[go: up one dir, main page]
Scribd
Upload
111 views15 pages

Symmetric Matrices and Orthogonal Diagonalization

1. The document discusses symmetric matrices and their orthogonal diagonalization. 2. The main theorem is that any symmetric matrix is diagonalizable with real eigenvalues and orthogonal eigenvectors. 3. The process of orthogonal diagonalization of a symmetric matrix involves finding the eigenvalues and corresponding orthonormal eigenvectors, and using them to form an orthogonal matrix that diagonalizes the original matrix.

Uploaded by

panna1
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
111 views15 pages

Symmetric Matrices and Orthogonal Diagonalization

1. The document discusses symmetric matrices and their orthogonal diagonalization. 2. The main theorem is that any symmetric matrix is diagonalizable with real eigenvalues and orthogonal eigenvectors. 3. The process of orthogonal diagonalization of a symmetric matrix involves finding the eigenvalues and corresponding orthonormal eigenvectors, and using them to form an orthogonal matrix that diagonalizes the original matrix.

Uploaded by

panna1
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/ 15

SYMMETRIC MATRICES

AND ORTHOGONAL
DIAGONALIZATION
Objectives
• Determine an orthogonal diagonalization of a symmetric
matrix.
Review
• If an 𝑛 × 𝑛 matrix 𝐴 has 𝑛 distinct eigenvalues, then 𝐴 is
diagonalizable.
• Why? Because we had a theorem that said the 𝑘 eigenvectors
belonging to 𝑘 distinct eigenvalues are linearly independent.
Another theorem stated that 𝐴 is diagonalizable if and only if 𝐴 has
𝑛 linearly independent eigenvectors.

• If a matrix does NOT have 𝑛 distinct eigenvalues, we


don’t know if 𝐴 is diagonalizable!
• We saw examples for both outcomes, diagonalizable and not
diagonalizable.
Question
• Assume that 𝐴 does not have distinct eigenvalues. Are
there any conditions which will guarantee that 𝐴 is
diagonalizable?
Real Spectral Theorem
Theorem: If 𝐴 is an 𝑛 × 𝑛 symmetric matrix, then the
following properties are true.
1. 𝐴 is diagonalizable
2. All eigenvalues of 𝐴 are real.
3. If 𝜆 is an eigenvalue of 𝐴 with multiplicity 𝑘, then 𝜆 has 𝑘
linearly independent eigenvectors. That is, the
dimension of the eigenspace corresponding to 𝜆 has
dimension 𝑘.
Recall
A square matrix 𝐴 is symmetric if it is equal to its transpose:
𝐴 = 𝐴𝑇
Note
• Your book proves the real spectral theorem this by diving
into Hermitian matrices. I am not going to cover
Hermitian matrices, but feel free to read over the section,
if interested! You will only need to know the theorems in
these slides!
Orthogonality of Eigenvectors of
Symmetric Matrix
Theorem: Let 𝐴 be an 𝑛 × 𝑛 symmetric matrix. If 𝜆1 , 𝜆2 are
distinct eigenvalues of 𝐴, then their corresponding
eigenvectors 𝐱1 , 𝐱2 are orthogonal.

Proof Concepts:
Use fact (𝐴𝐴)𝑇 = 𝐵𝑇 𝐴𝑇 and calculate 𝜆1 𝐱1 , 𝐱2 = 𝜆2 𝐱1 , 𝐱 2
Orthogonally Diagonalizable
Definition: A matrix 𝐴 is orthogonally diagonalizable if
there exists an orthogonal matrix 𝑄 such that 𝐷 = 𝑄−1 𝐴𝐴 is
diagonal.

We will say that 𝐴 has an orthogonal diagonalization if 𝐴 is


orthogonally diagonalizable.
Fundamental Theorem
of Symmetric Matrices
Theorem: Let 𝐴 be an 𝑛 × 𝑛 matrix. Then 𝐴 is orthogonally
diagonalizable and has real eigenvalues if and only if 𝐴 is
symmetric.
Orthogonal Diagonalization of a
Symmetric Matrix
Let 𝐴 be an 𝑛 × 𝑛 symmetric matrix.
1. Find all eigenvalues of 𝐴 and determine the multiplicity of
each.
2. For each eigenvalue of multiplicity 1, choose a unit
eigenvector. (Choose any eigenvector and normalize it.)
3. For each eigenvalue of mulitplicity 𝑘 ≥ 2, find a set of 𝑘
linearly independent eigenvectors. If the set is not
orthonormal, apply the Gram-Schmidt orthonormalization
process.
4. The composite of steps 2 and 3 produces an orthonormal set
of 𝑛 eigenvectors. Use those eigenvectors to form the
columns of 𝑃. Then 𝐷 = 𝑄−1 𝐴𝐴 = 𝑄𝑇 𝐴𝑄 will be diagonal,
and the diagonal entries in 𝐷 are the corresponding
eigenvalues of 𝐴.
Fitting it Together
• What have we done?

• Given a linear transformation 𝐿: ℝ𝑛 → ℝ𝑛 that has a


symmetric matrix representation, we have found an
orthonormal basis 𝐵 such that the matrix representation of
𝐿 relative to 𝐵 is a diagonal matrix!
• Remember that we like orthonormal bases! Remember why?
• Remember that we like diagonal matrix representations!
Remember why?

• Question: If a given matrix representation for 𝐿 is


symmetric, are all matrix representations for
𝐿 symmetric??
Examples
• Find an orthogonal matrix 𝑄 that orthogonally
diagonalizes the matrix:
−2 2
𝐴=
2 1
Facts of the Day
A symmetric matrix is orthogonally diagonalizable.
Keywords
Symmetric, orthogonally diagonalizable

You might also like