Thanks to visit codestin.com
Credit goes to www.scribd.com

Scribd
Upload
193 views2 pages

An Example of Jordan Basis

The document provides two examples of computing the Jordan normal form and Jordan basis for linear maps on vector spaces of polynomials. In the first example, the linear map α maps polynomials of degree ≤ 2 to their second derivatives minus 4 times the original polynomial. Computing the matrix representation of α with respect to the standard basis reveals its Jordan normal form is a 2x2 block with eigenvalue -4. A Jordan basis is then computed. The second example considers the linear map L that takes the derivative of polynomials of degree ≤ 5. Without computation, properties of the derivative imply the Jordan normal form is a 6x6 block with eigenvalue 0, and a Jordan basis is constructed.

Uploaded by

bebs_988
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
193 views2 pages

An Example of Jordan Basis

The document provides two examples of computing the Jordan normal form and Jordan basis for linear maps on vector spaces of polynomials. In the first example, the linear map α maps polynomials of degree ≤ 2 to their second derivatives minus 4 times the original polynomial. Computing the matrix representation of α with respect to the standard basis reveals its Jordan normal form is a 2x2 block with eigenvalue -4. A Jordan basis is then computed. The second example considers the linear map L that takes the derivative of polynomials of degree ≤ 5. Without computation, properties of the derivative imply the Jordan normal form is a 6x6 block with eigenvalue 0, and a Jordan basis is constructed.

Uploaded by

bebs_988
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Jordan basis: An example

There is a problem from exam for 2006 which asks to compute the Jordan normal form in a
relatively simple situation, but which still gives occasion to re-iterate some important basic points
from the theory. In fact, we will solve the problem here in two different ways and also compute a
Jordan basis for the vector space.
Let
V = {a0 + a1 x + a2 x2 | ai C},
the vector space of polynomials of degree 2 with complex coefficients and

: V V

be the linear map defined by


d2
(f ) =f 4f
dx2
The first method is to reduce the problem to manipulation of ordinary matrices. That is, if B =
{1, x, x2 }, the obvious basis for V , then

(1) = 4, (x) = 4x, (x2 ) = 2 4x2

so that
4 0 2
A := []B = 0 4 0
0 0 4
From this one computes easily that
ch (x) = (x + 4)3
So 4 is the only eigenvalue, and for the minimal polynomial,

0 0 2
A + 4I = 0 0 0
0 0 0

so that (4I + A)2 = 0. Therefore, m (x) = (x + 4)2 . So we see that the size of the largest Jordan
block is 2 2, and the Jordan normal form must be

4 1 0
J = 0 4 0
0 0 4

Now we compute a Jordan basis in C3 with respect to A. First we find a basis for

Ker(A + 4I) = 0

that is, the 4 eigenspace. The condition for (x, y, z)t to be in this kernel is just 2z = 0, so that a
basis is given by (1, 0, 0)t and (0, 1, 0)t . Since (A + 4I)2 = 0, the second generalized eigenspace is all
of C3 , so that we can enlarge the previous basis B1 to B = B1 B2 where B2 = {(0, 0, 1)t }. We have
obtained thereby a pre-Jordan basis. Now we change this to a Jordan basis by applying (A + 4I) to
B2 to get (2, 0, 0)t and changing B1 thereby to {(2, 0, 0)t , (0, 1, 0)t }. The corresponding Jordan basis
must be ordered as
BJ = {b1 = (2, 0, 0)t , b2 = (0, 0, 1)t , b3 = (0, 1, 0)t }.
Then we have
(A + 4I)(b1 ) = 0, (A + 4I)(b2 ) = b1 , (A + 4I)b3 = 0

1
so that the corresponding matrix is

0 1 0
[A + 4I]JB = 0 0 0
0 0 0

Therefore,
4 1 0
[A]JB = 4I + [A + 4I]JB = 0 4 0
0 0 4
which is the Jordan canonical form.
For the original vector space V , the Jordan basis, of course, is then {2, x2 , x}.
d2
Here is another way to find this Jordan basis. Note that + 4I = dx 2 so that by general properties

of the derivative of polynomials, the kernel of + 4I is spanned by 1, x. Also, ( + 4I)2 is then the
fourth derivative, so that it clearly kills the whole vector space V . Thus, {1, x, x2 } is a pre-Jordan
basis. But then ( + 4I)(x2 ) = 2 so that the eventual Jordan basis becomes {2, x2 , x}.
Let us work out the Jordan basis in an even simpler case to get a feel for less computational
approaches. Let
d
L= : W W
dx
where W is the Cvector space of complex polynomials of degree 5. Then {1, x, . . . , x5 } is a basis
and W has dimension 6. Of course L6 = 0. If l is an eigenvalue with eigenvector v, then 0 = L6 v = l6 v,
so we must have l = 0. Thus
chL (X) = x6 .
On the other hand, L5 (x5 ) = 5! 6= 0. So
mL (x) = x6
This tell us that the Jordan normal form is

0 1 0 0 0 0
0 0 1 0 0 0

0 0 0 1 0 0

0 0 0 0 1 0

0 0 0 0 0 1
0 0 0 0 0 0

To find a Jordan basis, note that L(x5 ) = 5x4 , L(x4 ) = 4x3 , and so on, so that the given basis
is already almost a Jordan basis (actually, it is a pre-Jordan basis). We just need to adjust the
coefficients to get Jordan basis: Replace, x4 by 5x4 . Replace x3 by 4 5x3 , and so on. That is,

{b1 = 5!, b2 = 5!x, b3 = 5 4 3x2 , b4 = 5 4x3 , b5 = 5x4 , b6 = x5 }

is clearly a Jordan basis. That is,

(L + 0I)b6 = b5 , (L + 0I)b5 = b4 , , (L + 0I)b2 = b1 , (L + 0I)b1 = 0

You might also like