BasisAndSubspaces

Basis and Subspaces
Section 5.4
Nicholas Moore, PhD
The case of repeated eigenvalues is a special one as we’ve already seen in the simpler 2D case. In the 2D
case we had a repeated eigenvalue (with multiplicity 2, since we were in R2 ) but a single eigenvector. This
situation generalizes in the Rn case as well, but we need a little more sophisticated machinery to deal with it.
Recall from our previous discussions that the collection of all linear combinations of a set of vectors is called
a subspace. This idea of a subspace is important and one of the foundational principles around the concept
of a subspace is that of a basis.
1 Basis
Let S be a subspace of Rn . A collection of vectors V1 , V2 , . . . , Vk is a basis for S if the Vj are linearly
independent and span S . Another way to state this is to say that a basis for S is a set of linearly independent
vectors V1 , . . . , Vk such that
S = {α1 V1 + α2 V2 + · · · + αk Vk | α j ∈ R}
Note that a subspace always has a basis, since a subspace is the set of all linear combinations of a set of
vectors. To see this, assume we have a set of vectors: P = { P1 , P2 , . . . , Pk }. The span of this set (that is, the set
of all linear combinations) is given as
span P = {α1 P1 + α2 P2 + · · · + αk Pk }
If P is already a linearly independent set, then P is a basis for the subspace that it spans. If not, then we can
find one of the vectors, let’s say for the sake of convenience in notation that vector is Pk for which Pk is a
linear combination of the remaining vectors. That is
Pk = β 1 P1 + β 2 P2 + · · · + β k−1 Pk−1
Then, we can still represent span P as the linearly combination of P1 , P2 , . . . , Pk−1 . In essence, since Pk can
be already be written as the linearly combination of the other vectors, we don’t need it because it’s not
contributing anything to the set of linear combinations. We can continue removing vectors until the all the
remaining vectors are linearly independent, in which case we have a basis.
Tip
I like to think of a basis as a “minimum set” of vectors that you need to span the subspace. Removing
any vector from the basis will “remove information” and therefore the new set will not be able to
span the same subspace.
2 An Important Proposition
Proposition 1. Every basis of a subspace S ⊂ Rn has the same number of elements.
1
Proof. The proof of this proposition is given in Hirsh, Smale, and Devaney on page 29.
Since the number of basis elements is the same, regardless of what basis we use, this allows us to define
the dimension of a subspace as the number of elements in a basis. This by the way, is why we call R2 a
two-dimensional space. Every basis of R2 has two vectors. We denote the dimension of subspace S with
dim S . Finally, the subspace which is the span of just the zero vector is defined to have dimension 0; the
reasoning behind this choice will be more apparent later.
3 Some Quick Examples

Example Every line through the origin in a one-dimensional subspace since a line through the origin is also
represented as the set of all αV for some vector V. Therefore the line the set of all linear combinations of the
single vector on that line.
Example All planes through the origin are two-dimensional subspaces since they can all be represented as
the linear combination of two vectors.
4 Linear Transformations
A function T : Rn → Rn is linear if T ( X ) = AX for some n × n matrix A. T is called a linear map or linear
transformation. We see from the properties of matrix operations that these function satisfy the general
notion of being linear:
T (αX + βY ) = αT ( X ) + βT (Y ) ∀α, β ∈ R and X, Y ∈ Rn
We sometimes denote the linear map applied to vector X with T ( X ), but sometimes, since the map is linear
we will use the notation TX instead (since the operator is described as a matrix multiplication anyway, we
might as well use that notation). We also say that a linear map is invertible if the matrix, A, that generates
the map is invertible.
For the study of linear systems of differential equations, the most important types of subspaces that we will
care about are the kernels and ranges of linear maps. The Kernel of a linear map, T, denoted ker T is the set
of all vectors mapped to 0 by T. That is, all V such that T (V ) = 0 (notice in both the previous cases, the 0
represents the zero vector, not the scalar as is evident by the fact that T maps Rn to Rn ). The range of T is
the set of all W such that there exists a vector, V for which W = T (V ). Note that the range of T is always a
subspace of Rn .
Example. Consider the linear map  

0 1 0
T(X) =  0 0 1X
0 0 0
 
x
If we define X =  y , then
z
 
y
T(X) =  z 
0
Therefore,  
α
ker T =  0  ∀α ∈ R
0
2
while  
β
Range T =  γ  ∀ β, γ ∈ R.
0
5 Connections Between the Kernel and Range of a Linear Map

One important proposition in linear algebra is as follows:
Proposition 2. Let T : Rn → Rn be a linear map. Then ker T and Range T are both subspaces of Rn . Moreover,
dim ker T + dim Range T = n
This fact will be utilized in the next set of lecture notes as we discuss the potential to have repeated
eigenvalues.

BasisAndSubspaces

Uploaded by

Copyright:

Available Formats

BasisAndSubspaces

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

BasisAndSubspaces

Uploaded by

Copyright:

Available Formats

Basis and Subspaces

Nicholas Moore, PhD

3 Some Quick Examples

T (αX + βY ) = αT ( X ) + βT (Y ) ∀α, β ∈ R and X, Y ∈ Rn

Example. Consider the linear map  

5 Connections Between the Kernel and Range of a Linear Map

dim ker T + dim Range T = n

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.