Basis
Basis
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 .
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 ).
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.
c1 · v1 + c2 · v2 + c3 · v3 = 0
3
Solution. We must check whether the equation
c1 · v1 + c2 · v2 + c3 · v3 = 0
c1 · v1 + ... + ck · vk = 0,
det(v1 v2 ... vn ) 6= 0.
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
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
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.
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!
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. 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.
form a basis of Rn .
Indeed, firstly they are linearly independent since the n × n matrix
(e1 e2 ... en )
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).
Why? Because more than n vectors are linearly dependent, and less than n
vectors can not span Rn .
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
is its basis.
Example. Similarly, the subspace of R3
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:
9
à !
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
(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.
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.
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 · (c · x) = c · (A · x) = c · 0 = 0.
x1 , x2 , ... , xr ,
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
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 .
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
14
Theorem 14 For a matrix A
(a) Row(A)⊥ = N ull(A).
(b) Col⊥ = N ullAT .
15
Exercises
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.
17