Spectral Theory: Kevin James

Spectral Theory
Kevin James
Kevin James Spectral Theory

Definition
Let A ∈ K n×n . A vector v ∈ K n satisfying Av = λv for some λ ∈ K is
called and eigenvector of A with respect to the eigenvalue λ.
Assumption
We will typically assume that the field K is algebraically closed. That is,
all polynomials f (x) ∈ K [x] have a root in K . Recall that C is such a
field.
Proposition
If A ∈ K n×n and K is algebraically closed, then A has an eigenvector.
Remark
By our theory of determinants, we know that (A − λI ) is singular if and
only if det(A − λI ) = 0.
Fact
λ ∈ K is an eigenvalue for A ∈ K n×n if and only if det(A − λI ) = 0.
Example

3 2
Find the eigenvalues and eigenvectors for A = . Analyze the
1 4
long term behavior of repeated application of A.
Definition
The characteristic polynomial of A ∈ K n×n is
pA (x) = det(xI − A).
Remark
For A ∈ K n×n , deg(pA (x)) = n.
Theorem
Eigenvectors of A ∈ K n×n corresponding to distinct eigenvalues are
linearly independent.

Corollary
Suppose that A ∈ K n×n has distinct eigenvalues λ1 , . . . , λt and that for
1 ≤ j ≤ t we have that Bj is a basis for Nul(A − λj ). Then B = ∪tj=1 Bj is
linearly independent and thus is a basis for Span(B).
Definition
If A ∈ K n×n is similar to a diagonal matrix D, we say that A is
diagonalizable.
Theorem
A matrix A ∈ K n×n is diagonalizable if and only if there is a basis
v1 , . . . , vn for K n of eigenvectors of A. (-i.e. iff Span(B) form above is all
of K n ).
Corollary
If A has distinct eigenvalues, then it is diagonalizable.

Fact
Similar matrices have the same characteristic polynomial.
Theorem
Suppose A ∈ K n×n has eigenvalues λ1 , . . . , λn , then
n
X n
Y
tr(A) = λi and det(A) = λi .
i=1 i=1
Corollary
Qt Qt
If pA (x) = i=1 (x − λi )mi , then det(A) = i=1 λm
i and
i
Pt
tr(A) = i=1 mi λi .

Theorem (Spectral Mapping Theorem)
Let A ∈ K n×n have eigenvalue λ and let q(x) ∈ K [x].
1 q(λ) is an eigenvalue of q(A).
2 Conversely, every eigenvalue of q(A) has the form q(λ).
Theorem (Cayley-Hamilton)
pA (A) = 0.
Lemma
Let P, Q ∈ K n×n [x] and let R(x) = P(x)Q(x) (under the usual definition
of polynomial multiplication). If A commutes with the coefficients of Q
then R(A) = P(A)Q(A) (under matrix multiplication).
Lemma
(K [x]))
n×n ∼
= K n×n [x] as rings.

Corollary
There is a unique monic polynomial mA (x) of minimal degree with
mA (A) = 0 and mA (x)|pA (x).
Example
Compute the characteristic and minimum polynomials and eigenvalues
and eigenvectors of the following.
1 A = In .

4 1
2 A= .
−1 2
Proposition
If A ∈ K n×n has only one eigenvalue λ and has n linearly independent
eigenvectors, then A = λI
Note
If A ∈ K 2×2 with tr(A) = 1 and det(A) = 1 then pA (x) = (x − 1)2 has a
single repeated root of λ = 1. There is a 2-parameter family of such
matrices and I is the only member which is diagonalizable.
Definition
A vector v is a generalized eigenvector of A corresponding to the
eigenvalue λ if (A − λI )m v = 0 for some m ∈ N.
Example

4 1
Let A = . Recall that there is only one eigenvalue λ = 3 and
−1
2
1
that spans the eigenspace. Find a generalized eigenvector.
−1

Definition
The algebraic multiplicity of an eigenvector λ of a matrix A is the largest
m such that (x − λ)m divides pA (x)
The geometric multiplicity of λ is dim(NA−λI ).
Note
Let Nm = N(A−λI )m . Then we have
1 (A − λI )Nj ⊆ Nj−1
2 {0} ⊂ N1 ⊂ N2 ⊂ · · · ⊂ Nk = Nk+1 = . . .
Definition
The index of λ is the minimum natural number d such that Nd = Nd+1 .

Definition
If A : X → X , the subspace Y ≤ X is invariant under A if A(Y ) ⊆ Y .
Theorem (Spectral Theorem)

Let A ∈ K n×n where K is algebraically closed. Then there is a basis for
K n of generalized eigenvectors.

Lemma
Suppose that p, q ∈ K [x] which share no common zero in K̄ . Then there
are a, b ∈ K [x] such that ap + bq = 1.
Lemma
Let p, q ∈ K [x] which share no common factor in K̄ and let A ∈ K n×n .
Let Np , Nq and Npq denote the nullspaces of p(A), q(A) and (pq)(A)
respectively. Then,
Npq = Np ⊕ Nq .
Corollary
Let p1 , . . . , pk ∈ K [x] be a collection of polynomials that are pairwise
without a common zero in K̄ . Denote the nullspace of the product
p1 (A)p2 (A) . . . pk (A) by Np1 ...pk . Then,
Np1 ...pk = Np1 ⊕ · · · ⊕ Npk .
Proof
Homework
Theorem
Let A be an n × n matrix with distinct eigenvalues λ1 , . . . , λk of indices
d1 , . . . , dk respectively. Then,
k
Y
mA (x) = (x − λi )di .
i=1
Corollary (Spectral Theorem Refinement)

Let N (j) := Ndj (x − λj )dj . Then the Spectral theorem yeilds
K̄ n = N (1) ⊕ · · · ⊕ N (k) .
Note that dim(N (j) ) is the algebraic multiplicity of λj (to be proved later)
Each N (j) is an invariant subspace of A.

Theorem
Suppose that A, B ∈ K n×n where K is an algebraically closed field and let
Nm (λ) = Nul((A − λI )m ), Mm (λ) = Nul((B − λI )m ),

dm (λ) = dimK (Nm (λ)) em (λ) = dimK (Mm (λ)).
Then, A is similar to B if and only if A and B have the same eigenvalues

λ1 , . . . , λt and for 1 ≤ k ≤ t and all m, dm (λk ) = em (λk ).
Lemma (Key Lemma)

Suppose that A ∈ K n×n with eigenvalue λ and let Nm = Nul((A − λI )m ).
Then,
(A − λI ) : Ni+1 /Ni ,→ Ni /Ni−1 .
Theorem
Let X be an n-dimensional vector space over an algebraically closed field
K . Suppose that A, B : X → X are commuting maps. Then, there is a
basis for X consisting of generalized eigenvalues of A and B.

Corollary
The previous theorem remains true for any number (even infinite) of
pairwise commuting maps.
Proof
Homework
Theorem
Let A, B : X → X be diagonalizable and AB = BA. Then A and B are
simultaneously diagonalizable. That is, ∃P ∈ K n×n and diagonal matrices
DA , DB ∈ K n×n such that
A = PDA P −1 B = PDB P −1 .
Theorem
Suppose that A ∈ K n×n . Then A is similar to T A.

Spectral Theory: Kevin James

Uploaded by

Copyright:

Available Formats

Spectral Theory: Kevin James

Uploaded by

Document Information

Original Description:

Original Title

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

Spectral Theory: Kevin James

Uploaded by

Copyright:

Available Formats

Spectral Theory

Kevin James Spectral Theory

pA (x) = det(xI − A).

Kevin James Spectral Theory

Kevin James Spectral Theory

Kevin James Spectral Theory

Kevin James Spectral Theory

Kevin James Spectral Theory

Kevin James Spectral Theory

Theorem (Spectral Theorem)

Kevin James Spectral Theory

Np1 ...pk = Np1 ⊕ · · · ⊕ Npk .

Corollary (Spectral Theorem Refinement)

Kevin James Spectral Theory

Nm (λ) = Nul((A − λI )m ), Mm (λ) = Nul((B − λI )m ),

Then, A is similar to B if and only if A and B have the same eigenvalues

Lemma (Key Lemma)

Kevin James Spectral Theory

Kevin James Spectral Theory

You might also like

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