CHAPTER 1

MATRICES, VECTORS, AND

SYSTEMS OF LINEAR EQUATIONS
1.1 Matrices and Vectors
In many occasions, we can arrange a number of values of interest
into an rectangular array. For example:
Example

We can represent the information

on 3July sales more simply as
2

6
4 15
45

8
20 5 .
64

elements of R, the set

of real numbers
Definitions
A matrix is a rectangular array of scalars.
If the matrix has m rows and n columns, we say that the size of the matrix is
m by n, written m n.
The matrix is called square if m = n.
The scalar in the ith row and jth column is called the (i, j)-entry of the matrix.

Notation:
2

a11
6
A = 4 ...
am1

..
.

a1n
.. 7 = [a ] 2 M
ij
mn
. 5
amn

Example:

We use Mmn to denote the set that contains all matrices whose sizes are
m n.
2

Equality of matrices
equal: We say that two matrices A and B are equal if they

have the same size and have equal corresponding entries.

Let A, B 2 Mmn .
Then A = B , aij = bij , 8i = 1, . . . , m, j = 1, . . . , n.

Example

Submatrices
submatrix: A submatrix is obtained by deleting from a

matrix entire rows and/or columns.

For example,
E=

15
45

20
64

is a submatrix of

6
B = 4 15
45

8
20 5
64

Matrix addition
Sum of matrices

Definition

Example

1
4 3
5

2 3
2
1 1
4 5+4 1 1 5=4 4 5 5
6 8
6
1 2

Scalar multiplication
Definition

Example

2 3
6
9
3 4 1 4 5 = 4 3 12 5
1 5
3 15

Zero matrices
zero matrix: matrix with all zero entries, denoted by O (any

size) or Omn.
For example, a 2-by-3 zero matrix can be denoted
O23 =

0 0 0
0 0 0

Property
A = O + A for all A

Property
0 A = O for all A
7

Question
Let A =

1
3

2
4

2 M22

3
1 2
and B = 4 3 4 5 2 M32 .
5 6

Then both 0 A and 0 B can be denoted by O, that is,

0 A = O, and 0 B = O.

Can we conclude that

0 A = 0 B?

Matrix Subtraction
Definition

Example 2 2

4 1
1

3
2 3
2
4 5 = ( 1) 4 1 4 5 = 4 1
5
1 5
1

3
4 5
5

Question
For any m n matrices A and B (i.e., 8A, B 2 Mmn ), will
A+B =B+A
always be true?

Question
For any m n matrices A and B (i.e., 8A, B 2 Mmn ) and any real number s
(i.e., 8s 2 R), will
s(A + B) = sA + sB
always be true?

10

Question
For any m n matrices A and B (i.e., 8A, B 2 Mmn ), will
A+B =B+A
always be true?

Answer: Yes! It is always true.

Proof:

11

Theorem 1.1

(Properties of Matrix Addition and Scalar Multiplication)

Proof: All proofs can follow from basic arithmetic laws in R and
previous definitions. Please do all of them yourself (homework).
By (b), sum of multiple matrices are written as A + B + + M

12

Transpose
Definition

Property
C 2 Mmn ) C T 2 Mnm

Example

7
C = 4 18
52

9
7 18 52
31 5 ) C T =
9 31 68
68

Question
Is C = C T always wrong?

Question
Is 8A, B 2 Mmn , (A + B)T = AT + B T always true?
13

Theorem 1.2

Proof:

14

(Properties of the Transpose)

Vectors
A row vector is a matrix with one row.

A column vector is a matrix with one column.

2

1
6 2 7

6
7 or 1
4 3 5
4

The term vector can refer to either a row vector or a

column vector.
(Important) In this course, the term vector always refers to
a column vector unless being explicitly mentioned
otherwise.
15

Vectors

Rn :We denote the set of all column vectors with n entries

by Rn .
In other words,

Rn = Mn1
components: the entries of a vector.
Let v 2 Rn and assume

6
6
v=6
4

v1
v2
..
.
vn

7
7
7.
5

Then the ith component of v refers to vi .

16

Vector Addition and Scalar Multiplication

Definitions of vector addition and scalar multiplication of

vectors follow those for matrices.

0 is the zero vector (any size), and u + 0 = u, 0u = 0 for all

u Rn.
A matrix is often regarded as a stack of row vectors or a cross list of column
vectors. For any C 2 Mmn , we can write

C = c1 cj cn
2

6
6
where cj = 6
4
17

c1j
c2j
..
.

cmj

3
7
7
7
5

Geometrical Interpretations
Vectors for geometry in R2

in R3

18

vector addition

scalar multiplication
for a vector

Section 1.1 (Review)

Matrix
Rows and columns
Size (m-by-n).
Square matrix.
(i,j)-entry
Matrix
Equality,
Addition, Zero Matrix
Scalar multiplication, subtraction

Vector
Row vectors, column vectors.
components

19

1.2 Linear Combinations, Matrix-Vector Products, and Special Matrices

Definition

Example:

2
8

1
1

+4

1
3

+1

1
1

Given the coefficients ({-3,4,1}), it is easy to compute the

combination ([2 8]T), but the inverse problem is harder.
Example:

20

4
1

= x1

2
3

+ x2

3
=
1

2x1
3x1

3x2
1x2

2x1 + 3x2
3x1 + x2

To determine x1 and x2, we must solve a system of linear equations,

which has a unique solution [x1 x2]T = [-1 2]T in this case.

Geometrical view point:

manage to form a parallelogram

Example: to determine if [-4 -2]T is a linear combination of [6 3]T

and [2 1]T, we must solve
6x1 + 2x2

3x1 + x2

which has infinitely many solutions, as the geometry

suggests.

21

Example: to determine if [3 4]T is a linear combination of [3 2]T

and [6 4]T, must solve
3x1 + 6x2

2x1 + 4x2

which has no solutions, as the geometry suggests.

22

If u and v are any nonparallel vectors in R2, then every vector in

R2 is a linear combination of u and v (unique linear combination).
algebraically, this means that u and v are
nonzero vectors, and u cv.
What is the condition in R3? in Rn?

23

Standard vectors
The standard vectors of Rn are defined as
2

6
6
e1 = 6
4

1
0
..
.
0

7
6
7
6
7 , e2 = 6
5
4

0
1
..
.
0

7
6
7
6
7 , , en = 6
5
4

0
0
..
.
1

7
7
7.
5

Obviously, every vector in Rn may be uniquely linearly combined

by these standard vectors.

24

Matrix-Vector Product
Definition

Note that we can write:

Example:
2

1
Av = 4 3
5
25

6
6
Av = 6
6 a1
4

1 2
Let A = 4 3 4
5 6
3
2

2
7
4 5
= 74
8
6

5,v =
3

a2

7
8

an

32
76
76
76
74
5

v1
v2
..
.
vn

3
7
7
7
5

. Then Av =?
3

1
2
7
16
23
3 5 + 8 4 4 5 = 4 21 5 + 4 32 5 = 4 53 5
5
6
35
48
83

Property: A0 = 0 and Ov = 0 for any A and v.

Let A 2 M23 and v 2 R3 . Then

2
3

v1
a11 a12 a13 4
v2 5 =
Av =
a21 a22 a23
v3

v1

a11
a21

+ v2

6
6
Av = 6
4

a11
a21
..
.

a12
a22
..
.

a1n
a2n
..
.

am1

am2

amn

32
76
76
76
54

v1
v2
..
.
vn

3
7
7
7
5

6
6
v1 6
4
2
6
6
6
4

a11
a21
..
.
am1

+ v3

a13
a23

6
7
6
7
7 + v2 6
4
5

a12
a22
..
.
am2

7
6
7
6
7 + + v3 6
5
4

a11 v1 + a12 v2 + + a1n vn

a21 v1 + a22 v2 + + a2n vn
..
.

am1 v1 + am2 v2 + + amn vn

The ith component of Av is

26

a12
a22

a11 v1 + a12 v2 + a13 v3

a21 v1 + a22 v2 + a23 v3

More generally, when A 2 Mmn and v 2 Rn . Then

2

ai1

ai2

a1n

3
7
7
7
5

6
6
6
4

a1n
a2n
..
.
amn

v1
v2
..
.
vn

3
7
7
7
5

3
7
7
7
5

Identity Matrix
Definition

Example:

1
I3 = 4 0
0

0
1
0

3
0
0 5
1

Sometimes In is simply written as I (any size).

Property:

27

In v = v for any v 2 Rn

Stochastic Matrix
Definition
An n n matrix A 2 Mnn is called a stochastic matrix if all entries of A
are nonnegative and the sum of all entries in each column is unity.

Example:

A=

28

0.85
0.15

0.03
0.97

is a 2 2 stochastic matrix.

Example: stochastic matrix

From
City Suburbs

probability matrix of a sample

.85 .03
To
City
= A persons residence movement
.15 .97
Suburbs

500
p=
: current population of the city and suburbs
700
Suburbs
700 thousand

City
500 thousand

This year

15%
85%

Next year

Ap =

29

A(Ap)

0.85 x 500 + 0.03 * 700 = 446

City

.85 .03
.15 .97

500
700

3%

97%

0.15 x 500 + 0.97 x 700 = 754

Suburbs

: population distribution in the next year

: population distribution in the year following the next

Example: rotation matrix

A =

cos
sin

sin
cos

P =

x0
y0
P =

x
y

=
=
=
=

30

cos
sin

sin
cos

x cos y sin
x sin + y cos

x0
y0

cos
x
sin

+y

x
y

sin
cos

x
y

Question
Is the statement
(A + B)u = Au + Bu, 8A, B 2 Mmn , u 2 Rn
always true?

Question
Let A 2 Mmn and ej be the jth standard vector in Rn . Then what is
Aej ?

31

Theorem 1.3

(Properties of Matrix-Vector Products)

Proof for (e): If B A, then (B - A)ej 0, i.e., Bej Aej, for some j.

32

Problems for practice (1.1~1.2)

Section 1.1: Problems 1, 5, 7, 11, 13, 19, 25, 27, 29, 31, 37,
39, 41, 43, 45, 51, 53, 55.
Section 1.2: Problems 3, 5, 8, 9, 15, 34, 42, 44, 83-87

33