Polar decomposition theorem
Statement and proof
As before, \(\mc{M}^{n \times n}, \mc{P}^{n \times n}\) denote the sets of \(n \times n\) matrices and \(n \times n\) symmetric positive semidefinite matrices.
Definition 1 Orthogonal matrices
A matrix \(\bsf{R} \in \mc{M}^{n \times n}\) is orthogonal if the column vectors and row vectors are orthonormal vectors.
Example 1 Verify that the following matrix is orthogonal: \[ \begin{bmatrix} 1/\sqrt{2} & -1/\sqrt{2} & 0 \\ 1/\sqrt{2} & 1/\sqrt{2} & 0 \\ 0 & 0 & 1 \end{bmatrix} \]
Theorem 1 Characterizations of orthogonal matrices
Let \(\bsf{R}\) be an orthogonal matrix. Then the following are true:
1. \(\norm{\bsf{R}\bsf{v}} = \norm{\bsf{v}}\) for all \(\bsf{v} \in \mbb{R}^{n}\).
1. \(\bsf{R} \bsf{u} \cdot \bsf{R} \bsf{v} = \bsf{u} \cdot \bsf{v}\) for all \(\bsf{u}, \bsf{v} \in \mbb{R}^n\).
1. \(\bsf{R}^{\msf{T}} \bsf{R} = \bsf{I}\) and \(\bsf{R} \bsf{R}^{\msf{T}} = \bsf{I}\).
1. \(\bsf{R}^{\msf T}\) is orthogonal.
1. \(\bsf{R}\) is invertible and \(\bsf{R}^{-1} = \bsf{R}^{\msf T}\).
Proof 1. Proof of the characterizations of orthogonal matrices
To be completed.
Definition 2 Special orthogonal group
The special orthogonal group, denoted by \(SO(n)\) is the set of orthogonal matrices whose determinant is +1.
By group, is meant a very specific structure for a set. Namely, there is an operation that acts of two elements of the set to give another element in the set. The operation must be associative, there must exist and identity element, and there must exist a unique inverse element for every element of the set. For \(SO(n)\), the operation is matrix multiplication.
We are now ready to state the polar decomposition theorem.
Theorem 2 Polar decomposition theorem
Let \(\bsf{F} \in \mc{M}^{n \times n}\) be injective. Then there exists a unique matrix \(\bsf{R} \in SO(n)\) and matrices \(\bsf{U}, \bsf{V} \in \mc{P}^{n \times n}\) such that \[ \bsf{F} = \bsf{R} \bsf{U} = \bsf{V} \bsf{R} \]
Proof 2. Proof of the polar decomposition theorem
To be written.