Generalized

Eigenvectors

Math 240

Definition

Computation
and Properties

Chains

Jordan
Generalized Eigenvectors
canonical form

Math 240 — Calculus III

Summer 2015, Session II

Thursday, July 23, 2015

 Generalized
Eigenvectors Agenda
Math 240

Definition

Computation
and Properties

Chains
1. Definition
Jordan
canonical form

2. Computation and Properties

3. Chains

4. Jordan canonical form

 Generalized
Eigenvectors Motivation
Math 240

Definition
Defective matrices cannot be diagonalized because they do not
Computation
and Properties possess enough eigenvectors to make a basis. How can we
Chains correct this defect?
Jordan
canonical form Example  
1 1
The matrix A = is defective.
0 1
1. Only eigenvalue is λ = 1.
 
0 1
2. A−I =
0 0
3. Single eigenvector v = (1, 0).
4. We could use u = (0, 1) to complete a basis.
5. Notice that (A − I)u = v and (A − I)2 u = 0.
Maybe we just didn’t multiply by A − λI enough times.
 Generalized
Eigenvectors Definition
Math 240

Definition Definition
Computation If A is an n × n matrix, a generalized eigenvector of A
and Properties

Chains
corresponding to the eigenvalue λ is a nonzero vector x
Jordan
satisfying
canonical form (A − λI)p x = 0
for some positive integer p. Equivalently, it is a nonzero
element of the nullspace of (A − λI)p .

Example
I Eigenvectors are generalized eigenvectors with p = 1.
I In the previous example we saw that v = (1, 0) and
u = (0, 1) are generalized eigenvectors for
 
1 1
A= and λ = 1.
0 1
 Generalized
Eigenvectors Computing generalized eigenvectors
Math 240

Definition
Example
Computation Determine generalized eigenvectors for the matrix
and Properties  
Chains
1 1 0
Jordan A = 0 1 2 .
canonical form
0 0 3

1. Characteristic polynomial is (3 − λ)(1 − λ)2 .

2. Eigenvalues are λ = 1, 3.
3. Eigenvectors are
λ1 = 3 : v1 = (1, 2, 2),
λ2 = 1 : v2 = (1, 0, 0).
4. Final generalized eigenvector will a vector v3 6= 0 such
that
(A − λ2 I)2 v3 = 0 but (A − λ2 I) v3 6= 0.
Pick v3 = (0, 1, 0). Note that (A − λ2 I)v3 = v2 .
 Generalized
Eigenvectors Facts about generalized eigenvectors
Math 240

Definition

Computation
and Properties

Chains

Jordan
How many powers of (A − λI) do we need to compute in order
canonical form to find all of the generalized eigenvectors for λ?
Fact
If A is an n × n matrix and λ is an eigenvalue with algebraic
multiplicity k, then the set of generalized eigenvectors for
λ
consists of the nonzero elements of nullspace (A − λI)k .
In other words, we need to take at most k powers of A − λI to
find all of the generalized eigenvectors for λ.
 Generalized
Eigenvectors Computing generalized eigenvectors
Math 240

Definition
Example
Computation
and Properties Determine generalized eigenvectors for the matrix
Chains  
1 2 0
Jordan
canonical form A = 1 1 2 .
0 −1 1

1. Single eigenvalue of λ = 1.
2. Single eigenvector v1 = (−2, 0, 1).
3. Look at  
2 0 4
(A − I)2 =  0 0 0
−1 0 −2
to find generalized eigenvector v2 = (0, 1, 0).
4. Finally, (A − I)3 = 0, so we get v3 = (1, 0, 0).
 Generalized
Eigenvectors Facts about generalized eigenvectors
Math 240

Definition

Computation The aim of generalized eigenvectors was to enlarge a set of

and Properties
linearly independent eigenvectors to make a basis. Are there
Chains
always enough generalized eigenvectors to do so?
Jordan
canonical form
Fact
If λ is an eigenvalue of A with algebraic multiplicity k, then
 
nullity (A − λI)k = k.

In other words, there are k linearly independent generalized

eigenvectors for λ.
Corollary
If A is an n × n matrix, then there is a basis for Rn consisting
of generalized eigenvectors of A.
 Generalized
Eigenvectors Computing generalized eigenvectors
Math 240

Definition

Computation
and Properties

Chains Example
Jordan Determine generalized eigenvectors for the matrix
canonical form  
1 2 0
A = 1 1 2 .
0 −1 1

1. From last time, we have eigenvalue λ = 1 and eigenvector

v1 = (−2, 0, 1).
2. Solve (A − I)v2 = v1 to get v2 = (0, −1, 0).
3. Solve (A − I)v3 = v2 to get v3 = (−1, 0, 0).
 Generalized
Eigenvectors Chains of generalized eigenvectors
Math 240

Definition
Let A be an n × n matrix and v a generalized eigenvector of A
Computation corresponding to the eigenvalue λ. This means that
and Properties

Chains
(A − λI)p v = 0
Jordan for a positive integer p.
canonical form
If 0 ≤ q < p, then
(A − λI)p−q (A − λI)q v = 0.
That is, (A − λI)q v is also a generalized eigenvector
corresponding to λ for q = 0, 1, . . . , p − 1.
Definition
If p is the smallest positive integer such that (A − λI)p v = 0,
then the sequence
(A − λI)p−1 v, (A − λI)p−2 v, . . . , (A − λI) v, v
is called a chain or cycle of generalized eigenvectors. The
integer p is called the length of the cycle.
 Generalized
Eigenvectors Chains of generalized eigenvectors
Math 240

Definition Example
Computation
and Properties
In the previous example,
 
Chains 0 2 0
Jordan A − λI = 1 0 2
canonical form
0 −1 0
and we found the chain
     
−1 0 −2
v =  0  , (A − λI)v = −1 , (A − λI)2 v =  0  .
0 0 1

Remark
The terminal vector in a chain is always an eigenvector.

Fact
The generalized eigenvectors in a chain are linearly
independent.
 Generalized
Eigenvectors Introduction to Jordan form
Math 240

Definition

Computation
and Properties
What’s the analogue of diagonalization for defective matrices?
Chains
That is, if {v1 , v2 , . . . , vn } are the linearly independent
Jordan
canonical form generalized eigenvectors of A occurring in  chains, what does
−1

the matrix S AS look like, where S = v1 v2 · · · vn ?

Suppose that v1 , v1 , . . . , vk is a chain of generalized

eigenvectors, so that (A − λI)vi = vi−1 for i > 1 and
(A − λI)v1 = 0. Then we have
Avi = λvi + vi−1 for i > 1
and Av1 = λv1 .
 Generalized
Eigenvectors Jordan blocks
Math 240

Definition

Computation
and Properties The matrix for T (x) = Ax with respect to a basis consisting of
Chains a chain of generalized eigenvectors will be a Jordan block:
Jordan
canonical form Definition
If λ is a real number, then the square matrix of the form
 
λ 1 0 0 ··· 0
0 λ 1
 0 ··· 0 
0 0 λ 1 ··· 0
Jλ =  . .
 
. . .. . .. ... .. 
. . . .
 
0 0 · · · · · · λ 1
0 0 ··· ··· 0 λ
is called a Jordan block corresponding to λ.
 Generalized
Eigenvectors Jordan canonical form
Math 240

Definition
In general, we will need to find more than one chain of
Computation
and Properties generalized eigenvectors in order to have enough for a basis.
Chains Each chain will be represented by a Jordan block.
Jordan
canonical form Definition
A square matrix consisting of Jordan blocks centered along the
main diagonal and zeros elsewhere is said to be in Jordan
canonical form (JCF).

Theorem
If S is the matrix whose columns are a basis of generalized
eigenvectors of A arranged in chains, then S −1 AS is a matrix
in JCF. It is unique up to a rearrangement of the Jordan blocks.
We may therefore refer to this matrix as the Jordan canonical
form of A, and we see that every matrix is similar to a matrix
in JCF.
 Generalized
Eigenvectors Examples
Math 240
I The matrix
Definition
 
Computation 2 1 0
and Properties 0 2 1 
Chains
0 0 2 
 
Jordan
 5 1 
canonical form 0 5
 
 

 7 1 

 0 7 
7
 
9

is in JCF. It contains five Jordan blocks.

I Any diagonal matrix is in JCF. All of its Jordan blocks are
1 × 1.
I The matrix " #
0 1 0
0 0 0
0 0 0
is in JCF. It has two blocks of sizes 2 and 1.
 Generalized
Eigenvectors Uses of JCF
Math 240

Definition

Computation
and Properties

Chains

Jordan
canonical form Theorem
Two n × n matrices are similar if and only if they have the
same Jordan canonical form (up to a rearrangement of the
Jordan blocks).

Our main use for JCF will be solving x0 = Ax when the matrix
A is defective.