Mathematical Analysis and Linear Algebra - lecture – Section 2 – Matrix Algebra

Section 2 Algebra of matrices

2.1. Matrix – definition and specific matrices.
2.2. Operations on matrices and their properties.
2.3. Determinant of a square matrix and its properties.
2.4. Elementary row or column operations.
2.5. Inverse of a matrix.
2.6. Orthogonal matrix.
2.7. Matrix equations.
Matrix theory is a powerful mathematical tool for dealing with data as a whole rather than individual
items. Many situations in both pure and applied mathematics deal with rectangular arrays of numbers. In fact,
in many branches of business and biological and social sciences, it is necessary to express and use a set of
numbers in a rectangular array. This array we called a matrix. Matrices can be added, subtracted and
multiplied. They also possess many of the algebraic properties of numbers. Matrix algebra is the study of these
properties.

2.1. Matrix – definition and specific matrices

Definition 2.1 (formal) A matrix 𝐴 of 𝑚 rows and 𝑛 columns is a mapping defined on a Cartesian
product of two sets into a nonempty set 𝑉 ≠ ∅: 𝐴 ∶ {1,2, … , 𝑚} × {1,2, … , 𝑛} → 𝑉
𝐴 ∶ (𝑖, 𝑗) → 𝑎𝑖𝑗 ∈ 𝑉

For 𝑉 = ℝ we have a real matrix.

For 𝑉 = ℂ we have a complex matrix.
We denote matrices using capital letters: 𝐴, 𝐵, 𝐶, …

Remarks.
 A matrix (plural: matrices) can be informally defined as a rectangular array of elements that are
enclosed by a pair of brackets:
𝑎11 𝑎12 … 𝑎1𝑗 … 𝑎1𝑛 ← 1𝑠𝑡 𝑟𝑜𝑤
𝑎21 𝑎22 … 𝑎2𝑗 … 𝑎2𝑛 ← 2𝑛𝑑 𝑟𝑜𝑤
⋮ ⋮ ⋱ ⋮ ⋱ ⋮ ⋮
𝑎𝑖1 𝑎𝑖2 ⋯ 𝑎𝑖𝑗 ⋯ 𝑎𝑖𝑛 ← 𝑖𝑡ℎ 𝑟𝑜𝑤
𝐴 = 𝐴(𝑚,𝑛) = [𝑎𝑖𝑗 ] = ⋮ ⋮ ⋱ ⋮ ⋱ ⋮ ⋮
[𝑎𝑚1 𝑎𝑚2 ⋯ 𝑎𝑚𝑗 ⋯ 𝑎𝑚𝑛 ] ← 𝑚𝑡ℎ 𝑟𝑜𝑤
↑ ↑ ↑ ↑
1𝑠𝑡 2𝑛𝑑 𝑗𝑡ℎ 𝑛𝑡ℎ 𝑐𝑜𝑙𝑢𝑚𝑛
 a row of a matrix is a horizontal vector in matrix 𝐴; [𝑎𝑖1 , 𝑎𝑖2 , … , 𝑎𝑖𝑛 ] = 𝑟𝑖 is an 𝑖th row,
 a column of a matrix is a vertical vector in matrix 𝐴; [𝑎1𝑗 , 𝑎2𝑗 , … , 𝑎𝑚𝑗 ] = 𝑐𝑗 is a 𝑗th column,
 the numbers 𝑎𝑖𝑗 in matrix 𝐴 are called the elements or entries or components of the matrix,
 each element 𝑎𝑖𝑗 of a matrix 𝐴 has two indices: the row index 𝑖, and the column index 𝑗 (𝑎𝑖𝑗 is an
entry in the 𝑖th row and 𝑗th column),
 the number of rows and columns, 𝑚 × 𝑛 (read as: “𝑚 by 𝑛”), is the dimension or size of the
matrix; in general, 𝑚 × 𝑛 matrix has 𝑚 rows and 𝑛 columns.
Example 2.1.

 Beata Ciałowicz 8
 Mathematical Analysis and Linear Algebra - lecture – Section 2 – Matrix Algebra
Specific matrices
1. A column matrix or column vector of height 𝑚 (𝑚 × 1 matrix, 𝑛 = 1) is a matrix with 1 column of
5
elements: 𝐴 = [−3]
2 3×1
A row matrix or row vector of length 𝑛 (1 × 𝑛 matrix, 𝑚 = 1) is a matrix with 1 row
of elements: 𝐴 = [7 2 + 𝑖 −1 3𝑖 ]1×4

2. A zero (null) matrix is a matrix in which all entries are zero. We use the symbol 0 to represent a zero
0 0 0
matrix of any dimensions: 𝟎=[ ]
0 0 0 2×3
𝑎11 𝑎12 … 𝑎1𝑛
𝑎21 𝑎22 … 𝑎2𝑛
3. A square matrix has as many rows as columns, i.e. 𝑚 = 𝑛: 𝐴 = 𝐴(𝑛,𝑛) = [ ⋮ ⋮ ⋱ ⋮ ]
𝑎𝑛1 𝑎𝑛2 … 𝑎𝑛𝑛 𝑛×𝑛
In the square matrix 𝐴 = [𝑎𝑖𝑗 ]𝑛×𝑛 the entries for which 𝑖 = 𝑗 (elements for which the row number and
column number are equal), namely {𝑎11 , 𝑎22 , … , 𝑎𝑛𝑛 }, 𝑛 ∈ ℕ, form a principal diagonal (or main
diagonal or primary diagonal or diagonal entries) of 𝐴.
Elements {𝑎1𝑛 , 𝑎2(𝑛−1) , … , 𝑎𝑛1 } form a secondary diagonal.
The trace of the square matrix 𝐴 (𝑡𝑟𝐴) is defined as the sum of elements on the main diagonal:
𝑡𝑟𝐴 = ∑𝑛𝑖=1 𝑎𝑖𝑖 = 𝑎11 + 𝑎22 + ⋯ + 𝑎𝑛𝑛
4. An upper triangular matrix is a square matrix with all the elements below the main diagonal equal
𝑎11 𝑎12 … 𝑎1𝑛
0 𝑎22 … 𝑎2𝑛
zero: ∀𝑖, 𝑗 ∈ {1,2, … , 𝑛}; 𝑖 > 𝑗 ∶ 𝑎𝑖𝑗 = 0 𝑈 = [ ⋱ ⋮ ]
⋮ ⋮
0 0 … 𝑎𝑛𝑛 𝑛×𝑛
An lower triangular matrix is a square matrix with all the elements above the main diagonal equal
𝑎11 0 … 0
𝑎21 𝑎22 … 0
zero: ∀𝑖, 𝑗 ∈ {1,2, … , 𝑛}; 𝑖 < 𝑗 ∶ 𝑎𝑖𝑗 = 0 𝐿=[ ]
⋮ ⋮ ⋱ ⋮
𝑎𝑛1 𝑎𝑛2 … 𝑎𝑛𝑛 𝑛×𝑛
5. A diagonal matrix is a square matrix with all non-diagonal elements equal zero:
𝑎11 0 … 0
0 𝑎22 5 0 0
… 0
∀𝑖, 𝑗 ∈ {1,2, … , 𝑛}; 𝑖 ≠ 𝑗 ∶ 𝑎𝑖𝑗 = 0 𝐷 = [ ] e.g. 𝐷 = [0 3 − 𝑖 0]
⋮ ⋮ ⋱ ⋮
… 𝑎𝑛𝑛 𝑛×𝑛 0 0 4 3×3
0 0
5. The identity (unit) matrix is a diagonal matrix such that every diagonal element is equal to 1:
1 0 … 0 1 0 0
0 for 𝑖 ≠ 𝑗 1 0
𝑎𝑖𝑗 = { 𝑰 = 𝐼𝑛 = [0 1 … 0] e.g. 𝐼2 = [ ] 𝐼3 = [0 1 0]
1 for 𝑖 = 𝑗 ⋮ ⋮ ⋱ ⋮ 0 1
0 0 … 1 𝑛×𝑛 0 0 1
Notation:
 𝑀(𝑚, 𝑛) is a set of all matrices of dimensions 𝑚 × 𝑛.
 𝑀𝑚×𝑛 (ℝ) is a set of all real matrices of dimensions 𝑚 × 𝑛.
 𝑀𝑛 (ℝ) is a set of all real square matrices of dimensions 𝑛 × 𝑛.

 Beata Ciałowicz 9
 Mathematical Analysis and Linear Algebra - lecture – Section 2 – Matrix Algebra
2.2 Operations on (with) matrices and their properties
 Multiplication by scalar (scalar multiplication)

Let 𝐴 = [𝑎𝑖𝑗 ] be an 𝑚 × 𝑛 matrix and 𝑘 ∈ ℝ be a real number, called a scalar. The product of the matrix 𝐴
by the scalar 𝑘, called scalar multiplication, is the 𝑚 × 𝑛 matrix:
𝑘 ∙ 𝐴 = [𝑘 ∙ 𝑎𝑖𝑗 ]𝑚×𝑛 .

Example 2.2 Multiply given matrix 𝐴 by scalar 𝑘.

 Transposition
The transpose of a matrix 𝐴, denoted by 𝐴𝑇 , is found by interchanging rows and columns of 𝐴.

Let 𝐴 = [𝑎𝑖𝑗 ] be an 𝑚 × 𝑛 matrix. The transpose of the matrix 𝐴 is the 𝑛 × 𝑚 matrix:

𝐴𝑇 = [𝑎𝑗𝑖 ]𝑛×𝑚

Example 2.3 Find the transpose of 𝐴.

Remark A symmetric matrix is a square matrix 𝐴 ∈ 𝑀(𝑚, 𝑛) that is equal to its transpose: 𝐴𝑇 = 𝐴.

Theorem 2.1 (properties of scalar multiplication and transposition)

Let 𝐴 ∈ 𝑀(𝑚, 𝑛), 𝐵 ∈ 𝑀(𝑛, 𝑝), 𝐼 ∈ 𝑀(𝑛, 𝑛) and 𝑘 ∈ ℝ. Then:
1. 𝐼 𝑇 = 𝐼
2. (𝐴𝑇 )𝑇 = 𝐴
3. (𝑘 ∙ 𝐴)𝑇 = 𝑘 ∙ 𝐴𝑇
4. (𝐴 ∙ 𝐵)𝑇 = 𝐵 𝑇 ∙ 𝐴𝑇
5. 𝑡𝑟(𝑘 ∙ 𝐴) = 𝑘 ∙ 𝑡𝑟(𝐴)

 Beata Ciałowicz 10
 Mathematical Analysis and Linear Algebra - lecture – Section 2 – Matrix Algebra
 Matrix addition and subtraction (sum of two matrices)

If 𝐴 = [𝑎𝑖𝑗 ]𝑚×𝑛 and 𝐵 = [𝑏𝑖𝑗 ]𝑚×𝑛 are two matrices of the same dimensions, then the sum and the
difference are defined as the 𝑚 × 𝑛 matrices: 𝐴 + 𝐵 = [𝑎𝑖𝑗 + 𝑏𝑖𝑗 ]𝑚×𝑛 , 𝐴 − 𝐵 = [𝑎𝑖𝑗 − 𝑏𝑖𝑗 ]𝑚×𝑛

Caution ! The sum or difference of two matrices DOES NOT EXIST if the dimensions of matrices are not
the same.
Example 2.4 Add and subtract given matrices 𝐴 and 𝐵.

Theorem 2.2 (properties of addition and subtraction)

Let 𝐴, 𝐵, 𝐶, 𝟎 ∈ 𝑀(𝑚, 𝑛) and 𝑘, 𝑙 ∈ ℝ. Then:
1. 𝐴 + 𝐵 = 𝐵 + 𝐴 commutativity of addition
2. 𝐴 + (𝐵 + 𝐶) = (𝐴 + 𝐵) + 𝐶 associativity of addition
3. 𝐴+𝟎=𝐴 the zero matrix is an additive identity
4. 𝑘 ∙ (𝐴 + 𝐵) = (𝑘 ∙ 𝐴) + (𝑘 ∙ 𝐵) scalar multiplication is distributive over matrix addition
5. (𝑘 + 𝑙) ∙ 𝐴 = (𝑘 ∙ 𝐴) + (𝑙 ∙ 𝐴) scalar multiplication is distributive over the addition of numbers
6. (𝑘 ∙ 𝑙) ∙ 𝐴 = 𝑘 ∙ (𝑙 ∙ 𝐴)
7. 𝑡𝑟(𝐴 + 𝐵) = 𝑡𝑟(𝐴) + 𝑡𝑟(𝐵)

 Matrix multiplication (product of two matrices)

Caution ! We DO NOT multiply two matrices by multiplying their corresponding members.

Matrix multiplication has certain rule: in the multiplication of two matrices, the matrices 𝐴 and 𝐵 can only
be multiplied together to form the product 𝐴 ∙ 𝐵 if the number of columns in 𝐴 is the same as the number of
rows in 𝐵, regardless of the other dimensions.

Remarks Let 𝐴 ∈ 𝑀(𝑚, 𝑛), 𝐵 ∈ 𝑀(𝑘, 𝑙).

1) 𝐴 ∙ 𝐵 exists if and only if 𝑛 = 𝑘.
2) 𝐴 ∙ 𝐵 ∈ 𝑀(𝑚, 𝑙).

 Beata Ciałowicz 11
 Mathematical Analysis and Linear Algebra - lecture – Section 2 – Matrix Algebra
A product of row matrix and column matrix:
𝑏11
𝑏
If 𝐴 = [𝑎11 𝑎12 … 𝑎1𝑚 ] is a row matrix (1 × 𝑚) and 𝐵 = [ 21 ] is a column matrix (𝑚 × 1), then we
⋮
𝑏𝑚1
define product: 𝐴 ∙ 𝐵 = [𝑐11 ] to be 1 × 1 matrix (a number), where
𝑐11 = 𝑎11 𝑏11 + 𝑎12 𝑏21 + 𝑎13 𝑏31 + ⋯ + 𝑎1𝑚 𝑏𝑚1 .
𝐵 ∙ 𝐴 = [𝑐𝑖𝑗 ] to be 𝑚 × 𝑚 matrix, where
𝑐𝑖𝑗 = 𝑏𝑖1 ∙ 𝑎1𝑗 .
The Falck’s diagrams:
𝑏11
𝑏
𝐵= [ 21 ] 𝐴 ∙ 𝐵 = 𝐴 ∘ 𝐵 𝑇 (inner product of two vectors)
⋮
𝑏𝑚1 𝑚×1
𝐴 = [𝑎11 𝑎12 … 𝑎1𝑚 ]1×𝑚 [𝑐11 ]1×1 = 𝐴 ∙ 𝐵

𝐴 = [𝑎11 … 𝑎1𝑗 … 𝑎1𝑚 ]1×𝑚

𝑏11 𝑐11 … … … 𝑐1𝑚
⋮ ⋮ ⋱ ⋮ ⋱ ⋮
𝑏
𝐵 = 𝑖1 … … 𝑐𝑖𝑗 … … =𝐵∙𝐴
⋮ ⋮ ⋮ ⋱ ⋮
[𝑏𝑚1 ]𝑚×1 [𝑐𝑚1 … 𝑐𝑚𝑗 … 𝑐𝑚𝑚 ]𝑚×𝑚

Example 2.5 a) For given matrices 𝐴 and 𝐵 compute 𝐴 ∙ 𝐵 and 𝐵 ∙ 𝐴.

 Beata Ciałowicz 12
 Mathematical Analysis and Linear Algebra - lecture – Section 2 – Matrix Algebra
A product of two matrices:
If 𝐴 = [𝑎𝑖𝑗 ]𝑚×𝑝 is an 𝑚 × 𝑝 matrix and 𝐵 = [𝑏𝑖𝑗 ]𝑝×𝑛 is a 𝑝 × 𝑛 matrix, then their product

𝐴 ∙ 𝐵 = [𝑐𝑖𝑗 ]𝑚×𝑛 is an 𝑚 × 𝑛 matrix with elements 𝑐𝑖𝑗 , such that:

𝑐𝑖𝑗 = 𝑎𝑖1 ⋅ 𝑏1𝑗 + 𝑎𝑖2 ⋅ 𝑏2𝑗 + 𝑎𝑖3 ⋅ 𝑏3𝑗 + ⋯ + 𝑎𝑖𝑝 ⋅ 𝑏𝑝𝑗

(𝑐𝑖𝑗 is an inner product of 𝑖th row of the first matrix and 𝑗th column of the second matrix: 𝑐𝑖𝑗 = 𝑟𝑖𝐴 ∘ 𝑐𝑗𝐵 ).

The Falck’s diagram:

𝑏11 … 𝑏1𝑗 … 𝑏1𝑛
𝑏21 … 𝑏2𝑗 … 𝑏2𝑛
𝐵=
⋮ ⋮ ⋱ ⋮
𝑏
[ 𝑝1 … 𝑏𝑝𝑗 … 𝑏𝑝𝑛 ]𝑝×𝑛
… … … … 𝑐11 … … … 𝑐1𝑛
𝑎𝑖1 𝑎𝑖2 … 𝑎𝑖𝑝 … … 𝑐𝑖𝑗 … …
𝐴=[ ⋮ ⋮ ⋱ ⋮ ] [ ⋮ ⋮ ⋱ ⋮ ] =𝐴∙𝐵
𝑎𝑚1 𝑎𝑚2 … 𝑎𝑚𝑝 𝑐𝑚1 … 𝑐𝑚𝑗 … 𝑐𝑚𝑛 𝑚×𝑛
𝑚×𝑝

Example 2.5 b) For given matrices 𝐴 and 𝐵 compute 𝐴 ∙ 𝐵 and 𝐵 ∙ 𝐴.

Theorem 2.3 (properties of matrix multiplication)

Let 𝐴 ∈ 𝑀(𝑚, 𝑛), 𝐵 ∈ 𝑀(𝑛, 𝑝), 𝐶 ∈ 𝑀(𝑝, 𝑟) , 𝐼 ∈ 𝑀(𝑛, 𝑛). Then:
1. 𝐴∙𝐵 ≠𝐵∙𝐴
2. 𝐴 ∙ (𝐵 ∙ 𝐶) = (𝐴 ∙ 𝐵) ∙ 𝐶 associativity of multiplication
3. 𝐼∙𝐴=𝐴∙𝐼 =𝐴 identity matrix is a multiplication identity
4. 𝑡𝑟(𝐴 ∙ 𝐵) = 𝑡𝑟(𝐵 ∙ 𝐴)

 Power of a matrix
Remark If 𝐴 ∈ 𝑀(𝑛, 𝑛)is a square matrix, we can define powers of it: 𝐴2 = 𝐴 ∙ 𝐴, 𝐴3 = 𝐴2 ∙ 𝐴, and so on.
In general for any positive integer 𝑘 we have 𝐴𝑘+1 = 𝐴𝑘 ∙ 𝐴, where we take 𝐴0 = 𝐼𝑛 .
Example 2.6 For a matrix 𝐴 compute 𝐴2 .

 Beata Ciałowicz 13
 Mathematical Analysis and Linear Algebra - lecture – Section 2 – Matrix Algebra
2.3 Determinant of a square matrix and its properties
Let 𝐴 ∈ 𝑀(𝑛, 𝑛).
A determinant is a number uniquely associated with a square matrix.
𝑎11 𝑎12 … 𝑎1𝑛
𝑎21 𝑎22 … 𝑎2𝑛
We will denote the determinant of the matrix 𝐴 by det 𝐴 or |𝐴| or | ⋮ ⋮ ⋱ ⋮ |
𝑎𝑛1 𝑎𝑛2 … 𝑎𝑛𝑛 𝑛×𝑛

Caution! Be careful not to confuse the notation for a matrix and that for a determinant. The symbol [… ]
(brackets) is used for a matrix; the symbol |… | (vertical bars) is used for the determinant of the matrix.

Remarks
 Determinant of a matrix 𝟏 × 𝟏 (a first-order determinant)
𝐴 = [𝑎11 ] det 𝐴 = 𝑎11
 Determinant of a matrix 𝟐 × 𝟐 (a second-order determinant)
𝑎11 𝑎12
𝐴 = [𝑎 ] det𝐴 = 𝑎11 𝑎22 −𝑎12 𝑎21
21 𝑎22
 Determinant of a matrix 𝟑 × 𝟑 (a third-order determinant)
𝑎11 𝑎12 𝑎13
𝐴 = [𝑎21 𝑎22 𝑎23 ] (the Sarrus’ rule – ONLY FOR A MATRIX 𝟑 × 𝟑)
𝑎31 𝑎32 𝑎33
𝑎11 𝑎12 𝑎13
𝑑𝑒𝑡𝐴 = | 21 𝑎22 𝑎23 | =
𝑎
𝑎31 𝑎32 𝑎33
𝑎11 𝑎12 𝑎13
𝑎21 𝑎22 𝑎23
= (𝑎11 𝑎22 𝑎33 + 𝑎12 𝑎23 𝑎31 + 𝑎21 𝑎32 𝑎13 ) − (𝑎13 𝑎22 𝑎31 + 𝑎12 𝑎21 𝑎33 + 𝑎23 𝑎32 𝑎11 )
Example 2.8 Evaluate the determinant of the matrix 𝐴 if:
2 5
a) 𝐴 = [ ]
1 3

2 −3 4
b) 𝐵 = [1 5 −2].
4 2 6

 Beata Ciałowicz 14
 Mathematical Analysis and Linear Algebra - lecture – Section 2 – Matrix Algebra
Definition 2.2
1. The minor 𝑀𝑖𝑗 of dimension 𝑛 − 1 (subdeterminant) belonging to the element 𝑎𝑖𝑗 of a square matrix 𝐴
of order 𝑛 ≥ 2 is the determinant of the matrix obtained by deleting the 𝑖th row and the 𝑗th column of 𝐴.
2. The cofactor 𝐶𝑖𝑗 of an element 𝑎𝑖𝑗 of a square matrix 𝐴 is given by
𝐶𝑖𝑗 = (−1)𝑖+𝑗 ∙ 𝑀𝑖𝑗 where 𝑀𝑖𝑗 is the minor of 𝑎𝑖𝑗 .

−2 3 0
Example 2.9 Given the determinant 𝑑𝑒𝑡 [ 5 1 −2] write the minor and cofactor of each of the
7 −4 8
following elements: 𝑎11 , 𝑎23 , 𝑎31 .
Solution:

Remark The matrix 𝐴𝐶 formed by all the cofactors of the elements in a matrix 𝐴 = [𝑎𝑖𝑗 ]𝑛×𝑛 is called the
cofactor matrix 𝐴: 𝐴𝐶 = [𝐶𝑖𝑗 ]𝑛×𝑛 .

Definition 2.3 (determinant of a matrix 𝒏 × 𝒏)

1. 𝑛 = 1
A first-order determinant is defined as follows: 𝐴 = [𝑎11 ] det𝐴 = 𝑎11
2. 𝑛 > 1
The value of a determinant may be found by multiplying every element in any row (or column) by its cofactor
and summing the products. This method is called expanding by cofactors (or Laplace expansion rule):
For the 𝑟th row of 𝐴, the determinant of 𝐴 is:
𝑎11 𝑎12 … 𝑎1𝑗 … 𝑎1𝑛
𝑎21 𝑎22 … 𝑎2𝑗 … 𝑎2𝑛
⋮ ⋮ ⋱ ⋮ ⋱ ⋮
det𝐴 = det 𝑎 𝑎 ⋯ 𝑎 𝑟𝑗 ⋯ 𝑎 𝑟𝑛
= 𝑎𝑟1 ∙ 𝐶𝑟1 + 𝑎𝑟2 ∙ 𝐶𝑟2 + ⋯ + 𝑎𝑟𝑛 ∙ 𝐶𝑟𝑛 .
𝑟1 𝑟2
⋮ ⋮ ⋱ ⋮ ⋱ ⋮
𝑎
[ 𝑚1 𝑎 𝑚2 ⋯ 𝑎 𝑚𝑗 ⋯ 𝑎 𝑚𝑛 ]

For the 𝑟th column of 𝐴, the determinant of 𝐴 is:

𝑎11 𝑎12 … 𝑎1𝑟 … 𝑎1𝑛
𝑎21 𝑎22 … 𝑎2𝑟 … 𝑎2𝑛
⋮ ⋮ ⋱ ⋮ ⋱ ⋮
det𝐴 = det 𝑎𝑖1 𝑎𝑖2 ⋯ 𝑎𝑖𝑟 ⋯ 𝑎𝑖𝑛 = 𝑎1𝑟 ∙ 𝐶1𝑟 + 𝑎2𝑟 ∙ 𝐶2𝑟 + ⋯ + 𝑎𝑛𝑟 ∙ 𝐶𝑛𝑟 .
⋮ ⋮ ⋱ ⋮ ⋱ ⋮
[𝑎𝑚1 𝑎𝑚2 ⋯ 𝑎𝑚𝑟 ⋯ 𝑎𝑚𝑛 ]

Remark
𝑎11 𝑎12 1+1 1+2
If 𝐴 = [𝑎
21 𝑎22 ] then det𝐴 = 𝑎11 𝐶11 +𝑎12 𝐶12 = 𝑎11 (−1) 𝑎22 +𝑎12 (−1) 𝑎21 = 𝑎11 𝑎22 −𝑎12 𝑎21 .

 Beata Ciałowicz 15
 Mathematical Analysis and Linear Algebra - lecture – Section 2 – Matrix Algebra
3 4 0 4
Example 2.10a Calculate a determinant of the matrix 𝐴 = [ 0 2 −2 0]
−2 1 7 0
5 −5 0 −1

Theorem 2.4 (properties of the determinants of matrices)

Let 𝐴, 𝐵, 𝐼 ∈ 𝑀(𝑛, 𝑛) and 𝑘 ∈ ℝ. Then:
1. det𝐼 = 1
2. det𝐴𝑇 = det 𝐴
3. det(𝑘𝐴) = k n det 𝐴
4. det(𝐴 ∙ 𝐵) = det 𝐴 ∙ det𝐵
5. If 𝐴 is a diagonal or a triangular matrix then det 𝐴 = 𝑎11 ∙ 𝑎22 ∙ … ∙ 𝑎𝑛𝑛
6. If any of two rows (columns) of 𝐴 are interchanged, the determinant of the new matrix equals −det 𝐴.
7. The determinant is zero if:
a) two rows (two columns) of 𝐴 are identical,
b) a matrix has a row (column) containing all zeros,
c) two rows (two columns) of a matrix are proportional.
8. If the result of multiplying a row (column) of 𝐴 by a constant is added to another row (column) of 𝐴, the
determinant of the new matrix is equal to det 𝐴.
 Beata Ciałowicz 16
 Mathematical Analysis and Linear Algebra - lecture – Section 2 – Matrix Algebra
2.4 Elementary row or column operations
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: before expanding by cofactors we may transform a determinant with the
help of elementary transformations into a form such that it contains as many zeros as possible or into a
triangular matrix.
2. Finding the inverse of a matrix (the Gauss-Jordan reduction method).
3. Calculating the rank of a matrix (Section 3).
4. Solving general system of linear equations (the Gauss and the Gauss-Jordan elimination methods -
Section 3).
Properties:
 Operation 1) changes the determinant of a matrix (det 𝐴′ = − det 𝐴).
 Operation 2) changes the determinant of a matrix (det 𝐴′ = 𝑘 det 𝐴).
 Operation 3) does NOT change the determinant of a matrix (det 𝐴′ = det 𝐴).
3 4 0 4
Example 2.10b Calculate a determinant of the matrix 𝐴 = [ 0 2 −2 0]
−2 1 7 0
5 −5 0 −1

 Beata Ciałowicz 17
 Mathematical Analysis and Linear Algebra - lecture – Section 2 – Matrix Algebra
2.5 Inverse of a matrix

Definition 2.5. Let 𝐴 ∈ 𝑀(𝑛, 𝑛). A matrix 𝐴−1 ∈ 𝑀(𝑛, 𝑛) is called an inverse matrix of 𝐴 with respect
to multiplication (in short: inverse of 𝐴), iff: 𝐴 ∙ 𝐴−1 = 𝐴−1 ∙ 𝐴 = 𝐼.
Remarks
 A square matrix 𝐴 has at most one inverse.
 A square matrix 𝐴 which has the inverse matrix is called invertible (or non-singular).
 If det 𝐴 = 0 then a matrix 𝐴 has no inverse (it is called singular).
Caution ! Not all square matrices are invertible.

Remark A square matrix 𝐴 is invertible iff det 𝐴 ≠ 0.

There are many ways of calculating the inverse of square matrix, if it exists. One of these method is the
cofactor method.

 Inverse matrix using cofactors

Theorem 2.5. Let 𝐴 ∈ 𝑀(𝑛, 𝑛), det 𝐴 ≠ 0 and 𝐴𝐶 ∈ 𝑀(𝑛, 𝑛) be its cofactor matrix. Then 𝐴 is invertible
1
and 𝐴−1 = [𝐴𝐶 ]𝑇 .
𝑑𝑒𝑡𝐴

Remark [𝐴𝐶 ]𝑇 is called the adjoint (or adjugate) matrix.

Procedure for calculating the inverse 𝐴−1 of a matrix 𝐴 using the cofactors:
a) check if det 𝐴 ≠ 0,
b) find the cofactor of each element,
c) replace each element by its cofactor,
d) find the transpose of the matrix found in (c),
1
e) multiply the matrix in (d) by 𝑑𝑒𝑡𝐴. The result is 𝐴−1 .

Example 2.11a If possible find inverse of given matrices:

6 3
𝐾=[ ]
4 2

3 5
𝐿=[ ]
2 4

2 −1 0
𝑀 = [1 −2 3]
4 0 2

 Beata Ciałowicz 18
 Mathematical Analysis and Linear Algebra - lecture – Section 2 – Matrix Algebra
 Inverse matrix using elementary row (column) operations (the Gauss-Jordan
reduction method)
Procedure for calculating the inverse 𝐴−1 of a matrix 𝐴 using elementary row (column) operations:
a) first we form a new matrix consisting, on the left, of the matrix 𝐴, on the right, of the corresponding
identity matrix 𝐼: [𝐴 ∶ 𝐼],
b) now we use elementary row operations to produce matrix:[𝐼 ∶ 𝐴−1 ],
c) we attempt to transform 𝐴 to an identity matrix, but whatever operations we perform, we do on the
identity matrix,
d) the matrix on the right will be 𝐴−1.

2 −1 0
Example 2.11b If possible find inverse of the given matrix 𝑀 = [1 −2 3].
4 0 2

 Beata Ciałowicz 19
 Mathematical Analysis and Linear Algebra - lecture – Section 2 – Matrix Algebra
2.6 Orthogonal matrix
Definition 2.6. The inner product of two vectors 𝑋 = (𝑥1 , 𝑥2 , … , 𝑥𝑛 ), 𝑌 = (𝑦1 , 𝑦2 , … , 𝑦𝑛 ) is a scalar:
𝑋 ∘ 𝑌 = < 𝑋|𝑌 > = 𝑥1 ⋅ 𝑦1 + 𝑥2 ⋅ 𝑦2 + ⋯ + 𝑥𝑛 ⋅ 𝑦𝑛 .

Definition 2.7. A set of vectors {𝑋1 , 𝑋2 , … , 𝑋𝑚 } ⊂ ℝ𝑛 is called:

1. orthogonal if 𝑋𝑖 ∘ 𝑋𝑗 = 0 (𝑋𝑖 ⊥ 𝑋𝑗 ) for each 𝑖, 𝑗 ∈ {1,2, … , 𝑚}, 𝑖 ≠ 𝑗,
2. orthonormal if this set is orthogonal and 𝑋𝑖 ∘ 𝑋𝑖 = 1 for each 𝑖 ∈ {1,2, … , 𝑚}.

Definition 2.8. A matrix 𝐴 is an orthogonal matrix if 𝐴 ⋅ 𝐴𝑇 = 𝐴𝑇 ⋅ 𝐴 = 𝐼.

Theorem 2.6. Let be given matrix 𝐴 ∈ 𝑀(𝑛, 𝑛). This matrix is an orthogonal matrix if and only if:
1. 𝐴−1 = 𝐴𝑇
2. their columns and rows (as elements of ℝ𝑛 ) form an orthonormal set of vectors.

Example 2.12 Determine if the given matrix is orthogonal:

1 0
𝐼=[ ]
0 1
1 0
𝐴=[ ]
0 −1

2.7 Matrix equations

A matrix equation is an equation in which a variable stands for a matrix. In solving a matrix equation, basic
matrix operations and an inverse of a matrix are used. Each matrix equation of the linear type can be reduced
to one of the following three forms:

1) 𝑨 ∙ 𝑿 = 𝑩 where 𝐴 ∈ 𝑀(𝑛, 𝑛), 𝐵 ∈ 𝑀(𝑛, 𝑝), 𝑑𝑒𝑡𝐴 ≠ 0. Solution: 𝑿 = 𝑨−𝟏 ∙ 𝑩, where 𝑋 ∈ 𝑀(𝑛, 𝑝).
2) 𝑿 ∙ 𝑨 = 𝑩 where 𝐴 ∈ 𝑀(𝑛, 𝑛), 𝐵 ∈ 𝑀(𝑝, 𝑛), 𝑑𝑒𝑡𝐴 ≠ 0. Solution: 𝑿 = 𝑩 ∙ 𝑨−𝟏 , where 𝑋 ∈ 𝑀(𝑝, 𝑛).
3) 𝑨 ∙ 𝑿 ∙ 𝑩 = 𝑪 where 𝐴 ∈ 𝑀(𝑛, 𝑛), 𝐵 ∈ 𝑀(𝑝, 𝑝), 𝐶 ∈ 𝑀(𝑛, 𝑝), 𝑑𝑒𝑡𝐴 ≠ 0, 𝑑𝑒𝑡𝐵 ≠ 0.
Solution: 𝑿 = 𝑨−𝟏 ∙ 𝑪 ∙ 𝑩−𝟏 .

4) 𝑨 ∙ 𝑿 = 𝑩 where 𝐵 ∈ 𝑀(𝑛, 𝑝), 𝐴 ∈ 𝑀(𝑛, 𝑛) and 𝑑𝑒𝑡𝐴 = 0 or 𝐴 ∉ 𝑀(𝑛, 𝑛).

Solution: We search the solution by substituting the form of the general matrix 𝑋 with specific dimensions for
which the multiplication 𝐴 ∙ 𝑋 is feasible and the dimensions of the resulting product were the same as the
dimensions of matrix 𝐵. After performing operations on matrices, we may find the solution by solving a
system of linear equations.
Remark. An equation of form 1) we can solve by substituting the form of the general matrix 𝑋 with specific
dimensions as in equations of type 4).

 Beata Ciałowicz 20
 Mathematical Analysis and Linear Algebra - lecture – Section 2 – Matrix Algebra
1 −2 4 1
Example 2.13 Given 𝐴 = [ ] 𝐵=[ ] determine matrices 𝑋 and 𝑌 satisfying:
3 4 0 −3
a) 𝐴 ∙ 𝑋 = 𝐵, b) 𝑌 ∙ 𝐴 = 𝐵.

𝑇
2 −3 3 4𝑇
Example 2.14. Solve the given equation for the matrix 𝑋: ([ ] ∙ 𝑋 + 3𝐼) = [ ] ∙ [ ] − 3𝐼 2 .
−1 4 −2 1

 Beata Ciałowicz 21