You are on page 1of 40

Chapter 3: VECTOR SPACES

AM
PH
Duong T. PHAM

T.
g.
APPLIED LINEAR ALGEBRA
on
Du

Duong T. PHAM October 16, 2019 1 / 40


Outline

1 Vector spaces

AM
2 Linear independence

PH
3 Bases and Dimension

T.
4 Linear Subspaces g.
on
5 Rank of matrices
Du

6 Null space - Solution spaces of homogeneous linear systems

Duong T. PHAM October 16, 2019 2 / 40


Vector spaces

Definition: A vector space over R is a set V together with two


operations: vector addition and scalar multiplication that satisfy:

AM
1 Associativity of addition: u + (v + w ) = (u + v ) + w
2 Commutativity of addition: u + v = v + u

PH
3 Identity element of addition: ∃0 ∈ V s.t. u + 0 = u ∀u ∈ V (Here,
0 is called zero vector of V )

T.
4 Inverse elements of addition: ∀u ∈ V , ∃(−u) ∈ V s.t. u + (−u) = 0
g.
on
5 Compatibility: α(βu) = (αβ)u ∀u ∈ V , ∀α, β ∈ R
Du

6 Identity element of scalar multiplication: 1u = u ∀u ∈ V


7 Distributivity: α(u + v ) = αu + αv ∀u, v ∈ V , ∀α ∈ R
8 Distributivity: (α + β)u = αu + βu ∀α, β ∈ R, ∀u ∈ V

Duong T. PHAM October 16, 2019 3 / 40


Vector spaces

Ex. V = Rn = {(x1 , x2 , . . . , xn )| xi ∈ R}. For any x, y ∈ Rn ;


x = (x1 , x2 , . . . , xn ); y = (y1 , y2 , . . . , yn )

AM
Addition: x + y = (x1 + y1 , x2 + y2 , . . . , xn + yn )

PH
Scalar multiplication: α · x = (αx1 , αx2 , . . . , αxn ) α ∈ R

T.
(Rn , +, ·) is a vector space over R.
g.
on
Ex. V = Mm,n (R) with the addition of matrices (+) and the
multiplication between a number and a matrix (·);
Du

(Mm,n (R), +, ·) is a vector space.

Duong T. PHAM October 16, 2019 4 / 40


Vector spaces

Ex. P[x] : set of polynomials with real coefficients;

AM
Pn [x] : set of polynomials with real coefficients and orders ≤ n ;

PH
Addition (+): addition of polynomials
Scalar multiplication (·): multiplication between a real number with a

T.
polynomial;
Then g.
on
(P[x], +, ·) is a vector space
Du

(Pn [x], +, ·) is a vector space

Duong T. PHAM October 16, 2019 5 / 40


Vector spaces

AM
Ex. C [a, b] : set of continuous functions on [a, b]

PH
Addition (+): addition of two functions

T.
Scalar multiplication (·): multiplication between a real number and a
function g.
on
Then (C [a, b], +, ·) is a vector space.
Du

Duong T. PHAM October 16, 2019 6 / 40


Basic Properties

Zero vector is unique


Suppose that there are two zero vector 01 and 02 . Then

AM
u + 01 = 01 + u = u ∀u ∈ V (1)
u + 02 = 02 + u = u ∀u ∈ V (2)

PH
(1) (2)
We have 02 = 01 + 02 = 01

T.
Inverse vector of a vector u is unique
g.
Suppose that there are two inverse vectors u 1 and u 2 of u. Then
on
Du

u 1 + u = u + u 1 = 0 (3)
u 2 + u = u + u 2 = 0 (4)

(3) (4)
We have u 2 = u 2 + 0 = u 2 + u + u 1 = 0 + u 1 = u 1

Duong T. PHAM October 16, 2019 7 / 40


Basic Properties

AM
∀u, v , w ∈ V : u + v = u + w =⇒ v = w

u + v = u + w =⇒ (−u) + (u + v ) = (−u) + (u + w )

PH
=⇒ (−u + u) + v = (−u + u) + w =⇒ 0 + v = 0 + w =⇒ v = w

T.
g.
α · 0 = 0 ∀α ∈ R and 0 · u = 0 ∀u ∈ V
on
0 + α · 0 = α · 0 = α · (0 + 0) = α · 0 + α · 0 =⇒ 0 = α · 0
Du

0 · u + 0 = 0 · u = (0 + 0) · u = 0 · u + 0 · u =⇒ 0 = 0 · u

Duong T. PHAM October 16, 2019 8 / 40


Basic Properties

α=0
α · u = 0 ⇐⇒
u=0

AM
If α = 0 or u = 0, then it is obvious that α · u = 0.

PH
Assume that α · u = 0 and α 6= 0. Then
1 1 1 

T.
· (α · u) = · 0 = 0 ⇐⇒ α ·u =0
α α α
g. ⇐⇒ 1 · u = 0⇐⇒ u = 0
on
(−1) · u = −u
Du

(−1) · u + u = (−1) · u + 1 · u = (−1 + 1) · u = 0 · u = 0


=⇒ (−1) · u = −u

Duong T. PHAM October 16, 2019 9 / 40


Linear Independence
Definition.
Given u, u 1 , u 2 , . . . , u n ∈ V ; we say that u is a linear combination of
u 1 , u 2 , . . . , u n if ∃ α1 , α2 , . . . , αn ∈ R s.t.

AM
u = α1 u 1 + α2 u 2 + . . . + αn u n

PH
Definition.

T.
. . u n ∈ V are said to be linearly dependent if ∃ α1 , α2 , . . . , αn ,
u 1 , u 2 , .P
where αi2 6= 0, satisfying
g.
on
α1 u 1 + α2 u 2 + . . . + αn u n = 0
Du

Definition.
u 1 , u 2 , . . . u n ∈ V are said to be linearly independent if they are not
linearly dependent.
⇐⇒ If α1 u 1 + α2 u 2 + . . . + αn u n = 0 then α1 = α2 = . . . = αn = 0.
Duong T. PHAM October 16, 2019 10 / 40
Linear Independence
Ex. Let u 1 = (1, 2, −1, 0), u 2 = (1, 0, 1, 3), u 3 = (−1, 1, 0, 2).
Decide if u 1 , u 2 , u 3 are linearly independent.
Suppose that α1 u 1 + α2 u 2 + α3 u 3 = 0. Then

AM
       
1 1 −1 0
2 0  1  0

PH
−1 + α2 1 + α3  0  = 0
α1        

0 3 2 0

T.
     
1 1 −1 0 1 1 −1 0 1 1 −1 0
 2 0 1 0 g.
 →
 0 −2 3 0 
 →
 0 −2 3 0 
on
 
 −1 1 0 0   0 2 −1 0   0 0 2 0 
0 13
Du

0 3 2 0 0 3 2 0 0 2 0
 
1 1 −1 0
 0 −2
→
3 0   ⇒ α1 = α2 = α3 = 0. Hence, u 1 , u 2 , u 3
 0 0 2 0  are linearly independent.
0 0 0 0
Duong T. PHAM October 16, 2019 11 / 40
Basic Properties

Basic properties
(
linearly independent ⇔ u 1 6= 0
{u 1 } is

AM
linearly dependent ⇔ u 1 = 0

PH
(
linearly independent ⇔ u 1 , u 2 are Not proportional
{u 1 , u 2 } is
linearly dependent ⇔ u 1 , u 2 are proportional

T.
A set of vectors containing 0 is linearly dependent
g.
on
Let S1 ⊂ S2 be two set of vectors. Then
Du

- S2 : linearly independent ⇒ S1 : linearly independent


- S1 : linearly dependent ⇒ S2 : linearly dependent

Duong T. PHAM October 16, 2019 12 / 40


Basic Properties

Basic Properties

AM
S is linearly dependent ⇒ a vector in S is a linear combination of
some other vectors in S.

PH
Let {u 1 , . . . , u n } is linearly independent;

T.
{u 1 , . . . , u n , u} is linearly independent if u is NOT a linear
g.
combination of u 1 , . . . , u n .
on
{u 1 , . . . , u n , u} is linearly dependent if u is a linear combination of
Du

u 1, . . . , u n.

Duong T. PHAM October 16, 2019 13 / 40


Basis
Definition.
Let V be a vector space; B be a subset of V . The set B is a basis of V
if:

AM
B is linearly independent : ∀v 1 , . . . , v n ∈ B; ∀α1 , . . . , αn ∈ R
α1 v 1 + . . . αn v n = 0 =⇒ α1 = . . . = αn = 0

PH
B is a spanning set of V :
∀v ∈ V ; ∃v 1 , . . . , v n ∈ B and ∃α1 , . . . , αn ∈ R s.t.

T.
v = α1 v 1 + . . . + αn v n
g.
on
{(1, 0), (0, 1)} is a basis of R2 since:
Du

It is linearly independent;
Every (x, y ) ∈ R2 is a linear
combination of (1, 0) and (0, 1), i.e.
(x, y ) = x(1, 0) + y (0, 1)
Duong T. PHAM October 16, 2019 14 / 40
Dimension

AM
Definition.

PH
Let V be a vector space; B be a basis of V .
If B has infinitely many elements, then V is a infinitely dimensional

T.
space
g.
If B has finite number of elements, then dimV := size(B)
on
Du

Duong T. PHAM October 16, 2019 15 / 40


Dimension

Ex. Consider Rn ; e 1 = (1, 0, . . . , 0), . . . , e n = (0, 0, . . . , 1)


{e 1 , . . . , e n } is linealy independent?

AM
Suppose that x1 , . . . , xn ∈ R such that x1 e 1 + . . . + xn e n = 0
  
1 0 ... 0 0

PH
x1 = 0


 0 1 ... 0 0  ..
=⇒   ... ... ... ... ...  =⇒ .

T.

x = 0
0 0 ... 1 0 n
g.
on
{e 1 , . . . , e n } spans Rn ? Because every (x1 , . . . , xn ) ∈ Rn ;
(x1 , . . . , xn ) = x1 e 1 + . . . + xn e n
Du

=⇒ {e 1 , . . . , e n } is a basis of Rn

dimRn = n

Duong T. PHAM October 16, 2019 16 / 40


Basis and Dimension

Ex. Consider Mm,n (R). (


Eijij = 1
Denote E ij ∈ Mm,n (R) defined by ij
if (k, `) 6= (i, j)

AM
Ek` =0
m X
X n
αij E ij = 0m,n =⇒ αij = 0 for all i = 1, m and j = 1, n

PH
i=1 j=1
=⇒ {E ij : i = 1, m, j = 1, n} is linearly independent.

T.
m X
n
g.
For any A = [aij ] ∈ Mm,n (R); A =
X
aij E ij
on
i=1 j=1
{E ij
Du

=⇒ : i = 1, m, j = 1, n} spans Mm,n (R)

Hence, {E ij : i = 1, m, j = 1, n} is a basis of Mm,n (R)

dimMm,n (R) = size{E ij : i = 1, m; j = 1, n} = mn

Duong T. PHAM October 16, 2019 17 / 40


Basis and Dimension
Ex. Pn [x] is the vector space of all polynomials of degrees ≤ n. Is
{1, x, x 2 , . . . , x n } a basis of Pn [x]?

AM
Assume that a0 1 + a1 x + a2 x 2 + . . . + an x n = 0 ∀x ∈ R
=⇒ a0 = a1 = . . . = an = 0.

PH
Hence, {1, x, . . . , x n } is linearly independent

Let f (x) ∈ Pn [x]. Then f is of the form

T.
f (x) = a0 + a1 x + . . . + an x n
g.
on
for some a0 , a1 , . . . , an ∈ R.
Du

Hence, {1, x, . . . , x n } spans Pn [x]

Therefore, {1, x, . . . , x n } is a basis of Pn [x]

and dimPn [x] = n + 1

Duong T. PHAM October 16, 2019 18 / 40


Basis and Dimension

AM
Ex. P[x] is the vector space of all polynomials with real coefficients.

PH
P[x] is a infinite dimensional vector space.

T.
{1, x, x 2 , . . . , x n , . . . , } is a basis of P[x]

dimP[x] = ∞ g.
on
Du

Duong T. PHAM October 16, 2019 19 / 40


Basis and Dimension

AM
Theorem.
Let V be a n-dimensional vector space. There hold

PH
A set which contains more than n vectors is linearly dependent
A set S which contains n vectors is a basis if and only if S satisfies at

T.
least one of the two following conditions:
- S is linearly independent;
g.
- S spans V ;
on
A set S of k linearly independent vectors ( k ≤ n ) can be added
Du

n − k vectors to make it a basis of V .

Duong T. PHAM October 16, 2019 20 / 40


Coordinates with respect to a basis

Definition.
Let V be a n-dimensional vector space and {u 1 , . . . , u n } be a basis. Every
u ∈ V can uniquely be written as

AM
u = x1 u 1 + . . . + xn u n .

PH
Then (x1 , . . . , xn ) is called coordinate of u with respect to {u 1 , . . . , u n }

T.
Notation: We write u/{u 1 ,...,u n } = (x1 , . . . , xn )
g.
on
Ex. In R2 , the set {(1, 1), (1, 2)} is a basis.
Du

2 ; Suppose (−2, 3) = x(1, 1) + y (1, 2)


( vector (−2, 3) ∈ R (
Consider
x + y = −2 x = −7
=⇒ =⇒
x + 2y = 3 y =5

(−2, 3)/{(1,1),(1,2)} = (−7, 5)


Duong T. PHAM October 16, 2019 21 / 40
Basis Transformation Matrix

Let U = {u 1 , u 2 , . . . , u n } be a basis of V . Let x ∈ V such that


x/U = (x1 , . . . , xn ) ⇐⇒ x = x1 u 1 + . . . + xn u n

AM
 
x1
  . 
⇐⇒ u 1 . . . u n  ..  = x

PH
xn

T.
Let U = {u 1 , . . . , u n } and V = {v 1 , . . . , v n } be two bases of V ;
g.
v 1 = a11 u 1 + a12 u 2 + . . . + a1n u n
on
v 2 = a21 u 1 + a22 u 2 + . . . + a2n u n
Du

..................................................
v n = an1 u 1 + an2 u 2 + . . . + ann u n

Duong T. PHAM October 16, 2019 22 / 40


Basis Transformation Matrix
 
a11 a21 . . . an1
a12 a2n . . . an2  is the basis transformation
Matrix: TUV =
 ... ...

...  matrix from U into V .

AM
a1n a2n . . . ann

PH
Proposition.
TUV is the basis transformation matrix from U to V ⇐⇒

T.
   
u 1 u 2 . . . u n TUV = v 1 v 2 . . . v n
g.
on
Remark:
−1 
Du

 
TUV = u 1 u 2 . . . u n v1 v2 . . . vn
| {z }| {z }
written in columns written in columns

−1
TV U = TUV

Duong T. PHAM October 16, 2019 23 / 40


Basis Transformation Matrix

Theorem.
Two bases of V : U = {u 1 , . . . , u n } and V = {v 1 , . . . , v n }.
Let u ∈ V and u/U = (x1 , . . . , xn ); u/V = (y1 , . . . , yn );

AM
We have    
x1 y1

PH
 ..   .. 
 .  = TUV  .  or u/U = TUV · u/V

T.
xn yn

Proof.   g.    
x1 y1 y1
on
  .    .     .. 
u 1 ... u n  ..  = u = v 1 ... v n  ..  = u 1 ... u n TUV
Du

.
xn yn yn
   
x1 y1
 ..   .. 
=⇒  .  = TUV  .  
xn yn
Duong T. PHAM October 16, 2019 24 / 40
Linear Subspace

Definition.
Let (V , +, ·) be a vector space and U ⊂ V . If U satisfies:
U is closed under addition : ∀u, v ∈ U, u + v ∈ U,

AM
U is closed under scalar multiplication : ∀u ∈ U, ∀α ∈ R, αu ∈ U,

PH
then (U, +, ·) is a vector space and called subspace of V .

T.
Ex. Let U1 = {(x1 , x2 , x3 ) ∈ R3 | x1 + x2 + x3 = 0} and
g.
Note that (R3 , +, ·) is a vector space
on
U1 is a subspace of R3 since:
Du

Let x = (x1 , x2 , x3 ), y = (y1 , y2 , y3 ) ∈ U1


=⇒ x1 + x2 + x3 = 0 and y1 + y2 + y3 = 0
=⇒ (x1 + y1 ) + (x2 + y2 ) + (x3 + y3 ) = 0 =⇒ x + y ∈ U1
Let x = (x1 , x2 , x3 ) ∈ U1 and α ∈ R
=⇒ x1 + x2 + x3 = 0
=⇒ αx1 + αx2 + αx3 = 0 =⇒ αx ∈ U1
Duong T. PHAM October 16, 2019 25 / 40
Linear Subspaces

AM
Ex. U2 = {(x1 , x2 , x3 ) ∈ R3 | x1 + x2 + x3 ≥ 0} is NOT a subspace of R3

PH
(1, 1, 1) ∈ U2 since 1 + 1 + 1 ≥ 0

T.
But −1(1, 1, 1) ∈
/ U2 since −1 − 1 − 1 = −3 < 0 =⇒ U2 is NOT
closed under scalar multiplication
g.
on
U2 is a not vector space and it is not a subspace of R3 .
Du

Duong T. PHAM October 16, 2019 26 / 40


Linear Subspaces

Theorem.
Let U be a subspace of vector space V and dim V = n. Then

AM
dim U ≤ n

PH
dim U = n ⇐⇒ U = V

T.
Definition.
g.
Let V be a vector space and S = {u 1 , u 2 , . . . , u m }⊂ V . The span of S
on
is the set of all linear combinations of S, i.e.
Du

span(S) = {α1 u 1 + α2 u 2 + . . . + αm u m : αi ∈ R, i = 1, . . . , m}

Notation: We denote the span of S by span(S) or hu 1 , u 2 , ..., u m i

Duong T. PHAM October 16, 2019 27 / 40


Linear Span

Proposition.
Let V be a vector space and S = {u 1 , u 2 , ..., u m }⊂ V . Then span(S) is a

AM
linear subspace of V

PH
Proof. Let a, b ∈ span(S) and α ∈ R. Then
a = a1 u 1 + a2 u 2 + . . . am u m

T.
b = b1 u 1 + b2 u 2 + . . . bm u m
g.
for some ai , bi ∈ R, i = 1, 2, ..., m. We have
on
Du

a + b = (a1 + b1 )u 1 + (a2 + b2 )u 2 + . . . (am + bm )u m ∈ span(S)


αa = αa1 u 1 + αa2 u 2 + . . . αam u m ∈ span(S).

Hence, span(S) is a linear subspace of V .

Duong T. PHAM October 16, 2019 28 / 40


Linear Subspace

Proposition.

AM
Let V be a vector space and S = {u 1 , u 2 , ..., u m }⊂ V . Then
span(S) is the smallest subspace of V which contain S.

PH
S is the spanning set of span(S)

T.
Proof.
Assume that U is a subspace of V and S ⊂ U. For any
g.
α1 , α2 , ..., αm ∈ R , since u 1 , ..., u m ∈ U and U is a vector space,
on
Du

α1 u 1 + α2 u 2 + ... + αm u m ∈ U.

Hence, span(S) ⊂ U.

Duong T. PHAM October 16, 2019 29 / 40


Intersection

Proposition.

AM
Given U, W : subspaces of a vector space V , then U ∩ W is also a

PH
subspace of V .

T.
Proof: Let v , w ∈ U ∩ W and let α ∈ R.
g.
v , w ∈ U and U: vector space =⇒ v + w ∈ U and αv ∈ U
on
v , w ∈ W and W : vector space =⇒ v + w ∈ W and αv ∈ W
Du

Thus, v + w ∈ U ∩ W and αv ∈ U ∩ W . Hence, U ∩ W is a


subspace of V . 

Duong T. PHAM October 16, 2019 30 / 40


Sum of Linear Subspaces

Proposition.
Let U1 , U2 be linear subspaces of V . We define

AM
U1 + U2 = {a 1 + a 2 : a 1 ∈ U1 , a 2 ∈ U2 }.

PH
Then, U1 + U2 is a linear subspace of V .

Proof. Let a, b ∈ U1 + U2 and α ∈ R. Then

T.
a = a1 + a2g. for some a 1 ∈ U1 , a 2 ∈ U2
b = b1 + b2 for some b 1 ∈ U1 , b 2 ∈ U2 .
on
Du

Since U1 and U2 are vector spaces,


a + b = (a 1 + b 1 ) + (a 2 + b 2 ) ∈ U1 + U2
αa = (αa 1 ) + (αa 2 ) ∈ U1 + U2 .

Hence, U1 + U2 is a linear subspace. 


Duong T. PHAM October 16, 2019 31 / 40
Sum of Linear Subspaces

AM
Corollary.

PH
Let U1 , ..., Uk be subspaces of a vector space V . Then

T.
U1 + ... + Uk = {u 1 + ... + u k : u 1 ∈ U1 , ..., u k ∈ Uk }

is a linear subspace of V .
g.
on
Du

Duong T. PHAM October 16, 2019 32 / 40


Sum of Linear Subspaces

Proposition.
If U = hu 1 , . . . , u n i and V = hv 1 , . . . , v m i. Then

AM
U + V = hu 1 , . . . , u n , v 1 , . . . , v m i

PH
Proof.
Let a ∈ U + V . Then a = u + v for some u ∈ U and v ∈ V .

T.
U, V ⊂ hu 1 , . . . , u n , v 1 , . . . , v m i =⇒ u, v ∈ hu 1 , . . . , u n , v 1 , . . . , v m i
g.
=⇒ a = u + v ∈ hu 1 , . . . , u n , v 1 , . . . , v m i
on
=⇒ U + V ⊂ hu 1 , . . . , u n , v 1 , . . . , v m i
Du

Let a ∈ hu 1 , . . . , u n , v 1 , . . . , v m i. Then ∃α1 , ..., αn , β1 , ..., βm ∈ R


a = α1 u 1 + ... + αn u n + β1 v 1 + ... + βm v m ∈ U + V .
=⇒ hu 1 , . . . , u n , v 1 , . . . , v m i ⊂ U + V . 
Duong T. PHAM October 16, 2019 33 / 40
Sum of Linear Subspaces

AM
PH
Theorem.
Let U1 and U2 be two subspaces of a vector space V . Then

T.
dim(U1 + U2 ) = dimU1 + dimU2 − dim(U1 ∩ U2 )
g.
on
Du

Duong T. PHAM October 16, 2019 34 / 40


Rank of Matrix

Definition.
Given A ∈ Mm,n (R); the rank of A is an natural number r satisfying:

AM
∃ a submatrix B ∈ Mr (R) of A s.t. det B 6= 0

PH
All submatrices of A whose sizes are greater than r have zero
determinants.

T.
Notation: rank A
Remark. g.
on
(
rank A = n ⇐⇒ det A 6= 0
If A ∈ Mn (R), then
Du

rank A < n ⇐⇒ det A = 0


rank(Om,n ) = 0
If A ∈ Mm,n (R), then 0 ≤ rank A ≤ min{m, n}

Duong T. PHAM October 16, 2019 35 / 40


Computing rank of matrix

Using determinants:
Given A ∈ Mm,n (R);

AM
PH
Step 1: Find a square submatrix with biggest possible size (saying
= k) whose determinant is non-zero

T.
Step 2: Compute determinants of all submatrices of size k + 1
g.
which contain the submatrix in Step 1:
on
(i) If determinants of all these matrices of size k + 1 equal zero, then
Du

rank A = r
(ii) If ∃B ∈ Mk+1 (R) and det B 6= 0, then Repeat Step 2

Duong T. PHAM October 16, 2019 36 / 40


Computing rank of matrix

Using Gauss Elimination:

AM
Definition ( Row elementary operations ).
The following operations are called row elementary operations:

PH
(i) Adding into one row with a linear combination of other rows
(ii) Exchanging two rows

T.
(iii) Multiplicating or dividing a row with a non-zero number.
g.
on
Theorem.
Du

If we apply row and column elementary operations on a matrix A to obtain


matrix B, then rank B = rank A.

Duong T. PHAM October 16, 2019 37 / 40


Computing rank of a matrix

Ex. Find rank of the following matrix

AM
   
1 −3 4 5 2 1 −3 4 5 2
0 1 3 4 6  0 1 3 4 6
−3 5 −2 −3 −4 → 0 −4 10 12
A=  →

PH
 
2
−2 3 5 6 4 0 −3 13 16 8

T.
   
1 −3 4 5 2 1 −3 4 5 2
g.
0 1 3 4 6 0 1 3 4 6
0 0 22 28 26 → 0 0 22 28 26
on
   
Du

0 0 22 28 26 0 0 0 0 0

rank A = 3

Duong T. PHAM October 16, 2019 38 / 40


Null space–Solution spaces of homogeneous linear systems
   
x1 0
 ..   .. 
Consider the system: Am×n  .  = . (?)
xn n×1 0 m×1

AM
Proposition.

PH
The set of all solutions of (?) is a linear subspace of Rn .

T.
Proof. Assume that x=(x1 , ..., xn ) and y =(y1 , ..., yn ) are solutions of (?)
and α ∈ R. Then
g.
on
A(x + αy ) = Ax + αAy = 0m,1 + 0m,1 = 0m,1 .
Du

x + αy is a solution of (?). The set of all solutions of (?) is a linear


subspace of Rn . 
Definition.
Null space of a matrix A is the solution space of the system (?).
Duong T. PHAM October 16, 2019 39 / 40
Solution spaces of homogeneous linear systems

Theorem.
If rank A = r , then the dimension of solution space of (?) is n − r .

AM
Ex. Find basis and dimension of the solution space of the system

PH
T.
     
1 0 −1 0 0 1 0 −1 0 0 1 0 −1 0 0
 0 1 2 −1 0  →  0 1 2 −1 0  →  0 1 2 −1 0 
1 1 1 −1 0
g.
0 1 2 −1 0 0 0 0 0 0
on
Du

S = {(a, −2a + b, a, b) : a, b ∈ R}

A basis of S is the set consisting of α1 = (1, −2, 1, 0); α2 = (0, 1, 0, 1)


Hence, dim(S) = 2.

Duong T. PHAM October 16, 2019 40 / 40

You might also like