Reading [SB] Ch. 11, p. 237-250, Ch. 27, p. 750-771.

1 Basis
1.1 Linear Combinations
A linear combination of vectors v1 , v2 , ... , vm ∈ Rn with scalar coefficients
α1 , α2 , ... , αm ∈ R is the vector

α1 · v1 + α2 · v2 + ... + αm · vm .

The set of all linear combinations of vectors v1 , v2 , ... , vm ∈ Rn is denoted

as
L[v1 , v2 , ... , vm ] = {α1 · v1 + α2 · v2 + ... + αm · vm , αi ∈ R}.
It is evident that L[v1 , v2 , ... , vm ] ⊂ Rn is a subspace.
Example. For a one single nonzero vector v ∈ Rn

L[v] = {t · v, t ∈ R}

is the line generated or spanned by v: it passes trough the origin and has
direction of v.
Example. For any two nonzero vectors v, w ∈ Rn

L[v, w] = {s · v + t · w, s, t ∈ R}

is either:
the line generated (or spanned ) by v if v and w are collinear, that is if
w = k · v, k ∈ R,
or is the plane generated (or spanned ) by v and w, which passes trough the
origin, if v and w are non-collinear.
Example. For any two non-collinear vectors v, w ∈ R2

L[v, w] = {s · v + t · w, s, t ∈ R}

is whole R2 .
Example. For any three nonzero vectors u, v, w ∈ R2 s.t. v and w are
non-collinear
L[v, w] = L[u, v, w] = R2 .

1
1.2 Linear Dependence and Independence
Definition 1. A sequence of vectors v1 , v2 , ... , vn is called linearly dependent
if one of these vectors is linear combination of others. That is
∃i, vi ∈ L(v1 , ... , v̂i , ... , vn ).

Definition 1’. A sequence of vectors v1 , v2 , ... , vm is linearly dependent if

there exist α1 , ... , αm with at last one nonzero αk s.t.
α1 · v1 + α2 · v2 + ... + αm · vm = 0.

Why these definitions are equivalent?

Example. Any sequence of vectors which contains the zero vector is linearly
dependent. (Why?)
Example. Any sequence of vectors which contains two collinear vectors is
linearly dependent. (Why?)
Example. Any sequence of vectors of R2 which consists of more then two
vectors is linearly dependent. (Why?)
Example. A sequence consisting of two vectors v1 , v2 is linearly dependent if
and only if these vectors are collinear (proportional), i.e. v2 = k · v1 . (Why?)

Definition 2. A sequence of vectors v1 , v2 , ... , vn is called linearly indepen-

dent if it is not linearly dependent.

Definition 2’. A sequence of vectors v1 , v2 , ... , vn is called linearly inde-

pendent if non of these vectors is a linear combination of others.

Definition 2”. A sequence of vectors v1 , v2 , ... , vn is called linearly inde-

pendent if
α1 · v1 + α2 · v2 + ... + αm · vm = 0
is possible only if all αi -s are zero.

Why these definitions are equivalent?

Example. The vectors e1 = (1, 0, 0), e2 = (0, 1, 0), e3 = (0, 0, 1) ∈ R3 are

linearly independent.
Indeed, suppose α1 e1 + α2 e2 + α3 e3 = 0, this means
α1 · (1, 0, 0) + α2 · (0, 1, 0) + α3 · (0, 0, 1) =
(α1 , 0, 0) + (0, α2 , 0) + (0, 0, α3 ) = (α1 , α2 , α3 ) = (0, 0, 0),
thus α1 = 0, α2 = 0, α3 = 0.

2
1.2.1 Linear Independence and Systems of Linear Equations
How to check wether a given sequence of vectors v1 , v2 , ... , vm ∈ Rn is
linear dependent or independent?
Let  
v11 v21 ... vm1
 v v ... vm1 
 
A =  12 22 ,
 ... ... ... ... 
v1n v2n ... vmn
be the matrix whose columns are vj ’s.

Theorem 1 A sequence of vectors v1 , v2 , ... , vm is linear independent iff

the homogenous system Aα = 0 has only zero solution α = (0, ... , 0).

Example. Determine whether the sequence of vectors is linearly dependent

     
1 1 0
 0   0   0 
     
v1 =   , v2 =   , v3 =  .
 1   0   1 
0 1 1

Solution. We must check whether the equation

c1 · v1 + c2 · v2 + c3 · v3 = 0

has non-all-zero solution for c1 , c2 , c3 . In coordinates this equation looks as

a system 
 1 · c1 + 1 · c2


+ 0 · c3 = 0
 0·c + 0·c + 0 · c3 = 0
1 2
.


 1 · c1 + 0 · c 2 + 1 · c3 = 0

0 · c1 + 1 · c2 + 1 · c3 = 0
The matrix of the system  
1 1 0
 0 0 0 
 
 
 1 0 1 
0 1 1
has maximal rank 3. So there are no free variables, and the system has only
zero solution. Thus this sequence of vectors is linearly independent.

Example. Determine whether the sequence of vectors is linearly dependent

     
1 1 1
 0   0   0 
v1 = 

 
 , v2 = 
 
 , v3 = 

.
 1   −1   0 
0 0 0

3
Solution. We must check whether the equation

c1 · v1 + c2 · v2 + c3 · v3 = 0

has non-all-zero solution for c1 , c2 , c3 . In coordinates this equation looks as

a system 

 1 · c1 + 1 · c2 + 1 · c3 = 0

 0·c + 0·c
1 2 + 0 · c3 = 0
.


 1 · c 1 − 1 · c2 + 0 · c3 = 0

0 · c1 + 0 · c2 + 0 · c3 = 0
The matrix of the system  
1 1 1
 0 0 0 
 
 
 1 −1 0 
0 0 0
has the rank 2. So there is free variable, and the system has non-zero solutions
too. Thus this sequence of vectors is linearly dependent.

Theorem 2 A set of vectors v1 , v2 , ... , vk in Rn with k > n is linearly

dependent.

Proof. We look at a nonzero solution c1 , ... , ck of the equation

c1 · v1 + ... + ck · vk = 0,

or, equivalently, of the system



 v11 · c1 + ... + vk1 · ck = 0

 v12 · c1 + ... + vk2 · ck = 0
.


 ... ... ... ... ... ...

v1n · c1 + ... + vkn · ck = 0

This homogenous system has k variables and n equations. Then rank ≤

n < k, so there definitely are free variables, consequently there exists nonzero
solution c1 , ... , ck .

Theorem 3 A set of vectors v1 , v2 , ... , vn in Rn is linearly independent iff

det(v1 v2 ... vn ) 6= 0.

Proof. We look at a nonzero solution for c1 , ... , cn of the equation

c1 · v1 + ... + cn · vn = 0.

The system which corresponds to this equation has n variables and n equa-
tions and is homogenous. So it has a non-all-zero solutions iff its determinant
is zero.

4
1.3 Span
Let v1 , ... , vk be a sequence of m vectors from Rn .
The set of all linear combinations of these vectors

L[v1 , ... , vk ] = {α1 · v1 + α2 · v2 + ... + αk · vk , α1 , ... , αk ∈ R}

is called the set generated (or spanned) by the vectors v1 , ... , vk .

Example. The vectors v1 = (1, 0, 0), v2 = (0, 1, 0) span the xy plane (the
plane given by the non-parameterized equation z = 0) of R3 . Indeed, any
point p = (a, b, 0) of this plane is the following linear combination

av1 + bv2 = a(1, 0, 0) + b(0, 1, 0) = (a, 0, 0) + (0, b, 0) = (a, b, 0).

Example. The vectors v1 = (1, 2), v2 = (3, 4) span whole R2 . Indeed, let’s
take any vector v = (a, b). Our aim is to find c1 , c2 s.t.

c1 · v1 + c2 · v2 = v.

In coordinates this equation looks as a system

(
c1 · 1 + c2 · 3 = a
.
c1 · 2 + c2 · 4 = b

The determinant of this system 6= 0, so this system has a solution for each a
and b.

Example. Different sequences of vectors can span the same sets. For exam-
ple R2 is spanned by each of the following sequences:
(a) v1 = (1, 0), v2 = (0, 1);
(b) v1 = (−1, 0), v2 = (0, 1);
(c) v1 = (1, 1), v2 = (0, 1);
(d) v1 = (1, 2), v2 = (2, 1);
(e) v1 = (1, 0), v2 = (0, 1), v3 = (2, 3).

Check this!

For a given sequence of vectors v1 , ... , vk ∈ Rn form the n × k matrix

whose columns are vi ’s:
 
v11 v21 ... vk1
 v12 v22 ... vk1 
A=


,
 ... ... ... 
v1n v2n ... vkn

here vi = (vi1 , vi2 , ... , vin ).

5
Theorem 4 Let v1 , ... , vk ∈ Rn be a sequence of vectors. A vector b ∈ Rn
lies in the space L(v1 , ... , vk ) if and only if the system A·c = b has a solution.

Proof. Evident: A · c = b means c1 v1 + ... + ck vk = b.

Corollary 1 A sequence of vectors v1 , ... , vk ∈ Rn spans Rn if and only if

the system A · c = b has a solution for any vector b ∈ Rn .

Corollary 2 A sequence of vectors v1 , ... , vk ∈ Rn with k ≤ n can not span

Rn .

Proof. In this case the matrix A has less columns than rows. Choosing
appropriate b we can make rank(A|b) > rank(A) (how?), this makes the
system A · c = b non consistent for this b.

1.4 Basis and Dimension

A sequence of vectors v1 , ... , vn ∈ Rn forms a basis of Rn if
(1) they are linearly independent;
(2) they span Rn .

Example. The vectors

e1 = (1, 0, ... , 0), e2 = (0, 1, ... , 0), ... , en = (0, 0, ... , 1)

form a basis of Rn .
Indeed, firstly they are linearly independent since the n × n matrix

(e1 e2 ... en )

is the identity, thus it’s determinant is 1 6= 0.

Secondly, they span R2 : any vector v = (x1 , ... , xn ) is the following linear
combination
v = x1 · e1 + ... + xn · en .

A basis v1 , ... , vn ∈ Rn is called orthogonal if vi · vj = 0 for i 6= j. This

means that all vectors are perpendicular to each other: vi · vj = 0 for i 6= j.
An orthogonal basis v1 , ... , vn ∈ Rn is called orthonormal if vi · vi = 1.
This means that each vector of this basis has the length 1. In other words:
vi · vj = δi,j where δij is famous Kroneker’s symbol
(
1 if i = j
δij = .
0 if i 6= j

The basis e1 , ... , en is orthonormal.

6
Theorem 5 Any two non-collinear vectors of R2 form a basis.

For example e1 = (1, 0), e2 = (0, 1) is a basis. Another basis is, say
e01 = (1, 0), e02 = (1, 1).

Theorem 6 Any basis of Rn contains exactly n vectors.

Why? Because more than n vectors are linearly dependent, and less than n
vectors can not span Rn .

The dimension of a vector space is defined as the number of vectors in its

basis. Thus
dim Rn = n.

Theorem 7 Let v1 , ... , vk ∈ Rn and A be the matrix whose columns are

vj ’s:
 
v11 v21 ... vn1
 v v22 ... vn1 
A= 
12 
.
 ... ... ... 
v1n v2n ... vnn
Then the following statements are equivalent
(a) v1 , ... , vn are linearly independent;
(b) v1 , ... , vn span Rn ;
(c) v1 , ... , vn is a basis of Rn ;
(d) det A 6= 0.

Example. R3 is 3 dimensional : we have here a basis consisting of 3 vectors

v1 = (1, 0, 0), v2 = (0, 1, 0), v3 = (0, 0, 1)

Generally, the dimension of Rn is n: it has a basis consisting of n elements

e1 = (1, 0, ... , 0), e2 = (0, 1, ... , 0), ... , en = (0, 0, ... , 1).

1.5 Subspace
A subset V ∈ Rn is called subspace if V is closed under vector operations -
summation and scalar multiplication, that is:

v, w ∈ V, c ∈ R ⇒ v + w ∈ V, c · v ∈ V.

Example. The line x(t) = t · (2, 1), that is all multiples of the vector
v = (2, 1) which passes trough the origin is a subspace. But the line x(t) =
(1, 1) + t · (2, 1) is not.

7
Theorem 8 Let w1 , ... , wk ∈ Rn be a sequence of vectors. Then the set of
all linear combinations
L[w1 , ... , wk ] ⊂ Rn
is a subspace.

Why?
Example. The subspace of R3

{(a, b, 0), a, b ∈ R},

which is the xy plane, has dimension 2:

v1 = (1, 0, 0), v2 = (0, 1, 0)

is its basis.
Example. Similarly, the subspace of R3

{(a, 0, 0), a ∈ R},

which is the y line, has dimension 1:

v1 = (1, 0, 0)

is its basis.

1.5.1 How to find the dimension and the basis of L(v1 , ... , vk )?
Let v1 , ... , vk ∈ Rn be a sequence of vectors from Rn , and L(v1 , ... , vk ) ⊂ Rn
be the corresponding subspace. How can we find the dimension and basis of
this subspace?

Let  
v11 v21 ... vk1
 v12 v22 ... vk1 
A=


,
 ... ... ... 
v1n v2n ... vkn
be the matrix whose columns are vj ’s. Let r be the rank of this matrix and
M be a corresponding main r ×r minor. Then the dimension of L(v1 , ... , vk )
is r and its basis consists of those vi -s, who intersect M (why?).

1.6 Conclusion
Let v1 , v2 , ... , vk be a sequence of vectors from Rn and let
 
v11 v21 ... vk1
 v12 v22 ... vk1 
A=


,
 ... ... ... 
v1n v2n ... vkn

8
be the matrix whose columns are vj ’s. Let r be the rank of this matrix. The
following table shows when this sequence is linearly independent or spans Rn
depending on value of k:

k<n k=n n<k

independent r=k r=n no
spans Rn no r=n r=n

2 Spaces Attached to a Matrix

Let  
a11 ... a1n
 
A =  ... ... ... 
am1 ... amn
be an m × n matrix. There are three vector spaces attached to A: the
column space Col(A) ⊂ Rm , the row space Row(A) ⊂ Rn and the null space
N ull(A) ⊂ Rn .

2.1 Column Space

The column space Col(A) is defined as a subspace of Rm spanned by column
vectors of A, that is
     
a11 a12 an1
 a21   a22   an2 
Col(A) = L(


,


 
 , ... , 

).
 ...   ...   ... 
am1 am2 anm
Theorem 9
dim Col(A) = rank A.

2.1.1 How to Find a Basis of Col(A)

Just find a basic minor of A. Then all the columns that intersect this minor
form a basis of Col(A).
Example. Find a basis of Col(A) for
 
2 3 1 4
 
A =  2 3 7 9 .
2 3 13 14
Solution.
à ! Calculation shows that a basic minor here can be chosen as
1 4
. So the basis of Col(A) consists of last two columns
7 9
   
1 4
   
 7 ,  9 .
13 14

9
 Ã !
3 7
Of course we can choose as a basic minor . In this case we obtain
3 13
   
3 1
   
a basis of Col(A) consisting of second and third columns  3  ,  7  .
3 1

2.1.2 The Role of Column Space

(a) The system A · x = b has a solution for a particular b ∈ Rm if b belongs
to column space Col(A).

(b) The system A·x = b has a solution for every b ∈ Rm if and only if rank A
equals of number of equations m.

(c) If A · x = b has a solution for every b, then

number of equations = rank A ≤ number of variables.

2.2 Row Space

The row space Row(A) is defined as a subspace of Rn spanned by row vectors
of A, that is
Row(A) = L(w1 , w2 , ... , wm )
where w1 , ... , wm are the row vectors of A:
w1 = (a11 , ... , a1n )
...
wm = (am1 , ... , amn ).
Theorem 10
dim Row(A) = rankA.
So the dimensions of the column space Col(A) and the row space Row(A)
both equal to rank A.

2.2.1 How to Find a Basis of Row(A)

Just find a basic minor of A. Then all the rows that intersect this minor
form a basis of Row(A)
Example. Find a basis of Row(A) for
 
2 3 1 4
 
A =  2 3 7 9 .
2 3 13 14
à !
1 4
Calculation shows that a basic minor here can be chosen as . So
7 9
the basis of Row(A) consists of first two rows (2, 3, 1, 4), (2, 3, 7, 9).

10
2.3 Null-space
Previous two attached spaces Col(A) and Row(A) are defined as subspaces
generated by some vectors.
The third attached space N ull(A) is defined as the set of all solutions of
the system A · x = 0, i.e.

N ull(A) = {x = (x1 , ... , xn ) ∈ Rn , A · x = 0}.

But is this set a subspace? Yes, yes! But why?

Theorem 11 A subset N ull(A) is a subspace.

Proof. We must show that N ull(A) is closed with respect to addition and
scalar multiplication. Indeed, suppose x, x0 ∈ N ull(A), that is A · x =
0, A · x0 = 0. Then

A(x + x0 ) = A(x) + A(x0 ) = 0 + 0 = 0.

Furthermore, let x ∈ N ull(A) and c ∈ R. Then

A · (c · x) = c · (A · x) = c · 0 = 0.

2.3.1 How to Find a Basis of N ull(A)

To find a basis of null-space N ull(A)) just solve a system A · x = 0, that
is express basic variables in terms of free variables. As we know there are
r = rank(A) basic variables, say

x1 , x2 , ... , xr ,

and consequently n − r free variables, in this case

xr+1 , xr+2 , ... , xn .

Express the basic variables in terms of free variables, and find n − r following
particular solutions particular solutions which form a basis of N ull(A)
       
x11 x21 x1n−r−1 x1n−r
 x12   x22   x2n−r−1   xn−r 
       2 
       


... 



... 



... 



... 

 x1r   x2r   n−r−1
xr   xn−r 
       r 
       
v1 =  1  , v2 =  0  , ... , vn−r−1 =  0  , vn−r =  0 .
       
 0   1   0   0 
       
 ...   ...   ...   ... 
       
       
 0   0   1   0 
0 0 0 1

11
Example. Find a basis for the null-space of the matrix
 
1 −1 3 −1
 1 4 −1 1 
 
A= .
 3 7 1 1 
3 2 5 −1
Solution. First solve the homogenous system


 x1 − x2 + 3x3 − 1x4 = 0

 x1 + 4x2 − x3 + x4 = 0
.


 3x1 + 7x2 + x3 + x4 = 0

3x1 + 2x2 + 5x3 − x4 = 0
Computation gives rank A = 2, so dim N ull(A) = 4 − rank A = 4 − 2 = 2,
and the solution gives
x1 = −2.2x3 + 0.6x4 , x2 = 0.8x3 − 0.4x4 , x3 = x3 , x4 = x4
So the general solution is
   
x1 −2.2x3 + 0.6x4
 x2   0.8x3 − 0.4x4 
   
 = 
 x3   x3 
x4 x4
Substituting

x3 = 1, x4 = 0 we obtain the first basis vector of null space
−0.22
 0.8 
 
v1 =  .
 1 
0
Now substituting

x3 = 0, x4 = 1 we obtain the second basis vector of
0.6
 −0.4 
null space v2 = 

.
 0 
1
   
−0.22 0.6
 0.8   −0.4 
   
So the basis of N ull(A) is   , v2 =  .
 1   0 
0 1

2.4 Fundamental Theorem of Linear Algebra

The column space of A, spanned by n column vectors, and the row space of
A, spanned by m row vectors, have the same dimension equal to rankA.
The Fundamental Theorem of Linear Algebra describes the dimension of
the third subspace attached to A:
Theorem 12 dim Null(A)+rank A=n.

12
2.5 Solutions of Systems of Linear Equations
We already know how to express all solutions of homogenous system A·x = 0:
just find a basis of N ull(A)
v1 , v2 , ... , vn−r ,
then any solution, since it is an element of N ull(A), is a linear combination
x = α1 v1 + ... + αn−r vn−r .

Now turn to non-homogenous systems.

Let A · x = b, x ∈ Rn , b ∈ Rm be a system of linear equations and
A · x = 0 be the corresponding homogenous system.
Theorem 13 Let c be a particular solution of A · x = b. Then, every other
solution c0 of A · x = b can be written as c0 = c + w where w is a vector from
N ull(A), that is a solution of homogenous system A · x = 0.
Proof. Since c and c0 are solutions, we have A · c = b, A · c0 = b. Let’s define
w = c0 − c. Then
A · w = A · (c0 − c) = A · c0 − A · c = b − b = 0,
so w = c0 − c is a solution of A · x = 0. Thus c0 = c + w.

According to this theorem in order to know all solutions of A · x = b it

is enough to know one particular solution of A · x = b and all solutions of
A · x = 0. Then any solution is given by
{c + α1 · v1 + ... + αn−r · vn−rank A }.

But how to find one particular solution of A · x = b? Just take (for

example) the following free variables xr+1 = 0, xr+2 = 0, ... , xn = 0 and
solve x1 , ... , xr .

Example. Express general solution of the system



 x1 − x2 + 3x3 − 1x4 = 1

 x1 + 4x2 − x3 + x4 = 6
.


 3x1 + 7x2 + x3 + x4 = 13

3x1 + 2x2 + 5x3 − x4 = 8
Solution. We already know general solution of corresponding homogenous
system A · x = 0: a basis of N ull(A) is
   
−0.22 0.6
 0.8   −0.4 
v1 = 


, v2 = 


,
 1   0 
0 1

13
so the general solution of homogenous system is given by
     
x1 −0.22 0.6
 x2   0.8   −0.4 
     
  = α1 ·  + α2 · .
 x3   1   0 
x4 0 1
Now we need one particular solution of non-homogenous system. Take
x3 = 0, x4 = 0, we obtain
(
x1 − x2 = 1
.
x1 + 4x2 = 6
This gives x1 = 2, x2 = 1. So a particular solution is
   
x1 2
 x2   1 
   
 = .
 x3   0 
x4 0
Finally, the general solution of nonhomogenous system is given by
       
x1 2 −0.22 0.6
 x2   1   0.8   −0.4 
       
 =  + α1 ·  + α2 · .
 x3   0   1   0 
x4 0 0 1

2.6 Orthogonal Complement

For a subspace V ⊂ Rn its orthogonal complement V ⊥ ⊂ Rn is defined as
the set of all vectors w ∈ Rn that are orthogonal to every vector from V , i.e.
V ⊥ = {w ∈ Rn , v · w = 0 f or ∀ v ∈ V }.
Proposition 1 For any subspace V ⊂ Rn
(a) V ⊥ is a subspace.
T
(b) V V ⊥ = {0}.
(c) dim V + dim V ⊥ = n.
(d) (V ⊥ )⊥ = V .
(e) Suppose V, W ∈ Rn are subspaces, din V + din W = n and for each
v ∈ V, w ∈ W one has v · w = 0. Then W = V ⊥ .
Proof of (a). 1. Suppose w ∈ V ⊥ , i.e. w · v = 0 f or ∀v ∈ V . Let us show
that kw ∈ V ⊥ . Indeed
kw · v = k(w · v) = k · 0 = 0.
2. Suppose w, w0 ∈ V ⊥ , i.e. w · v = 0, w0 · v = 0 f or ∀v ∈ V . Let us show
that w + w0 ∈ V ⊥ . Indeed
(w + w0 ) · v = w · v + w0 · v = 0 + 0 = 0.

14
Theorem 14 For a matrix A
(a) Row(A)⊥ = N ull(A).
(b) Col⊥ = N ullAT .

Example. In R3 , the orthogonal complement to xy plane is the z-axes. Prove

it!

15
 Exercises

Exercises from [SB]

11.2, 11.3, 11.9, 11.10, 11.12, 11.13, 11.14
27.1, 27.2, 27.3, 27.4, 27.5, 27.6, 27.7, 27.8, 27.10
27.12, 27.13, 27.14, 27.17

Homework

1. Exercise 11.12.
2. Show that the vectors from 11.14 (b) do not span R3 : present at last
one vector which is NOT their linear combination.
3. Show that the vectors from 11.14 (b) are linearly dependent: find their
linear combination with non-all-zero coefficients which gives the zero vector.
4. Show that if v ∈ Row(A), w ∈ N ull(A) then v · w = 0. Actually this
proofs Row(A)⊥ = N ull(A).
5. Exercise 27.10 (d).

16
 Summary
 
v11 v21 ... vm1
n  
Let v1 , v2 , ... , vm ∈ R and A =  ... ... ... ...  be the matrix
v1n v2n ... vmn
whose columns are vj ’s.
Linear Combinations: L[v1 , v2 , ... , vm ] = {α1 · v1 + ... + αm · vm .
Linearly dependent: ∃i, vi ∈ L(v1 , ... , v̂i , ... , vm ) or ∃ (α1 , ... , αm ) 6=
(0, ... , 0) s.t. α1 · v1 + α2 · v2 + ... + αm · vm = 0.
Linearly independent: α1 · v1 + α2 · v2 + ... + αm · vm = 0 ⇒ ∀ αk = 0,
or Aα = 0 has only zero solution.
(v1 , ... , vk ) ∈ Rn spans Rn if L[v1 , v2 , ... , vm ] = Rn or Aα = b has a
solution for ∀ b = (b1 , ... , bn ).
(v1 , ... , vk ) ∈ Rn is a basis if it is lin. indep. and spans Rn .
n lin. indep. vectors span Rn , so they form a basis. n vectors spanning
Rn are lin. indep., so they form a basis.
Subspace V ⊂ Rn : v, w ∈ V, c ∈ R ⇒ v + w ∈ V, c · v ∈ V.
Dimension and basis of L[v1 , v2 , ... , vm ]: dimension is rank A, basis
- the columns intersecting main minor.

Spaces Attached to a Matrix

 
a11 ... a1n
Let A =  
 ... ... ...  .
am1 ... amn
     
a11 a12 an1
Column space: Col(A) = L[    
 ...  ,  ...  , ... ,  ... ].


am1 am2 anm

b ∈ Col(A) iff Ax = b has a solution.
dim Col(A) = rank A, basis - columns that intersect main minor.
Row Space: Row(A) = L[(a11, ... , a1n ), ... , (am1, ... , amn )]. dim Row(A) =
rank A, basis - rows that intersect main minor.
Null-space: N ull(A) = {x ∈ Rn , A·x = 0}. dim N ull(A) = n−rank A.
Basis of N ull(A) - the following solutions of Ax = 0
       
x11 x21 x1n−r−1 x1n−r
       
 ...   ...   ...   ... 
       

 x1r 


 x2r 



n−r−1
xr 


 xn−r
r


 1   0   0   0 
v1 = 

 
 , v2 = 
 
 , ... , vn−r−1 = 
 
 , vn−r = 

.

 0 


 1 


 0 


 0 

 ...   ...   ...   ... 
       
       
 0   0   1   0 
0 0 0 1
Orthogonal complement: V ⊥ = {w ∈ Rn , v · w = 0 f or ∀ v ∈ V }.
Row(A)⊥ = N ull(A), Col⊥ = N ullAT .

17