Canonical Form 2

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 19

Mathematics I

[Canonical Forms]

Dr. Anirban Lakshman


Assistant Professor

Department of Mathematics
IIIT Kalyani

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 1 / 19


O UTLINE

1 J ORDAN CANONICAL REPRESENTATION

2 I NVARIANT S UBSPACES

3 R EFERENCES

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 2 / 19


J ORDAN CANONICAL REPRESENTATION

Jordan Block
Consider the following the r-square matrices, where a 6= 0 :
   
λ 1 0 ... 0 0 λ a 0 ... 0 0
0 λ 1 . . . 0 0  0 λ a . . . 0 0 
   
 .. .. .. .. .. ..  .. .. .. .. .. .. 
J(λ, r ) =  . . . . and A =

 . . 
. . . .
 . . 
0 0 0 . . . λ 1  0 0 0 . . . λ a 
0 0 0 ... 0 λ 0 0 0 ... 0 λ
The first matrix, called a Jordan Block, has λ’s on the diagonal,
1’s on the super diagonal , and 0’s elsewhere.
The second matrix A has 1’s on the diagonal, a’s on the super
diagonal, and 0’s elsewhere.
f (t) = (t − λ)r is both the characteristic and minimal polynomial of
both J(λ, r ) and A.

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 3 / 19


J ORDAN CANONICAL REPRESENTATION

Jordan canonical representation

An operator T can be put into Jordan canonical form if its


characteristic and minimal polynomials factor into linear polynomials.
This is always true if the field F is the complex field C.
We can always extend the base field F to afield in which the
characteristic and minimal polynomials do factor into linear factors.
Thus in a broad sense, every operator has a Jordan Canonical
Form.
Analogously every matrix is similar to a matrix in Jordan
Canonical Form.

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 4 / 19


J ORDAN CANONICAL REPRESENTATION
Theorem: Let T : V −→ V be a linear operator whose
characteristic and minimal polynomials are, respectively,
∆(t) = (t − λ1 )n1 (t − λ2 )n2 . . . (t − λr )nr and
m(t) = (t − λ1 )m1 (t − λ2 )m2 . . . (t − λr )mr ;
where the λi are distinct scalars.
Then T has a block diagonal matrix representation J in which
each diagonal entry is a Jordan Block Jij = J(λi ).
For each λi , the corresponding Jij have the following properties:
(i) There is at least one Jij of order mi ; all other Jij are of order
≤ mi .
(ii) The sum of the orders of the Jij is ni .
(iii) The number of jij equals the geometric multiplicity of λi .
(iv ) The number of Jij of each possible order is uniquely
determined by T .
Dr. Anirban Lakshman (IIIT Kalyani) MAC101 5 / 19
J ORDAN CANONICAL REPRESENTATION
E XAMPLE
Suppose the characteristic and minimal polynomials of an operator
T are, respectively,
∆(t) = (t − 2)4 (t − 5)3 and m(t) = (t − 2)2 (t − 5)3
=⇒ Then the Jordan canonical representation of T is one of the
following block diagonal matrices:
  
    5 1 0 
2 1 2 1 
diag , , 0 5 1 or
 0 2 0 2
0 0 5

  
  5 1 0 
2 1  
diag , 2 , 2 0 5 1
 0 2
0 0 5

The first matrix occurs if T has two independent eigenvectors


belonging to the eigenvalue 2; and the second matrix occurs if T has
three independent eigenvectors belonging to the eigenvalue 2.
Dr. Anirban Lakshman (IIIT Kalyani) MAC101 6 / 19
O UTLINE

1 J ORDAN CANONICAL REPRESENTATION

2 I NVARIANT S UBSPACES

3 R EFERENCES

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 7 / 19


I NVARIANT S UBSPACES

Let V be a vector space over a field F and T : V → V be linear


operator on V . A subspace W of V is said to be invariant under T or
T -invariant if T maps W into itself i.e., if v ∈ W , then T (v ) ∈ W .

In this case, T restricted to W defines a linear operator on W ; i.e.,


T induces a linear operator T̂ : W → W defined by T̂ (w) = T (w);
for every w ∈ W

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 8 / 19


I NVARIANT S UBSPACES
Example:Let T : R3 → R3 be the following linear operatorr, which
rotates each vector v about the z-axis by an angle θ :
T (x, y , z) = (x cos θ − y sin θ, x sin θ + y cos θ, z)

F IGURE : Rotator

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 9 / 19


I NVARIANT S UBSPACES

Let W = xy - plane. we observe that each vector w = (a, b, 0) ∈ W


remains in W under the map T . Hence W is T - invariant.
Let U = z-axis= {(0, 0, z) | (0, 0, z) ∈ R3 }
Observe that U is also T - invariant.

The restriction of T to W rotates each vector about the origin O.


Also the restriction of T to U is the identity mapping of U.

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 10 / 19


I NVARIANT S UBSPACES

T HEOREM
Suppose W is an invariant subspace
 of T : V → V . Then T has a
A B
block matrix representation , where A is a matrix
0 C
representation of the restriction T̂ of T to W .

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 11 / 19


I NVARIANT S UBSPACES

Direct Sum
A vector space V is termed as the direct sum of subspaces
W1 , W2 , . . . , Wr ; if every vector v ∈ V can be written uniquely in the
form v = w1 + w2 + . . . + wr with wi ∈ Wi .
In this case V can be written as V = W1 ⊕ W2 ⊕ . . . ⊕ Wr

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 12 / 19


I NVARIANT S UBSPACES

T HEOREM
Suppose W1 , W2 , . . . , Wr are subspaces of V , and suppose
B1 = {w11 , w12 , . . . , w1n1 }, B2 = {w21 , w22 , . . . , w2n2 }, . . . ,
Br = {wr 1 , wr 2 , . . . , wrnr } are bases of W1 , W2 , . . . , Wr respectively.
Then V is the direct sum of the W1 , W2 , . . . , Wr iff
the union B = B1 ∪ B2 ∪ . . . ∪ Br is a basis of V .

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 13 / 19


I NVARIANT S UBSPACES

Now, suppose T : V → V is linear mapping and V is the direct sum


of (non-zero) T -invariant subspaces W1 , W2 , . . . , Wr ;
i.e., V = W1 ⊕ W2 ⊕ . . . ⊕ Wr and T (Wi ) ⊆ Wi , i = 1, 2, . . . , r .
Let Ti denote the restriction of T to Wi .
Then T is said to be decomposable into the operators T1 , T2 , . . . , Tr
or T is said to be the direct sum of the T1 , T2 , . . . , Tr .
Then T can be written as T = T1 ⊕ T2 ⊕ . . . ⊕ Tr .
Also the subspaces W1 , W2 , . . . , Wr are said to reduce T or to form a
T -invariant direct sum decomposition of V .

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 14 / 19


I NVARIANT S UBSPACES

Consider the special case where two subspaces U and W reduce


an operator T : V → V ; say dimU = 2 and dimW = 3, and suppose
{u1 , u2 } and {w1 , w2 , w3 } are bases of U and W ,respectively. If T1
and T2 denote the restrictions of T to U and W , respectively, then
T1 (u1 ) = a11 u1 + a12 u2 T2 (w1 ) = b11 w1 + b12 w2 + b13 w3
T1 (u2 ) = a21 u1 + a22 u2 T2 (w2 ) = b21 w1 + b22 w2 + b23 w3
T2 (w3 ) = b31 w1 + b32 w2 + b33 w3
 
  b11 b21 b31  
a11 a21 A 0
Let A = , B = b12 b22 b32  , M =
a12 a22 0 B
b13 b23 b33
Here A, B, M are the matrix representations of T1 , T2 , T respectively.

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 15 / 19


I NVARIANT S UBSPACES

The block diagonal matrix M results from the fact that


{u1 , u2 , w1 , w2 , w3 } is a basis of V , and that T (ui ) = T1 (ui )
and T (wj ) = T2 (wj )

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 16 / 19


I NVARIANT S UBSPACES

A generalisation of the above argument gives us the following theorem:


T HEOREM
Suppose T : V → V is linear mapping and suppose V is the direct
sum of T -invariant subspaces, say, W1 , W2 , . . . , Wr . If Ai is a matrix
representation of the restriction of T to Wi , then T can be
represented by the block diagonal matrix:
M = diag(A1 , A2 , . . . , Ar )

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 17 / 19


O UTLINE

1 J ORDAN CANONICAL REPRESENTATION

2 I NVARIANT S UBSPACES

3 R EFERENCES

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 18 / 19


R EFERENCES

Gilbert Strang, Linear Algebra and Its Applications.


Steven J. Leon, Linear Algebra with Applications, Pearson.
Carl D. Meyer, Matrix Analysis and Linear Algebra,Siam
Kenneth Hoffman; Ray Kunze, Linear Algebra, Pearson

Dr. Anirban Lakshman (IIIT Kalyani) MAC101 19 / 19

You might also like