 = 𝐼, 𝐴𝑛 = ⏟

𝐴𝐴 ⋯ 𝐴. If 𝐴 is
𝑛 factors
invertible, then (𝐴𝑛 )−1 = (𝐴−1 )𝑛 =𝐴 𝐴−1 −1 −1
⋯𝐴

Theorem 17: If 𝐴 is a square matrix and 𝑟, 𝑠 are integers, then

𝐴𝑟 𝐴𝑠 = 𝐴𝑟+𝑠 , (𝐴𝑟 ) 𝑠 = 𝐴𝑟𝑠 .
When 𝑟 or 𝑠 is negative, 𝐴 should be invertible.

Theorem 18: I
f 𝐴 is invertible, then
(a) 𝐴−1 is invertible and (𝐴−1 )−1 = 𝐴.
(b) 𝐴𝑛 is invertible and (𝐴𝑛 )−1 = (𝐴−1 )𝑛
1
(c) If 𝑘 ≠ 0 (a scalar), then 𝑘𝐴 is invertible and (𝑘𝐴)−1 = 𝑘 𝐴−1

Proof:
(a) Since 𝐴𝐴−1 = 𝐴−1 𝐴 = 𝐼, 𝐴−1 is invertible and (𝐴−1 )−1 = 𝐴, is invertible.
1 1 1
(c) (𝑘𝐴) ( 𝐴−1 ) = (𝑘𝐴)𝐴−1 = ( 𝑘) 𝐴𝐴−1 = 1𝐼 = 𝐼.
𝑘 𝑘 𝑘
1 1
Similarly, (𝑘 𝐴 ) (𝑘𝐴) = 𝐼. So that 𝑘𝐴 is invertible and (𝑘𝐴)−1 = 𝑘 𝐴−1 .
−1

Example 19: Let 𝐴 be a square matrix such that 𝐴2 − 2𝐴 + 3𝐼 = 0.

(a) Show that 𝐴 is invertible and find 𝐴−1 in terms of 𝐴.
(b) Show that 𝐴 + 4𝐼 is invertible and find (𝐴 + 4𝐼)−1 in terms of 𝐴. (Exercise).

Proof: Observe that 𝐴2 − 2𝐴 + 3𝐼 = 0⇒𝐴2 − 2𝐴 = −3𝐼

(𝐴−2𝐼)
(a) 𝐴(𝐴 − 2𝐼) = −3𝐼⇒𝐴 −3 = 𝐼
(𝐴−2𝐼) 1
⇒𝐴−1 = = − 3 (𝐴 − 2𝐼).
−3

Theorem 20: If 𝐴, 𝐵 are matrices of appropriate sizes, then

(a) (𝐴𝑡 )𝑡 = 𝐴 (b) (𝐴 + 𝐵)𝑡 = 𝐴𝑡 + 𝐵 𝑡
(c) (𝑘𝐴)𝑡 = 𝑘𝐴𝑡 (d) (𝐴𝐵)𝑡 = 𝐵 𝑡 𝐴𝑡

Theorem 21: If 𝐴 is invertible, the 𝐴𝑡 is invertible and (𝐴𝑡 )−1 = (𝐴−1 )𝑡 .

Proof: Note that 𝐴𝑡 (𝐴−1 )𝑡 = (𝐴−1 𝐴)𝑡 = 𝐼 𝑡 = 𝐼. Similarly, (𝐴−1 )𝑡 𝐴𝑡 = 𝐼.

 18

Definition 22: If 𝐴 is an 𝑛 × 𝑛 matrix, then the trace of 𝐴, is defined by

𝑛

tr 𝐴 = ∑ 𝑎𝑖𝑖 (the sum of the diagonal entries of 𝐴)

𝑖=1

Properties of trace 23: Let 𝐴, 𝐵 be 𝑛 × 𝑛 matrices and let 𝑐 be a real number.

Then
(1) tr(𝐴 + 𝐵) = tr 𝐴 + tr 𝐵
(2) tr(𝑐𝐴) = 𝑐 tr 𝐴
(3) tr(𝐴𝐵) = tr(𝐵𝐴)
(4) tr(𝑆 −1 𝐴𝑆) = tr 𝐴 for every invertible matrix 𝑆
(5) tr 𝟎 = 0 and tr 𝐼𝑛 = 𝑛
Example 24: Show that there are no 𝑛 × 𝑛 matrice 𝐴, 𝐵 such that 𝐴𝐵 − 𝐵𝐴 = 𝐼𝑛 .
(Hint: Use a trace argument).
 19

Section 1.5: Elementary Matrices and a Method for Finding 𝑨−𝟏

Definition 1: An 𝑛 × 𝑛 matrix is called an elementary matrix if it can be obtained

from 𝐼𝑛 by performing a single elementary row operation.

Example 2: Each of the following is an elementary matrix:

1 0
(i) [ ] (Multiply the 2nd row of 𝐼2 by −3)
0 −3
1 0 0 0
0 0 0 1
(ii) [ ] (Interchange the 2nd and 4th rows of 𝐼4 )
0 0 1 0
0 1 0 0
1 0 3
(iii) [0 1 0] (Add 3 times the 3rd row of 𝐼3 to the 1st row)
0 0 1
1 0 0 ⋯ 0
0 1 0 ⋯ 0
(iv) 0 0 1 ⋯ 0 (Multiply the 1st row of 𝐼𝑛 by 1)
⋮ ⋮ ⋮ ⋱ ⋮
[0 0 0 ⋯ 1]

Theorem 3:If the elementary matrix 𝐸 results from performing a certain row
operation on 𝐼𝑚 and if 𝐴is an 𝑚 × 𝑛matrix, then the product 𝐸𝐴 is
the matrix obtained when the same row operation is performed on 𝐴.
1 0 2 𝑎11 𝑎12 𝑎13
Example 4:Let 𝐸 = [0 1 0] and let 𝐴 = [𝑎21 𝑎22 𝑎23 ]. Then 𝐸 is obtained
0 0 1 𝑎31 𝑎32 𝑎33
from 𝐼3 by adding 2 times the 3rd row to the 1st row of 𝐼3 .
𝑎11 + 2𝑎31 𝑎12 + 2𝑎32 𝑎13 + 2𝑎33
Now, 𝐸𝐴 = [ 𝑎21 𝑎22 𝑎23 ].
𝑎31 𝑎32 𝑎33
 20

Theorem 5: Every elementary matrix is invertible, and the inverse is also an

elementary matrix.

Proof: Let 𝐸 be an elementary matrix obtained from 𝐼𝑛 by some elementary row

operations. Let 𝐸0 be the matrix obtained when the inverse of this operation is
performed on 𝐼𝑛 . Then by the previous theorem, we have 𝐸0 𝐸 = 𝐼𝑛 and
𝐸𝐸0 = 𝐼𝑛 . Hence, the elementary matrix 𝐸0 is the inverse of 𝐸.

Definition 6: Two matrices of the same size are called row equivalent if one can
be obtained by applying a finite sequence of elementary row
operations on the other.

Theorem 7: If 𝐴 is an 𝑛 × 𝑛 matrix, then the following statements are equivalent:

(a) 𝐴 is invertible
(b) 𝐴𝑥 = 0 has only the trivial solution
(c) 𝐴 is row equivalent to 𝐼𝑛 (the r.r.e.f. of 𝐴 is 𝐼𝑛 )
(d) 𝐴 is a product of elementary matrices

Proof: We shall prove (a) ⇒ (b) ⇒ (c) ⇒ (d) ⇒ (a).

(a) ⇒ (b): Assume 𝐴 is invertible and let 𝑥0 be any solution of 𝐴𝑥 = 𝟎 i.e.,
𝐴𝑥0 = 𝟎. Multiplying both sides of this equation by 𝐴−1 gives
𝐴−1 (𝐴𝑥0 ) = 𝟎. So, (𝐴−1 𝐴)𝑥0 = 𝟎 or 𝐼𝑛 𝑥0 = 0. Thus, 𝑥0 = 𝟎.

(b) ⇒ (c): Let 𝐴𝑥 = 𝟎 be the matrix form of the system

𝑎11 𝑥1 + 𝑎12 𝑥2 + ⋯ + 𝑎1𝑛 𝑥𝑛 = 0
𝑎21 𝑥1 + 𝑎22 𝑥2 + ⋯ + 𝑎2𝑛 𝑥2 = 0
]…………..(1)
⋮ ⋮ ⋮ ⋮
𝑎𝑛1 𝑥1 + 𝑎𝑛2 𝑥2 + ⋯ + 𝑎𝑛𝑛 𝑥𝑛 = 0
and assume that this system has only the trivial solution. Hence, solving
by Gauss-Jordan elimination, the system of equations corresponding to the
r.r.e.f. of the augmented matrix is
𝑥1 =0
⋯ 𝑥2 =0
……….(2)
⋱ =0
𝑥𝑛 = 0]

Thus, augmented matrix of (1) has the r.r.e.f. (the augmented matrix of (2))
Hence, 𝐴 is row equivalent to 𝐼𝑛 .
 21

(c) ⇒ (d): Assume 𝐴 is row equivalent to 𝐼𝑛 . Then there are elementary matrices
𝐸1 , 𝐸2 , … , 𝐸𝑘 such that 𝐸𝑘 ⋯ 𝐸2 𝐸1 𝐴 = 𝐼𝑛
⇒𝐴−1 = 𝐸𝑘 ⋯ 𝐸2 𝐸1⇒𝐴 = (𝐴−1 )−1 = (𝐸𝑘 ⋯ 𝐸2 𝐸1 )−1 .
But the inverses of E1 , 𝐸2 , … , 𝐸𝑘 are elementary matrices. Hence, 𝐴 =
𝐸1−1 𝐸2−1 ⋯ 𝐸𝑘−1 is a product of elementary matrices.

(d) ⇒ (a): If 𝐴 is the product of elementary matrices, then 𝐴 is the product of

invertible matrices, and so 𝐴 is invertible.

Remark 8: Part (c) says that 𝐼𝑛 is the r.r.e.f. of 𝐴. From the proof of the theorem
we have 𝐴−1 = 𝐸𝑘 ⋯ 𝐸2 𝐸1 and so 𝐴−1 = 𝐸𝑘 ⋯ 𝐸2 𝐸1 𝐼𝑛 , that is 𝐴−1 can
be obtained from 𝐼𝑛 by a sequence of elementary row operations. This
sequence is the same one which reduces 𝐴 to 𝐼𝑛 .
2 0 1
Example 9: Find the inverse of 𝐴 = [2 1 −1] if it exists.
3 1 −1

Solution: We start with [𝐴|𝐼3 ] and reduce it (if possible) to [𝐼3 |𝐵], so we get
𝐴−1 = 𝐵.
1
R
2 0 1 1 0 0 2 1 1 0 1/2 1/2 0 0
[𝐴|𝐼3 ] = [2 1 −1| 0 1 0] ⃗⃗⃗ [2 1 −1 | 0 1 0]
3 1 −1 0 0 1 3 1 −1 0 0 1

−2R1 + R 2 1 0 1/2 1/2 0 0 −R + R 1 0 1/2 1/2 0 0

2 3
⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ [0 1 −2 | −1 1 0] [0 1 −2 | −1 1 0]
⃗⃗⃗
−3R1 + R 3 0 1 −5/2 −3/2 0 1 0 0 −1/2 −1/2 −1 1

−2R 3 1 0 1/2 1/2 0 0 2R 3 + R 2 1 0 0 0 −1 1

[0 1 −2 | −1 1 0 ] ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗
1
|
0 1 0| 1 5 −4 = [𝐼3 |𝐴−1 ].
⃗⃗⃗ − 2 R 3 + R1
0 0 1 1 2 −2 0⏟ 0 1 ⏟1 2 −2
[ 𝐼3 𝐴−1 ]
0 −1 1
−1
Thus, 𝐴 = [1 5 −4]. (Check that 𝐴−1 𝐴 = 𝐴𝐴−1 = 𝐼3 )
1 2 −2
 22

2 0 1
Example 10: Find the inverse (if it exists) of the matrix 𝐴 = [ 3 1 −1].
−2 −2 4

Solution:
1
R
2 0 1 1 0 0 2 1 1 0 1/2 1/2 0 0
[𝐴|𝐼3 ] = [ 3 1 −1| 0 1 0] ⃗⃗⃗ [ 3 1 −1 | 0 1 0]
−2 −2 4 0 0 1 −2 −2 4 0 0 1

−3R1 + R 2 1 0 1/2 1/2 0 0 2R + R 1 0 1/2 1/2 0 0

2 3
⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ [0 1 −5/2| −3/2 1 0] [0 1 −5/2| −3/2 1 0]
⃗⃗⃗
2R1 + R 3 0 −2 5 1 0 1 0 0 0 −2 2 1
This shows that 𝐴 is not invertible.

Remark 11: If 𝐴 is invertible, then the homogeneous system 𝐴𝑥 = 𝟎 has only the
trivial solution. This can be seen as follows:
𝐴𝑥 = 𝟎⇒𝐴−1 (𝐴𝑥) = 𝐴−1 𝟎⇒(𝐴−1 𝐴)𝑥 = 𝟎⇒𝐼𝑛 𝑥 = 𝟎⇒𝑥 = 𝟎.

Example 12: Show that the homogeneous system

2𝑥1 + 𝑥3 = 0
2𝑥1 + 𝑥2 − 𝑥3 = 0
3𝑥1 + 𝑥2 − 𝑥3 = 0
has only the trivial solution 𝑥1 = 0, 𝑥2 = 0, 𝑥3 = 0
Solution: The matrix form of this system is 𝐴𝑥 = 𝟎, where
2 0 1
𝐴 = [2 1 −1]
3 1 −1
Since 𝐴 is invertible (see Example 9), it follows that the homogeneous system has
only the trivial solution.
 23

Section 1.6:Further Results on Systems of Equations and

Invertibility
Theorem 1: Every system of linear equations has either no solutions, exactly one
solution, or infinitely many solutions.

Theorem 2: If 𝐴 is an 𝑛 × 𝑛 invertible matrix then for each 𝑛 × 1 matrix 𝑏, the

system 𝐴𝑥 = 𝑏 has exactly one solution, namely 𝑥 = 𝐴−1 𝑏.

Proof: Since 𝐴(𝐴−1 𝑏) = 𝑏, 𝑥 = 𝐴−1 𝑏 is a solution of 𝐴𝑥 = 𝑏.

Assume that 𝑥0 is an arbitrary solution of the system 𝐴𝑥 = 𝑏. Then 𝐴𝑥0 = 𝑏. So,
𝐴−1 (𝐴𝑥0 ) = 𝐴−1 𝑏⇒(𝐴−1 𝐴)𝑥0 = 𝐴−1 𝑏⇒𝐼𝑛 𝑥0 = 𝐴−1 𝑏⇒𝑥0 = 𝐴−1 𝑏.

Example 3: Show that the system

2𝑥1 + 𝑥3 = 2
2𝑥1 + 𝑥2 − 𝑥3 = 0
3𝑥1 + 𝑥2 − 𝑥3 = −1
has a unique solution and find this solution.

Solution: The system has the form 𝐴𝑥 = 𝑏, where

2 0 1 𝑥1 2
𝑥
𝐴 = [2 1 −1] , 𝑥 = [ 2 ] , and 𝑏 = [ 0 ].
3 1 −1 𝑥3 −1
By a previous example , since 𝐴−1 exists, we have 𝐴 is invertible, so the system has
0 −1 1 2 −1
−1
a unique solution. To find this solution: 𝑥 = 𝐴 𝑏 = [1 5 −4] [ 0 ] = [ 6 ]
1 2 −2 −1 4
𝑥1 −1
⇒[𝑥2 ] = [ 6 ]⇒ The unique solution is 𝑥1 = −1, 𝑥2 = 6, 𝑥3 = 4.
𝑥3 4

Theorem 4: Let 𝐴 be a square matrix.

(a) If 𝐵 is a square matrix such that 𝐵𝐴 = 𝐼, then 𝐵 = 𝐴−1 .
(b) If B is a square matrix such that 𝐴𝐵 = 𝐼, then 𝐵 = 𝐴−1 .

Proof: We prove part (a). The proof of part (b) is an exercise. Assume 𝐵𝐴 = 𝐼.
We show that 𝐴 is invertible by proving that 𝐴𝑥 = 𝟎 has only the trivial
solution. If 𝐴𝑥 = 𝟎, then 𝐵(𝐴𝑥) = 𝐵𝟎. So,(𝐵𝐴)𝑥 = 𝟎 or 𝐼𝑥 = 𝟎. That
is,𝑥 = 𝟎. Hence, 𝐴 is invertible.
From 𝐵𝐴 = 𝐼 we get 𝐵𝐴𝐴−1 = 𝐼𝐴−1 , and so 𝐵 = 𝐴−1 .
 24

Theorem 5: Let 𝐴 and 𝐵 be square matrices of the same size. If 𝐴𝐵 is invertible,

then both 𝐴 and 𝐵 are invertible.

Theorem 6: If 𝐴 is an 𝑛 × 𝑛 matrix, then the following are equivalent:

(a) 𝐴 is invertible
(b) 𝐴𝑥 = 𝟎 has only the trivial solution.
(c) 𝐴 is row equivalent to 𝐼𝑛 (i.e., the r.r.e.f. of 𝐴 is 𝐼𝑛 ).
(d) 𝐴 is a product of elementary matrices.
(e) 𝐴𝑥 = 𝑏 is consistent for every𝑛 × 1 matrix 𝑏.
(f) 𝐴𝑥 = 𝑏 has exactly one solution for every 𝑛 × 1 matrix 𝑏.

Example 7: Find the conditions on 𝑏1 , 𝑏2 , 𝑏3 so that the system is consistent.

𝑥1 + 𝑥2 − 𝑥3 = 𝑏1
−2𝑥1 + 𝑥2 +3𝑥3 = 𝑏2
−𝑥1 +2𝑥2 +2𝑥3 = 𝑏3
Solution:
1 1 −1 𝑏1 2R + R 1 1 −1 𝑏1
1 2
[−2 1 3 𝑏2 ] [0 3 1 2𝑏1 + 𝑏2 ]
⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗
R 1 + 𝑅 3
−1 2 2 𝑏3 0 3 1 𝑏1 + 𝑏3
1 −1 1 𝑏1
−R 2 + R 3
3 [0
1 2𝑏1 + 𝑏2 ].
⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ . 0
0 0 −𝑏1 − 𝑏2 + 𝑏3
So, the system is consistent if −𝑏1 − 𝑏2 + 𝑏3 = 0, i.e., if 𝑏3 = 𝑏1 + 𝑏2 .

Example 8:Show that the system

𝑥1 + 𝑥2 +2𝑥3 = 𝑏1
𝑥1 −2𝑥2 + 𝑥3 = 𝑏2
2𝑥1 + 𝑥2 − 𝑥3 = 𝑏3
is consistent for all 𝑏1 , 𝑏2 , 𝑏3 and solve the system.

Solution:Observe that the matrix form of the system is 𝐴𝑥 = 𝑏, where

1 1 2 𝑥1 𝑏1
𝐴 = [1 −2 1 ] , 𝑥 = [𝑥2 ] , 𝑏 = [𝑏2 ].
2 1 −1 𝑥3 𝑏3
1 3 5
14 14 14
−1 3 5 1
Note that the matrix 𝐴 is invertible (why?) with 𝐴 = − 14 .
14 14
5 1 3
[14 − 14]
14
 25

So, the system is consistent.

To solve the system: Note that 𝐴𝑥 = 𝑏⇒𝐴−1 𝐴𝑥 = 𝐴−1 𝑏⇒𝑥 = 𝐴−1 𝑏

1 3 5 𝑏1 +3𝑏2 +5𝑏3 𝑏1 +3𝑏2 +5𝑏3
14 14 14 𝑏1 14 𝑥1 14
3 5 1 3𝑏1 −5𝑏2 +𝑏3 3𝑏1 −5𝑏2 +𝑏3
⇒𝑥 = 14
− 14 14
[𝑏2 ] = ⇒[𝑥2 ] =
14 14
5 1 3 𝑏3 5𝑏1 +𝑏2 −3𝑏3 𝑥3 5𝑏1 +𝑏2 −3𝑏3
[14 −
14 14] [ 14 ] [ 14 ]
𝑏1 +3𝑏2 +5𝑏3 3𝑏1 −5𝑏2 +𝑏3 5𝑏1 +𝑏2 −3𝑏3
∴ The solution is 𝑥1 = , 𝑥2 = , 𝑥3 =
14 14 14

Section 1.7: Diagonal, Triangular, and Symmetric Matrices

Definition 1: An 𝑛 × 𝑛 matrix 𝐴 is called:

𝑎11 0 0 ⋯ 0
0 𝑎22 0 ⋯ ⋮
(1) diagonal if it is of the form 𝐴 = 0 0 𝑎33 ⋮ . That is all the
⋮ ⋮ ⋱ 0
[ 0 ⋯ ⋯ 0 𝑎𝑛𝑛 ]
entries off the main diagonal of 𝐴 are zero.
(i.e., 𝐴 = [𝑎𝑖𝑗 ] is diagonal ⇔𝑎𝑖𝑗 = 0, 𝑓𝑜𝑟 𝑖 ≠ 𝑗).

𝑎11 𝑎12 ⋯ ⋯ 𝑎1𝑛

0 𝑎22 ⋯ ⋯ ⋮
(2) upper triangular if it is of the form A = 0 0 𝑎33 ⋯ ⋮ . That is
⋮ ⋮ ⋱ ⋮
[ 0 ⋯ ⋯ 0 𝑎𝑛𝑛 ]
all the entries below the main diagonal of 𝐴 are zero.
(i.e., 𝐴 = [𝑎𝑖𝑗 ] is upper triangular ⇔𝑎𝑖𝑗 = 0, 𝑓𝑜𝑟 𝑖 > 𝑗).

𝑎11 0 0 ⋯ 0
𝑎21 𝑎22 0 ⋯ ⋮
(3) 𝐴 is lower triangular if it is of the form A = ⋮ ⋮ 𝑎33 ⋮ .
⋮ ⋮ ⋱ 0
[𝑎𝑛1 ⋯ ⋯ 𝑎𝑛𝑛 ]
That is all the entries above the main diagonal of 𝐴 are zero.
(i.e., 𝐴 = [𝑎𝑖𝑗 ] is lower triangular ⇔𝑎𝑖𝑗 = 0, 𝑓𝑜𝑟 𝑖 < 𝑗).
 26

(4) If 𝐴 is upper triangular or lower triangular, then 𝐴 is called triangular, that

is 𝐴 is called triangular if it is either upper triangular or lower triangular.

❖ Note that diagonal matrices are both upper and lower triangular.

Example 2:
2 0 0
(1) 𝐴 = [0 1 0] is diagonal. So, 𝐴 is both upper and lower triangular.
0 0 6

2 0 4
(2) 𝐴 = [0 1 0] is upper triangular
0 0 6

2 0 0 0
𝑒 1 0 0
(3) 𝐴 = [ ] is lower triangular
4 0 6 0
3 0 𝜋 −1
Theorem 3:
(a) The transpose of a lower triangular matrix is upper triangular, and the transpose
of an upper triangular matrix is lower triangular.
(b) The product of lower triangular matrices is lower triangular, and the product of
upper triangular matrices is upper triangular.
(c) A triangular matrix is invertible if and only if its diagonal entries are all
nonzero.
(d) The inverse of an invertible lower triangular matrix is lower triangular, and the
inverse of an invertible upper triangular matrix is upper triangular.
1 3 −1 3 −2 2
Theorem 4: Let 𝐴 = [0 2 4 ] and 𝐵 = [0 0 −1]. Then 𝐴, 𝐵 are upper
0 0 5 0 0 1
3 7
1 −
2 5 3 −2 −2
−1 1 2
triangular. Also, 𝐴 = 0 2
−5 and 𝐴𝐵 = [0 0 2 ] are upper
1 0 0 5
[0 0 5 ]
triangular.
 27

Definition 5:
(a) A square matrix 𝐴 is said to be symmetric if 𝐴 = 𝐴𝑡 .
(b) A square matrix 𝐴 is said to be skew-symmetric if 𝐴 = −𝐴𝑡 .
Remark 6: An 𝑛 × 𝑛 matrix 𝐴 = [𝑎𝑖𝑗 ] is:
(a) symmetric if 𝑎𝑖𝑗 = 𝑎𝑗𝑖 for all 𝑖, 𝑗 = 1,2, … , 𝑛.
(b) skew-symmetric if 𝑎𝑖𝑗 = −𝑎𝑗𝑖 for all 𝑖, 𝑗 = 1,2, … , 𝑛. Note that the diagonal
entries of a skew-symmetric matrix are 0.
Example 7:
7 −3 7 −3
(1) 𝐴 = [ ] is symmetric because 𝐴𝑡 = [ ]=𝐴
−3 8 −3 8
1 4 5 1 4 5
𝑡
(2) 𝐴 = [4 −5 0] is symmetric because 𝐴 = [4 −5 0] = 𝐴
5 0 7 5 0 7
𝑑1 0 0 0 𝑑1 0 0 0
0 𝑑2 0 0 0 𝑑2 0 0
(3) 𝐴 = [ ] is symmetric because 𝐴𝑡 = [ ]=𝐴
0 0 𝑑3 0 0 0 𝑑3 0
0 0 0 𝑑4 0 0 0 𝑑4
0 −4 5 0 4 −5
𝑡
(4) 𝐴 = [ 4 0 0 ] is skew-symmetric because 𝐴 = [−4 0 0 ] = −𝐴
−5 0 0 5 0 0
0 0
(5) 𝐴 = [ ] is both symmetric and skew-symmetric because 𝐴𝑡 = 𝐴, 𝐴𝑡 = −𝐴
0 0
 28

Example 8: Show that if 𝐴 is a square matrix, then 𝐴 + 𝐴𝑡 is symmetric and

𝐴 − 𝐴𝑡 is skew-symmetric. Use this to show that A can be written as
the sum of a symmetric matrix and a skew-symmetric matrix.
Solution:
➢ (𝐴 + 𝐴𝑡 )𝑡 = 𝐴𝑡 + (𝐴𝑡 )𝑡 = 𝐴𝑡 + 𝐴 = 𝐴 + 𝐴𝑡 ⇒𝐴 + 𝐴𝑡 is symmetric
➢ (𝐴 − 𝐴𝑡 )𝑡 = 𝐴𝑡 − (𝐴𝑡 )𝑡 = 𝐴𝑡 − 𝐴 = −(𝐴 − 𝐴𝑡 )⇒𝐴 − 𝐴𝑡 is skew-symmetric.

Theorem 9: If 𝐴 and 𝐵 are symmetric matrices with the same size, and if 𝑘 is any
scalar, then
(a) 𝐴𝑡 is symmetric.
(b) 𝐴 + 𝐵 and 𝐴 − 𝐵 are symmetric.
(c) 𝑘𝐴 is symmetric
Proof:𝐴 and 𝐵 are symmetric⇒𝐴𝑡 = 𝐴 and 𝐵 𝑡 = 𝐵
(a) (𝐴𝑡 )𝑡 = 𝐴 = 𝐴𝑡 ⇒(𝐴𝑡 )𝑡 = 𝐴𝑡 ⇒𝐴𝑡 is symmetric.
(b) (𝐴 + 𝐵)𝑡 = 𝐴𝑡 + 𝐵 𝑡 = 𝐴 + 𝐵⇒𝐴 + 𝐵 is symmetric.
(𝐴 − 𝐵)𝑡 = 𝐴𝑡 − 𝐵 𝑡 = 𝐴 − 𝐵⇒𝐴 − 𝐵 is symmetric.
(c) (𝑘𝐴)𝑡 = 𝑘𝐴𝑡 = 𝑘𝐴⇒𝑘𝐴 is symmetric.

It is not true, in general, that the product of symmetric matrices is

symmetric. To see why this is so, let 𝐴 and 𝐵 be symmetric matrices with the
same size. Then
(𝐴𝐵)𝑡 = 𝐵 𝑡 𝐴𝑡 = 𝐵𝐴. Thus,(𝐴𝐵)𝑡 = 𝐴𝐵⇔𝐴𝐵 = 𝐵𝐴⇔𝐴 and 𝐵 commute.
1 2 1 1
Example 10: Let 𝐴 = [ ] and 𝐵 = [ ]. Then 𝐴, 𝐵 are symmetric and they
2 3 1 0
do not commute (𝐴𝐵 ≠ 𝐵𝐴) ⇒𝐴𝐵 is not symmetric.
1 2 1 1 3 1
Observe that 𝐴𝐵 = [ ][ ]=[ ] is not symmetric.
2 3 1 0 5 2

In summary, we have the following result.

 29

Theorem 11: The product of two symmetric matrices is symmetric if and only if
the matrices commute.
Proof:Let 𝐴, 𝐵 be symmetric matrices of the same size ⇒𝐴𝑡 = 𝐴 and 𝐵 𝑡 = 𝐵.
Then
𝐴𝐵 is symmetric ⇔(𝐴𝐵)𝑡 = 𝐴𝐵⇔𝐵 𝑡 𝐴𝑡 = 𝐴𝐵⇔𝐵𝐴 = 𝐴𝐵
⇔𝐴, 𝐵 commute.

❖ The Products 𝑨𝑨𝒕 and 𝑨𝒕 𝑨 are symmetric: Matrix products of the form 𝐴𝐴𝑡
and 𝐴𝑡 𝐴 arise in a variety of applications. If 𝐴 is an 𝑚 × 𝑛 matrix, then 𝐴𝑡 is an
𝑛 × 𝑚 matrix, so the products 𝐴𝐴𝑡 and 𝐴𝑡 𝐴 are both square matrices. The
matrix 𝐴𝐴𝑡 has size 𝑚 × 𝑚, and the matrix 𝐴𝑡 𝐴 has size 𝑛 × 𝑛. Such products
are always symmetric since
➢ (𝐴𝐴𝑡 )𝑡 = (𝐴𝑡 )𝑡 𝐴𝑡 = 𝐴𝐴𝑡 (that is, 𝐴𝐴𝑡 is symmetric)
(𝐴𝑡 𝐴)𝑡 = 𝐴𝑡 (𝐴𝑡 )𝑡 = 𝐴𝑡 𝐴 (that is, 𝐴𝑡 𝐴 is symmetric).

❖ Invertibility of symmetric matrices: In general, a symmetric matrix need not

be invertible. For example, a diagonal matrix with a zero on the main diagonal
is symmetric but not invertible. However, the following theorem shows that if a
symmetric matrix happens to be invertible, then its inverse must also be
symmetric.
Theorem 12: If 𝐴 is an invertible symmetric matrix, then 𝐴−1 is symmetric.

Proof: Assume that 𝐴 is symmetric and invertible. We have

(𝐴−1 )𝑡 = (𝐴
⏟ 𝑡 )−1 = 𝐴−1 , which proves that 𝐴−1 is symmetric.
since 𝐴𝑡 =𝐴

Theorem 13: If 𝐴 is an invertible matrix, then 𝐴𝐴𝑡 and 𝐴𝑡 𝐴 are also invertible.

Proof: Since 𝐴 is invertible, so is 𝐴𝑡 . Thus, 𝐴𝐴𝑡 and 𝐴𝑡 𝐴 are invertible, since they
are the products of invertible matrices.