042 jordex
PDF · 4 pages · 83.3 KB
Open PDF file
A section of linear algebra course notes, apparently from a text by Matthews, titled Two Jordan Canonical Form Examples. Example (a) takes a 4x4 rational matrix with characteristic polynomial (x-2)^2(x-3)^2, uses dot diagrams, kernels and cyclic subspaces to build a basis P, and verifies P^-1AP = J1(2)+J1(2)+J2(3). Example (b) treats a 6x6 matrix with minimal polynomial x^3 and given nullities, giving J3(0)+J2(0)+J1(0) and a left-to-right algorithm for choosing the basis.
AI-written summary; may contain errors.
Extracted text (machine-read; may contain errors)
4.2 Two Jordan Canonical Form Examples
4.2.1 Example (a):
LetA=2
6644 0 1 0
2 2 3 0
1 0 2 0
4 0 1 23
7752M44(Q).
We ndchT= (x 2)2(x 3)2=p2
1p2
2, wherep1=x 2; p2=x 3.
CASE 1,p1=x 2:
p1(A) =A 2I4=2
6642 0 1 0
2 0 3 0
1 0 0 0
4 0 1 03
775!2
6641 0 0 0
0 0 1 0
0 0 0 0
0 0 0 03
775;
so(p1(A)) =
1= 2. Hence b1= 1 and the corresponding dot diagram has
height 1, width 2, with associated Jordan blocks J1(2)J1(2):
N1;x 2
We ndv11=2
6640
1
0
03
775andv12=2
6640
0
0
13
775form a basis for Ker p1(TA) =
N(A 2I4) andmTA;v11=mTA;v12=x 2. Also
Ker (pb1
1(TA)) =N(p1(A)) =N(A 2I4) =CTA;v11CTA;v12:
Note thatCTA;v11andCTA;v12have Jordan bases 11:v11and12:v12
respectively.
CASE 2,p2=x 3:
p2(A) =A 3I4=2
6641 0 1 0
2 1 3 0
1 0 1 0
4 0 1 13
775!2
6641 0 0 1
3
0 1 01
3
0 0 11
3
0 0 0 03
775;
so(p2(A)) = 1 =
2; also(p2
2(A)) = 2. Hence b2= 2 and we get a
corresponding dot diagram consisting of two vertical dots, with associated
Jordan block J2(3):
N2;x 3
N1;x 3
71
We have to nd a basis of the form p2(TA)(v21) = (A 3I4)v21for Kerp2(TA) =
N(A 3I4).
To ndv21we rst get a basis for N(A 3I4)2. We have
p2
2(A) = (A 3I4)2=2
6640 0 0 0
3 1 4 0
0 0 0 0
1 0 2 13
775!2
6641 0 2 1
0 1 10 3
0 0 0 0
0 0 0 03
775
and we nd X1=2
6642
10
1
03
775andX2=2
6641
3
0
13
775is such a basis. Then we have
N2;p2=hp2X1; p2X2i
=hp2(A)X1; p2(A)X2i=h(A 3I4)X1;(A 3I4)X2i
=*2
6643
3
3
93
775;2
6641
1
1
33
775+
=*2
6643
3
3
93
775+
:
Hence we can take v21=X1. ThenmTA;v21= (x 3)2. Also
Kerpb2
2(TA)) =N(p2
2(A)) =N(A 3I4)2=CTA;v21:
MoreoverCTA;v21has Jordan basis 21:v21;(A 3I4)v21.
Finally we have V4(Q) =CTA;v11CTA;v12CTA;v21and=11[12[
21is a basis for V4(Q). Then with
P= [v11jv12jv21j(A 3I4)v21] =2
6640 0 2 3
1 0 10 3
0 0 1 3
0 1 0 93
775
we have
P 1AP= [TA]
=J1(2)J1(2)J2(3) =2
6642 0 0 0
0 2 0 0
0 0 3 0
0 0 1 33
775:
72
4.2.2 Example (b):
LetA2M66(F) have the property that chA=x6; mA=x3and
(A) = 3; (A2) = 5;((A3) = 6):
Next, with h;x= dimFNh;xwe have
1;x=(A) = 3 =
1;
2;x=(A2) (A) = 5 3 = 2;
3;x=(A3) (A2) = 6 5 = 1:
Hence the dot diagram corresponding to the (only) monic irreducible factor
xofmAis
N3;x
N2;x
N1;x
Hence we read o that 9a non-singular P2M66(F) such thatP 1AP=
J3(0)J2(0)J1(0). To nd such a matrix Pwe proceed as follows:
(i) First nd a basis for N3;x. We do this by rst nding a basis for
N(A3):X1; X2; X3; X4; X5; X6. Then
N3;x=hA2X1; A2X2; A2X3; A2X4; A2X5; A2X6i:
We now apply the LRA (left{to{right algorithm) to the above spanning
family to get a basis A2v11forN3;x, whereA2v11is the rst non{zero vector
in the spanning family.
(ii) Now extend the linearly independent family A2v11to a basis for N2;x.
We do this by rst nding a basis Y1; Y2; Y3; Y4; Y5forN(A2). Then
N2;x=hAY1; AY 2; AY 3; AY 4; AY 5i:
We now attach A2v11to the head of this spanning family:
N2;x=hA2v11; AY 1; AY 2; AY 3; AY 4; AY 5i
and apply the LRA to nd a basis for N2;xwhich includes A2X1. This
will have the form A2v11; Av 12, whereAv12is the rst vector in the list
AY1;:::;AY 5which is not a linear combination of A2v11.
(iii) Now extend the linearly independent family A2v11; Av 12to a basis
forN1;x=N(A). We do this by rst nding a basis Z1; Z2; Z3forN(A).
73
Then place the linearly independent family A2v11; Av 12at the head of this
spanning family:
N1;x=hA2v11; Av 12; Z1; Z2; Z3i:
The LRA is then applies to the above spanning family selects a basis of the
formA2v11; Av 12; v13, wherev13is the rst vector among Z1; Z2; Z3which
is not a linear combination of A2v11andAv12.
ThenmTA;v11=x3; mTA;v12=x2; mTA;v13=x. Also
Kerpb1
1(TA) =N(A3) =CTA;v11CTA;v12CTA;v13:
Finally, if we take Jordan bases
11:v11; Av 11; A2v11;
12:v12; Av 12;
13:v13
for the three T{cyclic subspaces CTA;v11; CTA;v12; CTA;v13, respectively, we
then get the basis
=11[12[13
=v11; Av 11; A2v11;v12; Av 12;v13
forV6(F). Then if
P= [v11jAv11jA2v11jv12jAv12jv13]
we have
P 1AP= [TA]
=J3(0)J2(0)J1(0)
=2
66666640 0 0 0 0 0
1 0 0 0 0 0
0 1 0 0 0 0
0 0 0 0 0 0
0 0 0 1 0 0
0 0 0 0 0 03
7777775:
74