Chapter 02: Matrices and Matrix Operations

Prerequisites: Linear Equations (Ch01)

Chapter Outline: Definition and Notation → Basic Operations (Addition, Scalar Multiplication, Multiplication) → Non-commutativity of Multiplication → Transpose and its Properties → Special Matrices (Identity, Diagonal, Triangular, Symmetric) → Elementary Matrices & Row Operations → Inverse Matrices: Definition and Properties → Gauss-Jordan Method for Inversion → Block Matrix Operations → Trace of a Matrix

Extension: Matrices are not just containers for data but representation of linear operators; the definition of matrix multiplication reflects the composition of linear transformations (Ch05).

If systems of linear equations are the language of linear algebra, then matrices are its notation system. A matrix condenses complex linear relationships into concise rectangular arrays and endows them with a set of sophisticated algebraic rules. This chapter establishes the standard axioms of matrix algebra and explores the essential tool of the inverse matrix.

---

02.1 Basic Definitions and Operations

Definition 02.1 (Matrix)

An \(m \times n\) matrix is a rectangular array of \(m \cdot n\) elements arranged in \(m\) rows and \(n\) columns. Usually denoted by uppercase letters \(A, B\).

Definition 02.2 (Matrix Multiplication)

If \(A\) is an \(m \times n\) matrix and \(B\) is an \(n \times p\) matrix, their product \(C = AB\) is an \(m \times p\) matrix with entries: $\(c_{ij} = \sum_{k=1}^n a_{ik} b_{kj}\)$ WARNING: Matrix multiplication is generally not commutative, i.e., \(AB \neq BA\).

---

02.2 Special Matrix Classes

Definition 02.3 (Special Matrices)

1. Identity Matrix \(I\): Diagonal entries are 1, all others 0. Satisfies \(AI = IA = A\).
2. Symmetric Matrix: Satisfies \(A^T = A\).
3. Skew-symmetric Matrix: Satisfies \(A^T = -A\).
4. Triangular Matrix: Upper triangular (all entries below the main diagonal are 0) or Lower triangular.

---

02.3 Elementary Matrices and Inverses

Definition 02.4 (Inverse Matrix)

For a square matrix \(A\), if there exists a matrix \(B\) such that \(AB = BA = I\), then \(A\) is invertible (or non-singular). \(B\) is called the inverse of \(A\), denoted \(A^{-1}\).

Theorem 02.1 (Properties of Inverses)

1. \((A^{-1})^{-1} = A\)
2. \((AB)^{-1} = B^{-1} A^{-1}\) (Reversal law)
3. \((A^T)^{-1} = (A^{-1})^T\)

Algorithm 02.1 (Gauss-Jordan Inversion)

Construct the block matrix \([A | I]\). Apply elementary row operations to transform the left side into \(I\). The resulting right side is \(A^{-1}\): $\([A | I] \xrightarrow{\text{row operations}} [I | A^{-1}]\)$

---

02.4 Block Matrices

Technique: Block Operations

For large-scale matrices, it is useful to partition them into smaller sub-blocks. If block sizes are compatible, addition and multiplication rules are identical to those for standard matrices. This is vital for distributed computing and sparse matrix handling.

---

Exercises

---

Solution

---

Solution

---

Solution

---

Solution

---

Solution

---

Solution

---

Solution

---

Solution

---

Solution

---

Solution

Chapter Summary

Matrix operations construct the computational syntax of linear algebra:

****: The most fundamental difference between matrix and scalar multiplication, dictating that the order of operators cannot be arbitrarily swapped.

****: Transpose and inverse operations maintain internal logical consistency, providing the basis for solving operator equations.

****: Special matrices (Identity, Diagonal, Block) greatly simplify the analysis of complex systems and are key entry points for numerical algorithm optimization.