0% found this document useful (0 votes)
101 views6 pages

Control Systems: V I A V

This document provides solutions to a problem set on control systems. It finds Jordan forms and transformation matrices for three matrices A1, A2, and A3. It then computes powers and polynomials of these matrices, such as A199, A1100, and polynomials of the form g(A). It proves that two polynomials of the same matrix commute, meaning f(A)g(A)=g(A)f(A). It also proves that if v is an eigenvector of a matrix A with eigenvalue λ, then v is also an eigenvector of any polynomial f(A) of A with the corresponding eigenvalue f(λ).

Uploaded by

Igor Oliveira
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
101 views6 pages

Control Systems: V I A V

This document provides solutions to a problem set on control systems. It finds Jordan forms and transformation matrices for three matrices A1, A2, and A3. It then computes powers and polynomials of these matrices, such as A199, A1100, and polynomials of the form g(A). It proves that two polynomials of the same matrix commute, meaning f(A)g(A)=g(A)f(A). It also proves that if v is an eigenvector of a matrix A with eigenvalue λ, then v is also an eigenvector of any polynomial f(A) of A with the corresponding eigenvalue f(λ).

Uploaded by

Igor Oliveira
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
You are on page 1/ 6

Control Systems

Solution for problem set #5 (Modified from Jinming Chens homework)


1. Find Jordan-form representations and transformation matrix Q for the following
matrices:

1 1 0
2 0 0
1 0 0
A1 0 0 1 , A 2 1 6 4 , A 3 0 1 0
0 0 1
1 9 6
4 2 2
Solution:
For A1, the characteristic polynomial is:
( 1) 2 0 1 0

2, 3 1

For 1=0, we get: 0


0

1 v1 0 v1 1
0
1

1
0
0

For 2=1, from [A-2I]2v3=0, [A-2I]v30, we obtain:


0
0

1
1
0

1
0

1 v3 0 v3 1
1
0

so v 2 A 2 I v3 0

1 1
Q 1 0
0 0

0
A1 Q AQ 0
0
1

and

0
1
0

0
1
1

For A2, the characteristic polynomial is:


2 ( 2) 0 1 2

2,3 0

for 1=2, we get: 1


1

4
9

4
4 v1 0 v1 4
5
8

for 2=0, from [A-2I]2v3=0, [A-2I]2v30, we get:

4 0 0

4 0 0 v 3 0 v 3
5 0 0

0
0
0 , v 2 ( A 2 I )v 3 4
6
1

4 0 0
Q 4 4 0
5 6 1

2 0 0
A 2 Q 1 AQ 0 0 1
0 0 0

and

For A3, the characteristic polynomial is:


( 1) 2 ( 2) 0 1, 2 1

3 2

0 0 0
For 1,2=1. We have A-I= 0 0 0 . Since (A-1I)=1, we obtain
4 2 1
(A-1I)=2. (A-1I)v1=0 has two LI solutions. We choose

v1 2 , v 2
0

0
1 .
2
1

For 3=2, we get: (A-3I)v3 = 0

1
Q 2
0

0
1
2

0
0

and

0 0

1 0 v 3 0 v 3

2 0

1
A 3 Q AQ 0
0
1

0
1
0

0
0
2

0
0
1

2. For the matrices in Problem 1, compute A199, A1100, A29, A210, A311, A312.
For A1, assume f()=k, g()=0+1+22 then
g(0)=f(0), 0=0
g(1)=f(1), 1=1+2
g(1)=f(1), k=1+22
we get: 0=0, 1=2-k, and 2=k-1;
So g(A)=(2-k)A+(k-1)A2
1
A 0
0
2

1
0
0

1
1
1

For k=99, g ( A) 0
0
1

For k=100, g ( A) 0
0

1
0

98
1

1
1 99

0
0

1
1

For A2, assume f()=k, g()=0+1+22 then


g(0)=f(0), 0=0
g(0)=f(0), 0=1
g(2)=f(2), 2k=42
we get: 0=0, 1=0, and 2=2k-2
So g(A)=2k-2A2

4 0 0
A 2 4 0 0
5 0 0

512 0 0

For k=9, g ( A) 512 0 0

640 0 0
1024 0
For k=10, g ( A) 1024 0
1280 0

0
0
0

For A3, assume f()=k, g()=0+1+22 then


g(2)=f(2), 2k=0+21+42
g(1)=f(1), 1=0+1+2
g(1)=f(1), k=1+22
we get: 0=2k-2k, 1=2+3k-2k+1, and 2=2k-k-1
So:

4(2 1)

2(2 k 1)

g ( A) (2 2k ) I ( 2 3k 2 2 ) A (2 k 1) A
k

1
A 0
12
2

0
1
6

0
0
4

0
0
0
1
0
8188 4094 2048
1
0
0

1
0
For k=12, g ( A) 0
16380 8190 4096

For k=11, g ( A)

0
0

2 k

3. Let f(A) and g(A) be two polynomials of A. Show that f(A)g(A)=g(A)f(A), i.e., two
polynomials of the same matrix commute.
Proof:
n

Assume g(A)= i A and f(A)= k A k then


i

k 0

i 0

i 0

k 0

g ( A) f ( A) i A i k A k
n

i A i k A k
i 0 k 0
n m

i k A i k
i 0 k 0
n m

k A k i A i k A k i A i
i 0 k 0
m

k 0 i 0

k A k i Ai
k 0

i 0

f ( A) g ( A)

Another approach: use induction. First, it is easy to verify that


( 0 I 1 A)( 0 I 1 A) )( 0 I 1 A)( 0 I 1 A) 0 0 I ( 0 1 1 0 ) A 1 1 A 2
This validates f(A)g(A)=g(A)f(A) when both of them are first order polynomials.
It is also easy to see that An+1g(A) =g(A)An+1 for any n0 and any polynomial g.
Now assume that f(A)g(A)=g(A)f(A) is true for f(A) of degree n and g(A) of degree
m. Let f1(A)=n+1An+f(A) where f(A) has degree n. Then for any g(A) of degree m,

f1 ( A) g ( A) n 1 A n 1 g ( A) f ( A) g ( A)
g ( A) f 1 ( A) n 1 g ( A) A n 1 g ( A) f ( A) n 1 A n 1 g ( A) f ( A) g ( A) f1 ( A) g ( A)
This validates f1(A)g(A)=g(A)f1(A) for f1(A) of degree n+1 and g(A) of degree m.
Similarly we can show that f1(A)g1(A)=g1(A)f1(A) for f1(A) of degree n+1 and g1(A)
of degree m+1.
The result is proved by repeating the above procedure inductively starting with m=1
and n=1.

4. Let f(A) be a polynomial. Suppose that v is an eigenvector of A with corresponding


eigenvalue , then v is also an eigenvector of f(A) with corresponding eigenvalue
f().
Proof:
From the definition, we have: Av=v. Then
A2v=AAv=Av=Av=v=v,
A3v=AA2v=Av=Av=v=v,
..
Anv=nv,

f ( A)v i A i v
i

i v
i

( i i )v
i

f ( )v
so, we have f ( A)v f ( )v
This means that v is an eigenvector of f(A) with corresponding eigenvalue f(

You might also like