91Ó°ÊÓ

Linear systems are a fundamental concept in mathematics and engineering. They are equations that represent relationships between multiple variables using linear combinations. A typical linear system is expressed as:

• \( A \mathbf{x} = \mathbf{b} \)

Here, \( A \) represents the matrix of coefficients, \( \mathbf{x} \) is the vector of unknowns, and \( \mathbf{b} \) is the vector of constants.
The main goal when dealing with linear systems is to find the vector \( \mathbf{x} \) that satisfies this equation.
For a unique solution to exist, matrix \( A \) must be invertible, meaning it has an inverse that undoes the linear transformation applied by \( A \).
Linear systems are ubiquitous in various fields, such as:

• Physics for solving motion equations,
• Economics for balancing supply and demand,
• Engineering for circuit analysis.

Matrix norms are essential for measuring the size or length of a matrix, which is pivotal in linear algebra.
They help in understanding how much a matrix can stretch or shrink a vector it multiplies.
Common matrix norms include the Frobenius norm and the 2-norm (spectral norm), among others.
Any matrix norm \( \| A \| \) satisfies these properties:

• Non-negativity: \(\| A \| \geq 0\)
• Scalar multiplication: \(\| cA \| = |c| \| A \|\) for scalar \(c\)
• Triangle inequality: \(\| A + B \| \leq \| A \| + \| B \|\)
• Submultiplicativity: \(\| AB \| \leq \| A \| \| B \|\)

The ability to measure the impact of perturbations in a linear system directly ties to how matrix norms illustrate the behavior of matrices under multiplication and addition.
This property is crucial for bounding errors and proving inequalities in linear and numerical analysis.

Error propagation refers to how uncertainties or errors in the input of a mathematical operation affect the final output.
In the context of solving linear systems, a small error in the vector \( \mathbf{b} \) leads to a subsequent error in the solution vector \( \mathbf{x} \).
This is particularly important when dealing with numerical computations where precision is essential.
The relationship between changes in \(\mathbf{b}\) and changes in \(\mathbf{x}\) can be quantified using the relation:

• \( \Delta \mathbf{x} = A^{-1} \Delta \mathbf{b} \)

This equation shows the direct dependency of solution error on input error.
The condition number of the matrix \( A \) helps bound the relative error propagation in the system as follows:

• \( \frac{\| \Delta \mathbf{x} \|}{\| \mathbf{x} \|} \leq \operatorname{cond}(A) \frac{\| \Delta \mathbf{b} \|}{\| \mathbf{b} \|} \)

This inequality demonstrates how the condition number determines the sensitivity of the system to perturbations, emphasizing the need for well-conditioned matrices to minimize errors in numerical solutions.

An invertible matrix, also known as a non-singular or non-degenerate matrix, is a square matrix that has an inverse.
The existence of an inverse is a sign of many desirable properties:

• Unique solution for linear systems,
• Stability in numerical operations,
• Determinant is non-zero.

For a given matrix \( A \), its inverse \( A^{-1} \) satisfies:

• \( A A^{-1} = A^{-1} A = I \)
• where \( I \) is the identity matrix.

Invertibility ensures that any linear transformation represented by \( A \) can be reversed.
In practical applications, it implies that any deviation in solutions due to input changes can be reliably calculated and corrected.
When \( A \) is invertible, solving the system \( A \mathbf{x} = \mathbf{b} \) can be effectively tackled as \( \mathbf{x} = A^{-1} \mathbf{b} \), confirming that the inverse plays a key role in understanding and solving linear problems.