164 CHAPTER 7. CANONICAL FORMS

and so this must be it. This is the process used in beginning differential equations to findsolutions to a system when there is a repeated eigenvalue and the matrix is defective. Ithink it is a fairly easy way to remember things. Other cases are similar.

Example 7.6.4 Find the Jordan form for A =

3 3 0 0 −1−1 1 −3 −4 −32 2 7 7 50 −2 4 7 5−2 0 −10 −13 −9

 .

The eigenvalues are 2,1. This was of course cooked up. You can’t find eigenvaluesin general. The minimum polynomial is λ

3− 5λ2 + 8λ − 4. This can be found through

the methods described earlier. It equals (λ −1)(λ −2)2. The eigenvectors for λ = 2 are(− 3

2 t4− 12 t5 1

2 t4 + 12 t5 −t4− t5 t4 t5

)T. I can get two independent eigenvectors,

(−3 1 −2 2 0

)T,(−1 1 −2 0 2

)T.

However, this will not be enough so I look for generalized eigenvectors which yield theseeigenvectors. Two of these are respectively(

−3 0 −2 2 0)T

,(−1 0 0 0 0

)T.

An eigenvector for λ = 1 is(

1 2 −3 −4 8)T

. Then using this ordered basisordered as eigenvector followed by generalized eigenvector, we obtain the matrix of A withrespect to this basis as follows

−3 −3 −1 −1 11 0 1 0 2−2 −2 −2 0 −32 2 0 0 −40 0 2 0 8



−13 3 0 0 −1−1 1 −3 −4 −32 2 7 7 50 −2 4 7 5−2 0 −10 −13 −9

 ·−3 −3 −1 −1 11 0 1 0 2−2 −2 −2 0 −32 2 0 0 −40 0 2 0 8

=

2 1 0 0 00 2 0 0 00 0 2 1 00 0 0 2 00 0 0 0 1

which is the Jordan form.

7.7 Exercises1. In the discussion of Nilpotent transformations, it was asserted that if two n× n ma-

trices A,B are similar, then Ak is also similar to Bk. Why is this so? If two matricesare similar, why must they have the same rank?

164 CHAPTER 7. CANONICAL FORMSand so this must be it. This is the process used in beginning differential equations to findsolutions to a system when there is a repeated eigenvalue and the matrix is defective. Ithink it is a fairly easy way to remember things. Other cases are similar.3 3 0 0 -!lExample 7.6.4 Find the Jordan form for A = 2 2 7 7 5—2 0 —-10 —-13 —-9The eigenvalues are 2,1. This was of course cooked up. You can’t find eigenvaluesin general. The minimum polynomial is 4° — 547 + 8A —4. This can be found throughthe methods described earlier. It equals (A — 1) (A —2)*. The eigenvectors for A = 2 areT( —3t4 — 5ts 514 + 5ts —t4—ts ta ts ) . [can get two independent eigenvectors,(-3 1 2 2 0)'.(-1 1 2 0 2)".However, this will not be enough so I look for generalized eigenvectors which yield theseeigenvectors. Two of these are respectively30220), 10000).( )-( )TAn eigenvector for A = 1 is ( 1 2 -3 -4 8 ) . Then using this ordered basisordered as eigenvector followed by generalized eigenvector, we obtain the matrix of A withrespect to this basis as follows—3 -3 -1 -1 1 3 3 0 O -il1 0 1 0 2 -l tl -3 -4 -3—2 -2 -2 0 -3 2 2 7 7 52 2 0 0 -4 0 —-2 4 7 50 O 2 O 8 —2 0 -10 -13 —-9-3 -3 -1 -1 1 2 100 01 0 1 0 2 0 2 0 0 0—2 —-2 -2 0 -3 |=] 002 1 02 2 0 0 -4 000 2 00 O 2 O 8 000 0 1which is the Jordan form.7.7 Exercises1. In the discussion of Nilpotent transformations, it was asserted that if two n x n ma-trices A, B are similar, then A‘ is also similar to B‘. Why is this so? If two matricesare similar, why must they have the same rank?