Scribd
Upload
26 views

Ij Ji: Typeset by Ams-Tex

The document discusses normal operators and orthonormal bases of eigenvectors. It defines the adjoint of a linear operator T, denoted T*, and proves that if T and S are represented by matrices using orthonormal bases, then Sij = Tji. It then proves that a linear operator T on a finite-dimensional complex vector space has an orthonormal basis of eigenvectors if and only if T*T = TT*. Finally, it discusses special classes of normal operators, including self-adjoint, skew-adjoint, and unitary operators.

Uploaded by

salman suhail
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
26 views

Ij Ji: Typeset by Ams-Tex

The document discusses normal operators and orthonormal bases of eigenvectors. It defines the adjoint of a linear operator T, denoted T*, and proves that if T and S are represented by matrices using orthonormal bases, then Sij = Tji. It then proves that a linear operator T on a finite-dimensional complex vector space has an orthonormal basis of eigenvectors if and only if T*T = TT*. Finally, it discusses special classes of normal operators, including self-adjoint, skew-adjoint, and unitary operators.

Uploaded by

salman suhail
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/ 3

Normal Operators and

Orthonormal Bases of Eigenvectors


Let V and W be a finite-dimensional complex Euclidean vector spaces.
Theorem 1. Let T : V W be a linear map. Then there is a unique linear map
S : W V such that
(1)

(T v, w) = (v, Sw) for all v V and w W .

If we use orthonormal bases for V and for W to represent T and S by matrices, then
Sij = Tji .
Definition. The map S is called the adjoint of T and is denoted T .
Proof. Let e1 , . . . , en and f1 , . . . , fm be orthonormal bases for V and for W , respectively.
Then
X
X
(T v, w) = (T (
vi ei ),
wj fj )
X
=
vi wj (T ei , fj )
i,j

vi wj (the fj component of T ei )

vi wj Tji .

i,j

(2)

i,j

Almost exactly the same calculation shows that


X
(3)
(v, Sw) =
vi wj Sij .
i,j

Thus if we let Sij is the conjugate of Tji (for all i and j), then (2) and (3) will be equal
for all v and w. This establishes the existence of a map S with property (1).
Conversely, if S has the property (1), then (2) and (3) will be equal for all v and w. In
particular, they will be equal when v = ei and w = fj , in which case (2) is Tji and (3) is
the conjugate of Sij . This proves uniqueness. 
Theorem 2. Let V be a nite-dimensional complex Euclidean vector space. A linear map
T : V V has an orthonormal basis of eigenvectors if and only if T T = T T .
Remark. A linear transformation T : V V such that T T = T T is called normal.
Typeset by AMS-TEX
1

Lemma 1. If T : V V is normal, then |T v| = |T v| for all v.


Proof of lemma 1.
|T v|2 = (T v, T v) = (T T v, v) = (T T v, v) = (T v, T v) = |T v|2

Lemma 2. Suppose T : V V is normal. If v is an eigenvector of T with eigenvalue ,
In other words,
then it is also an eigenvector of T , with eigenvalue .

T v = v = T v = v.

Proof. If T is normal, then so is (T I). Thus by lemma 1 applied to T I,

|(T I)v| = |(T I) v| = |T v v|


Since T v = v, the left side is 0. Therefore the right side is 0, which means that T v =

v.

Proof of theorem 2. Any linear transformation of a complex vector space has at least one
eigenvector. So let v1 be a unit eigenvector for T .
Now suppose we have orthonormal eigenvectors v1 , . . . , vk for T . If k = dim V , we are
done. If not, let W be the set of all vectors that are orthogonal to v1 , . . . , vk :
W = {x V : (x, vi ) = 0 for i = 1, . . . , k}.
Then W is an n k dimensional subspace (where n = dim V .)
Claim: If x is perpendicular to vi , then so is T x.
Proof of claim. If x is perpendicular to vi , then
(T x, vi ) = (x, T vi ) = (x, i vi ) = (x, vi ) = 0.
Thus (T x, vi ) = 0, so T x is perpendicular to vi . This proves the claim.
Consequently, if x is perpendicular to all the vi s, then so is T x. In other words,
x W = T x W.
Now since T maps W into itself, W contains a unit eigenvector vk+1 .
Continuing in this way, we get an orthonormal basis of eigenvectors.

Special Classes of Normal Transformations


Note that if T = T , or if T = T , or if T = T 1 , then T is normal. These conditions
define the following classes of normal operators:
1. The eigenvalues of a normal operator T are real if and only if T = T . Such a T is
called self-adjoint or Hermitian.
2. The eigenvalues of a normal operator T are imaginary if and only if T = T . Such a
T is called skew-adjoint or skew-Hermitian.
3. The eigenvalues of a normal operator T are unit complex numbers (i.e., have norms 1)
if and only if T = T 1 . Such a T is called unitary.
Unitary operators have other important characterizations:
Proposition. The T : V V be an operator on the nite dimensional complex Euclidean
space V . The following are equivalent:
(1) T is unitary.
(2) (T u, T v) = (u, v) for all vectors u and v in V .
(3) |T v| = |v| for all vectors v in V .
Proof. Exercise.

You might also like