Eigenvalues, Eigenvectors, and Eigenspaces - Part 1
Eigenvalues, Eigenvectors, and Eigenspaces - Part 1
Eigenvalues, Eigenvectors, and Eigenspaces - Part 1
Question 1
Answer saved
The definitions of eigenvalues and eigenvectors, along with basic examples, are presented in the video Eigenvalues and eigenvectors 1.
Definition: Let A be an n × n matrix. A scalar λ (pronounced lambda) is called an eigenvalue of A if there is a non-zero vector x ∈ R
n
such
that
Ax = λx
What this definition means is that if x is an eigenvector, then multiplying x with A transforms x in the simplest possible way: it just multiplies it
by a scalar.
−1 −1 1
⎡ ⎤
Example: Let A = ⎢ 0 −2 0⎥ .
⎣ ⎦
2 −2 0
Thus,
1
λ = -2 is an eigenvalue of A and 1 is an eigenvector of A corresponding to λ = -2 .
Correct answer, well done. Correct answer, well done. Correct answer, well done.
Marks for this submission: 1.00/1.00. Marks for this submission: 1.00/1.00. Marks for this submission: 1.00/1.00.
−1 −1 1 1
-3 1
⎡ ⎤⎡ ⎤ ⎡ ⎤
⎢ 0 −2 0 ⎥⎢ 2 ⎥ = -4 ≠ λ ⎢ 2⎥
⎣ ⎦⎣ ⎦ ⎣ ⎦
2 −2 0 0
-2 0
Check
Question 2
Correct
Geometric Interpretation
Let A be an n × n matrix.
Geometrically, a non-zero vector x ∈ R
n
in an eigenvector of A if and only if Ax is parallel to x .
Is x an eigenvector of A? Yes
Is u an eigenvector of A? No
Is v an eigenvector of A? Yes
Multiplication by the matrix A is a linear transformation. More precisely, the map TA : Rn → Rn defined by TA (x) = Ax is a linear
transformation. So an eigenvector of A is a non-zero vector which is mapped by TA to a scalar multiple of itself.
In Block 4, we noted that matrices and linear transformations between Euclidean spaces (that is, the spaces Rn and Rm ) are in fact equivalent
concepts: every linear transformation is a matrix transformation and vice versa. So we also have equivalent concepts of eigenvalues and
eigenvectors for linear transformations. More precisely, a scalar λ is called an eigenvalue of a linear transformation T : Rn → Rn if there is
a non-zero vector x ∈ Rn such that T (x) = λx, in which case the vector x is called an eigenvector of T corresponding to the eigenvalue λ.
Indeed, λ is an eigenvalue of T if and only if it is an eigenvalue of the standard matrix [T ] of T . The same statement applies to eigenvectors as
well.
Example: Consider the linear operator T given by reflection about some line ℓ in R .
2 2 2
: R → R
-1
.
Check
Correct
Marks for this submission: 5.00/5.00.
Question 3
Answer saved
Exercises:
−8 10 1 2
(a) Let A = [ ,
] u = [ ] , and v = [ ] . Show that u and v are eigenvectors of A and find their corresponding eigenvalues.
−5 7 1 1
1 2 0 0 −3
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
(b) Let B = ⎢ 0 2 ,
0⎥ w = ⎢ 0⎥ , and x = ⎢ 0 ⎥ . Show that w and x are eigenvectors of B and find their corresponding
⎣ ⎦ ⎣ ⎦ ⎣ ⎦
−1 0 0 7 3
eigenvalues.
1 2 0 −3
-3 Correct answer, well done.
⎡ ⎤⎡ ⎤
Bx = ⎢ 0 2 0 ⎥⎢ 0 ⎥= 0 Marks for this submission: 1.00/1.00.
⎣ ⎦⎣ ⎦
−1 0 0 3
3
Thus Bx = x .
Check
Question 4
Correct
Question:
How do we compute the the eigenvalues of an n × n matrix A, that is, how do we find the values of the scalars λ for which Ax = λx , for some
non-zero vector x ∈ Rn ?
Let I be the n × n identity matrix. The argument below is explained in the video Eigenvalues and eigenvectors 3.
Ax = λx , for some x ≠ 0
⟺ Ax − λx = 0 , for some x ≠ 0
Computing the determinant of A − λI using cofactor expansion along any row or column produces a polynomial in the variable λ. (This will
become more clear after working out some examples: if you want, you may look at the first example worked out in Question 6 below and come
back here afterwards.) Therefore, the eigenvalues of an n × n matrix A are the solutions λ of the polynomial equation
det(A − λI ) = 0
C A (λ) = det(A − λI )
Question:
How do we compute the the eigenvalues of an n × n matrix A, that is, how do we find the values of the scalars λ for which Ax = λx , for some
non-zero vector x ∈ R
n
?
Answer:
To compute the eigenvalues of an n × n matrix A, we find the roots of the characteristic polynomial C A(λ) = det(A − λI ) of A.
We will go through some examples of how to compute det(A − λI ) shortly, but first we will do some examples that involve finding the
eigenvalues of A once we know the characteristic polynomial C A(λ) of A.
3 2
C A (λ) = −(λ − 8) (λ + 6)
-6
.
Example: Suppose that A is a 7 × 7 matrix with characteristic polynomial
2 3
C A (λ) = −λ(λ + 5)(λ + 1) (λ − 2)
2
C A (λ) = λ + λ − 12
2 5
C A (λ) = (λ + 2) (λ − 7) (λ − 9)
Then n =
Note: For a square matrix A with real entries, the characteristic polynomial C A(λ) may have some roots in the field of complex numbers, which
would mean that A has complex eigenvalues. We will discuss complex numbers, complex eigenvalues and eigenvectors in Block 6.
Check
Correct
Algebraic Multiplicity
Definition: The algebraic multiplicity of an eigenvalue λ0 of an n × n matrix A, denoted alg(λ0 ), is the number of times λ0 appears as a
root of the characteristic polynomial C A(λ) of A. Equivalently, it is the number of times λ − λ0 appears as a factor of C A(λ).
3 4 2
C A (λ) = −λ (λ − 2) (λ − 6) .
Thus, the eigenvalues of A are 0 , 2 , and 6 . These eigenvalues have algebraic multiplicities as follows:
.
λ = 6 has algebraic multiplicity alg(6) =
.
4 5 2 3
C A (λ) = −(λ + 5) (λ + 1) (λ − 1) (λ − 11)(λ − 19)
The eigenvalue λ =
11
19
-5
-1
Check
Answer saved
Another example about finding the eigenvalues of a 3x3 matrix, which is a bit more complicated than the one found below, is presented in the
video Eigenvalues and eigenvectors 4.
−1 −1 1
⎡ ⎤
Example: Let A = ⎢ 0 −2 0⎥ . Compute the characteristic polynomial C A(λ) and eigenvalues of A.
⎣ ⎦
2 −2 0
−1 −1 1 λ 0 0 −1 − λ −1 1
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
A − λI = ⎢ 0 −2 0⎥ − ⎢ 0 λ 0 ⎥ = ⎢ 0 −2 − λ 0 ⎥
⎣ ⎦ ⎣ ⎦ ⎣ ⎦
2 −2 0 0 0 λ 2 −2 −λ
Thus,
∣ −1 − λ −1 1 ∣
∣ ∣
C A (λ) = det(A − λI ) = 0 −2 − λ 0
∣ ∣
∣ 2 −2 −λ ∣
∣ −1 − λ 1 ∣ expand
= (−2 − λ) ∣ ∣ ( )
∣ 2 −λ ∣ along R2
2
= −(λ + 2)(λ + λ − 2)
2
= −(λ + 2) (λ − 1)
Exercises:
4 1
(a) Let A = [ ] . Find the characteristic polynomial and eigenvalues of A.
11 −6
4 1 λ 0 4 − λ 1
A − λI = [ ] − [ ] = [ ]
11 −6 0 λ 11 −6 − λ
3 1 −1
⎡ ⎤
(b) Let B = ⎢ −1 5 −1 ⎥ . Find the characteristic polynomial and eigenvalues of B.
⎣ ⎦
−4 4 0
Check
Question 7
Answer saved
0 0 0 0
⎡ ⎤
⎢ 1 4 0 0 ⎥
A = ⎢ ⎥
⎢ −1 5 2 0 ⎥
⎣ ⎦
0 3 −6 −7
0 0 0 0 λ 0 0 0 −λ 0 0 0
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎢ 1 4 0 0 ⎥ ⎢ 0 λ 0 0 ⎥ ⎢ 1 4 − λ 0 0 ⎥
A − λI = ⎢ ⎥ − ⎢ ⎥ = ⎢ ⎥
⎢ −1 5 2 0 ⎥ ⎢ 0 0 λ 0 ⎥ ⎢ −1 5 2 − λ 0 ⎥
⎣ ⎦ ⎣ ⎦ ⎣ ⎦
0 3 −6 −7 0 0 0 λ 0 3 −6 −7 − λ
Since A − λI is a triangular matrix, its determinant is the product of the entries on its main diagonal. Thus,
Theorem: The eigenvalues of a triangular n × n matrix A (upper, lower, or diagonal) are the entries on the main diagonal of A.
18 3 −1 7
⎡ ⎤
⎢ 0 −9 3 0 ⎥
Example: Let B = ⎢ ⎥ .
⎢ 0 0 5 1 ⎥
⎣ ⎦
0 0 0 27
Check
Question 8
Correct
Definition: Let A be an n × n matrix and let λ be an eigenvalue of A. The set of all eigenvectors of A corresponding to λ, together with the
zero vector, is called the eigenspace of A corresponding to λ and is denoted Eλ . More precisely,the eigenspace corresponding to λ is given by
Eλ = Nul(A − λI )
Indeed,
x ∈ Eλ ⟺ Ax = λx
⟺ Ax − λx = 0
⟺ (A − λI )x = 0
⟺ x ∈ Nul(A − λI )
Definition: Let A be an n × n matrix and let λ be an eigenvalue of A. The geometric multiplicity of λ, denoted geo(λ) , is the dimension of
the eigenspace Eλ corresponding to λ.
Check
Correct
Marks for this submission: 1.00/1.00.
Question 9
Answer saved
To describe explicitly any subspace, it is enough to find a basis. This is true, in particular, for eigenspaces. It is explained in the video Eigenvalues
and eigenvectors 2 how to find the basis of one eigenspace for a specific matrix. Another example is given below.
[ ]
−1 −1 1
Example: Let A = 0 − 2 0 . In Question 6, we determined that the characteristic polynomial of A is
2 −2 0
C A(λ) = − (λ + 2) 2(λ − 1)
and that the eigenvalues of A are λ = 1 and λ = − 2. The matrix A has two eigenspaces E 1 and E − 2. We will compute a basis for each eigenspace.
[ ] [ ]
1
−2 −1 1 0 1 0 −2 0
[A − I | 0 ] = 0 −3 0 0 → ⋯row operations⋯ → 0 1 0 0
2 −2 −1 0
0 0 0 0
1
x1 = t
2
x2 = 0
x 3 = t (free) t∈R
[]
1
2
x=t 0 , t∈R
1
{[ ]}
1
2
B
E1 = 0
1
is
{[ ]}
1
B
E1 = 0
2
[ ] [ ]
1 −1 1 0 → 1 −1 1 0
[A + 2I | 0 ] = 0 0 0 0 R 3 − 2R 1 0 0 0 0
2 −2 2 0 0 0 0 0
1 -1
Correct answer, well done.
x=s 1 +t 0 s, t ∈ R
Marks for this submission: 2.00/2.00.
0 1
1 -1
Correct answer, well done.
B
E −2 = { 1 , 0 } Marks for this submission: 2.00/2.00.
0 1
Check
Question 10
Answer saved
Exercise: Let \(A = {\left[\begin{array}{ccc} 1 & 2 & 0 \\ 0 & 2 & 0 \\ -1 & 0 & 0 \end{array}\right]}\).
Verify that the characteristic polynomial of \(A\) is
\[C_A(\lambda) = -\lambda(\lambda - 1)(\lambda - 2)\]
Thus, the eigenvalues of \(A\) are \(\lambda = 0\), \(\lambda = 1\) and \(\lambda = 2\).
Find bases for the three eigenspaces \(E_0\), \(E_1\), and \(E_2\) and state the algebraic multiplicity and geometric multiplicity of each eigenvalue
of \(A\).
For \(\lambda =0\), we have \(A - \lambda I = A \) and so \(E_{0} = \operatorname{Nul}(A) \).
A basis for the eigenspace \(E_{0}\) is
For \(\lambda =1\), we have \(A - \lambda I = A - I \) and so \(E_{1} = \operatorname{Nul}(A - I) \).
A basis for the eigenspace \(E_{1}\) is
For \(\lambda =2\), we have \(A - \lambda I = A - 2I \) and so \(E_{2} = \operatorname{Nul}(A - 2I) \).
A basis for the eigenspace \(E_{2}\) is
2
Correct answer, well
\(\mathcal{B}_{E_{2}} = \left\{ \begin{array}{c}\\ \\ \\ \(\left. \begin{array}{c}\\ \\ \\ \\
1 done.
\\ \end{array}\right. \) \end{array}\right\}\)
Marks for this submission:
-1
2.00/2.00.
Check
Jump to...