91Ó°ÊÓ

The concept of a matrix norm is essential in understanding how the size or length of a matrix is measured. Unlike vectors, which use simple norms like the Euclidean norm, matrix norms evaluate matrices using slightly more complex rules. A matrix norm is generally denoted as \( \|A\| \), where \( A \) is a given matrix. It must satisfy specific properties:

• Non-negativity: \( \|A\| \geq 0 \), and \( \|A\| = 0 \) if and only if \( A \) is a zero matrix.
• Scalar multiplication: \( \|cA\| = |c| \cdot \|A\| \) for any scalar \( c \).
• Triangle Inequality: \( \|A + B\| \leq \|A\| + \|B\| \) for any matrices \( A \) and \( B \).
• Submultiplicativity: \( \|AB\| \leq \|A\| \cdot \|B\| \) for any matrices \( A \) and \( B \).

These rules make it easier to analyze the behavior of matrices, especially when discussing their properties, such as convergence in a Neumann series or assessing their spectral properties. Matrix norms are particularly useful in numerical analysis for measuring errors and stability of matrix computations.
Understanding and calculating matrix norms help us deduce whether certain matrix expressions, like \( I - \lambda A \), can be expanded into a convergent series, which is central to proving invertibility.

Spectral properties of matrices refer to characteristics that arise from eigenvalues and eigenvectors. The eigenvalues of a matrix \( A \) can provide us with critical insights into its behavior and are central to understanding conditions for invertibility.
An eigenvalue \( \lambda \) of a matrix \( A \) satisfies the equation \( Av = \lambda v \), where \( v \) is a non-zero eigenvector. The collection of all eigenvalues forms the spectrum of the matrix. Matrix invertibility is tied to spectral properties because a matrix is invertible if zero is not an eigenvalue.

• If \( \lambda_0 \) exists such that \( \|I - \lambda_0 A\| < 1 \), then the eigenvalues of \( \lambda_0 A \) are constrained within a unit circle centered at 1.
• This information implies that the matrix \( I - \lambda_0 A \) cannot have eigenvalue zero, supporting its invertibility.

These spectral insights help us understand why the expression \( \inf _{\lambda \in \mathbb{R}}\|I-\lambda A\|<1 \) assures the invertibility of \( A \). By controlling the spectrum of the matrix transformation \( I - \lambda A \) via its norm, we determine vital properties about \( A \) itself.

A Neumann series is a powerful tool in linear algebra used to find the inverse of a matrix expression under specific conditions. The series is particularly useful when dealing with matrices that resemble the identity matrix. The Neumann series for a matrix \( B \) is given by: \[(I - B)^{-1} = I + B + B^2 + B^3 + \cdots \]This series converges if \( \|B\| < 1 \), implying that each subsequent power of \( B \) becomes increasingly negligible.
This concept is critical when proving the invertibility of \( A \) in the exercise:

• Given \( \|I - \lambda_0 A\| < 1 \), we can safely apply the Neumann series \( (I - \lambda_0 A)^{-1} \).
• This convergent series allows us to express the inverse of \( I - \lambda_0 A \) directly.
• Therefore, \( A \) is invertible, since \( I - \lambda_0 A \) can be inverted using the series.

Neumann series provide a practical approach to handle cases where direct matrix inversion is not straightforward, helping us establish the conditions for a matrix's invertibility through the convergence of sums of invertible matrices.