Lab 8. Jordan Canonical Form: Independent Study

This lab accompanies Chapter 8: Jordan Canonical Form.

The goal is to understand what happens when diagonalization fails:

  1. repeated eigenvalues may not give enough eigenvectors;
  2. generalized eigenvectors repair the missing directions;
  3. Jordan chains create Jordan blocks;
  4. powers and exponentials of Jordan blocks have polynomial correction terms;
  5. kernel dimensions reveal Jordan block sizes.

This is an independent-study lab. Each main question includes a worked solution and a similar practice question.

Python practice notebook

You may use the Jupyter notebook version for longer Python practice:

Interactive lab

Study guide and worked questions

Question 1. A matrix that is not diagonalizable

Let

\[ A=\begin{bmatrix}2&1\\0&2\end{bmatrix}. \]

Show that \(A\) is not diagonalizable.

Solution

The characteristic polynomial is

\[ \det(A-\lambda I)=(2-\lambda)^2. \]

So the only eigenvalue is \(\lambda=2\), with algebraic multiplicity \(2\).

Now

\[ A-2I=\begin{bmatrix}0&1\\0&0\end{bmatrix}. \]

The eigenspace is

\[ E_2=\ker(A-2I)=\operatorname{span}\left\{\begin{bmatrix}1\\0\end{bmatrix}\right\}. \]

It has dimension \(1\), not \(2\). Therefore \(A\) does not have enough eigenvectors to form a basis and is not diagonalizable.

Similar practice

Show that

\[ B=\begin{bmatrix}5&1\\0&5\end{bmatrix} \]

is not diagonalizable.

Answer

The only eigenvalue is \(5\). The eigenspace is

\[ \ker(B-5I)=\operatorname{span}\left\{\begin{bmatrix}1\\0\end{bmatrix}\right\}, \]

which is one-dimensional. Hence \(B\) is not diagonalizable.

Question 2. Find a Jordan chain

For

\[ A=\begin{bmatrix}2&1\\0&2\end{bmatrix}, \]

find a Jordan chain.

Solution

Let

\[ v_1=e_1=\begin{bmatrix}1\\0\end{bmatrix}, \qquad v_2=e_2=\begin{bmatrix}0\\1\end{bmatrix}. \]

Then

\[ (A-2I)v_1=0, \qquad (A-2I)v_2=v_1. \]

Thus \(v_1,v_2\) is a Jordan chain for the eigenvalue \(2\).

Similar practice

Find a Jordan chain for

\[ B=\begin{bmatrix}5&1\\0&5\end{bmatrix}. \]

Answer

Use \(v_1=e_1\) and \(v_2=e_2\). Then

\[ (B-5I)v_1=0, \qquad (B-5I)v_2=v_1. \]

Question 3. Powers of a Jordan block

Let

\[ J=\begin{bmatrix}3&1\\0&3\end{bmatrix}. \]

Compute \(J^m\).

Solution

Write

\[ J=3I+N, \qquad N=\begin{bmatrix}0&1\\0&0\end{bmatrix}, \qquad N^2=0. \]

Then

\[ J^m=(3I+N)^m=3^mI+m3^{m-1}N. \]

Therefore

\[ J^m=\begin{bmatrix}3^m&m3^{m-1}\\0&3^m\end{bmatrix}. \]

Similar practice

For

\[ K=\begin{bmatrix}2&1\\0&2\end{bmatrix}, \]

compute \(K^m\).

Answer

\[ K^m=\begin{bmatrix}2^m&m2^{m-1}\\0&2^m\end{bmatrix}. \]

Question 4. Kernel dimensions and Jordan block sizes

Suppose for one eigenvalue \(\lambda\) we have

\[ d_1=2, \qquad d_2=4, \qquad d_3=5, \]

where \(d_j=\dim\ker((A-\lambda I)^j)\). Find the Jordan block sizes.

Solution

Let \(d_0=0\). Then

\[ d_1-d_0=2, \qquad d_2-d_1=2, \qquad d_3-d_2=1. \]

So there are two blocks of size at least \(1\), two blocks of size at least \(2\), and one block of size at least \(3\). Hence the Jordan block sizes are

\[ 3 \quad\text{and}\quad 2. \]

Similar practice

Suppose

\[ d_1=3, \qquad d_2=5, \qquad d_3=6. \]

Find the Jordan block sizes.

Answer

The differences are

\[ 3,2,1. \]

Therefore the block sizes are

\[ 3,2,1. \]

Question 5. Matrix exponential of a Jordan block

Let

\[ J=\begin{bmatrix}\lambda&1\\0&\lambda\end{bmatrix}. \]

Find \(e^{tJ}\).

Solution

Write \(J=\lambda I+N\), where

\[ N=\begin{bmatrix}0&1\\0&0\end{bmatrix}, \qquad N^2=0. \]

Then

\[ e^{tJ}=e^{\lambda t}e^{tN}=e^{\lambda t}(I+tN). \]

Thus

\[ e^{tJ}=e^{\lambda t} \begin{bmatrix}1&t\\0&1\end{bmatrix}. \]

Similar practice

For

\[ J=\begin{bmatrix}2&1\\0&2\end{bmatrix}, \]

find \(e^{tJ}\).

Answer

\[ e^{tJ}=e^{2t}\begin{bmatrix}1&t\\0&1\end{bmatrix}. \]

Short reflection

In your own words, explain this sentence:

A Jordan block is an eigenvalue plus a nilpotent correction.

A good answer should mention the formula

\[ J_{\lambda,k}=\lambda I+N_k \]

and explain how the nilpotent part creates generalized eigenvector chains.