Why Are Symmetric & Non-Symmetric Matrices Not Unitarily Equivalent?

  • Thread starter namegoeshere
  • Start date
  • Tags
    Equivalent
In summary, if two matrices are similar and one is Hermitian (conjugate symmetric), then the other must also be Hermitian. This is because unitarily equivalent matrices must have the same symmetry properties. Similarly, if both matrices are real-valued, then you can replace "unitary" with "orthogonal" and "Hermitian" with "symmetric."
  • #1
namegoeshere
6
0
I heard this assertion during a discussion:

If two matrices are similar, but one is symmetric and the other is not, then the two matrices are not unitarily equivalent.

Why is this true? This was only mentioned in passing, and I definitely don't understand why.
 
Last edited:
Physics news on Phys.org
  • #2
I assume that when you say "symmetric" you really mean conjugate symmetric (Hermitian). A matrix ##A## is Hermitian if it equals the complex conjugate of its transpose: ##A = A^*##.

If ##A## and ##B## are unitarily equivalent, then there exists a unitary matrix ##U## (i.e., ##U## satisfies ##U^* U = U U^* = I##) such that ##A = U^* B U##.

If ##A## is Hermitian, then ##A = A^* = U^* B^* U##. Therefore,
$$U^* B U = U^* B^* U$$
Multiplying this equation on the left by ##U## and on the right by ##U^*##, we conclude that ##B = B^*##.

In exactly the same way, we can show that if ##B^* = B##, then ##A^* = A##.

Therefore ##A## is Hermitian if and only if ##B## is Hermitian.

If all the matrices involved are real-valued, then you can replace "unitary" with "orthogonal" and "Hermitian" with "symmetric."
 
  • #3
jbunniii said:
I assume that when you say "symmetric" you really mean conjugate symmetric (Hermitian). A matrix ##A## is Hermitian if it equals the complex conjugate of its transpose: ##A = A^*##.

If ##A## and ##B## are unitarily equivalent, then there exists a unitary matrix ##U## (i.e., ##U## satisfies ##U^* U = U U^* = I##) such that ##A = U^* B U##.

If ##A## is Hermitian, then ##A = A^* = U^* B^* U##. Therefore,
$$U^* B U = U^* B^* U$$
Multiplying this equation on the left by ##U## and on the right by ##U^*##, we conclude that ##B = B^*##.

In exactly the same way, we can show that if ##B^* = B##, then ##A^* = A##.

Therefore ##A## is Hermitian if and only if ##B## is Hermitian.

If all the matrices involved are real-valued, then you can replace "unitary" with "orthogonal" and "Hermitian" with "symmetric."

D'oh. Yes, I meant conjugate symmetric. Thanks. :)
 

FAQ: Why Are Symmetric & Non-Symmetric Matrices Not Unitarily Equivalent?

Why are symmetric and non-symmetric matrices not unitarily equivalent?

Symmetric and non-symmetric matrices are not unitarily equivalent because they have different structures and properties. Unitarily equivalent matrices must have the same eigenvalues and eigenvectors, which is not true for symmetric and non-symmetric matrices. Additionally, unitarily equivalent matrices must have the same determinant and trace, which is also not true for these two types of matrices.

Can a symmetric matrix be unitarily equivalent to a non-symmetric matrix?

No, a symmetric matrix cannot be unitarily equivalent to a non-symmetric matrix. This is because unitarily equivalent matrices must have the same eigenvalues and eigenvectors, which is not possible for a symmetric matrix and a non-symmetric matrix.

What is the significance of unitarily equivalent matrices?

Unitarily equivalent matrices have the same spectral properties, which means they have the same eigenvalues and eigenvectors. This is important in many applications, such as in quantum mechanics and signal processing, where the eigenvalues and eigenvectors of a matrix hold important information.

How can I determine if two matrices are unitarily equivalent?

To determine if two matrices are unitarily equivalent, you can calculate their eigendecompositions and compare the eigenvalues and eigenvectors. If they are the same, then the two matrices are unitarily equivalent. Alternatively, you can also check if the matrices have the same determinant and trace, as these properties are also preserved under unitary equivalence.

Are there any real-world applications of unitarily equivalent matrices?

Yes, there are many real-world applications of unitarily equivalent matrices, such as in quantum computing, image and signal processing, and linear algebra. In these fields, unitarily equivalent matrices are used to simplify calculations and solve problems more efficiently.

Similar threads

Replies
9
Views
2K
Replies
4
Views
7K
Replies
3
Views
11K
Replies
4
Views
7K
Replies
1
Views
2K
Back
Top