2.4 Matrices, Systems of Equations & Linear Algebra

2.4 Matrices, Systems of Equations & Linear Algebra

Linear algebra provides the structural framework connecting systems of equations, matrix transformations, and multi-dimensional geometric spaces. In secondary mathematics, students transition from solving 2 x 2 systems by substitution and elimination to matrix methods that scale to arbitrary dimensions. For the NBPTS AYA candidate, deep content knowledge encompasses both computational proficiency (Gaussian elimination, matrix inversion, determinant evaluations) and conceptual fluency (interpreting matrices as geometric mappings, understanding the algebraic consequences of non-commutativity, and classifying solution spaces).

---

1. Classification and Geometry of Linear Systems

A system of m linear equations in n variables represents the intersection of geometric hyperplanes in ℝⁿ.

Systems in Two Variables (2 x 2)

In ℝ², each linear equation ax + by = c represents a straight line. Three distinct geometric possibilities exist:

1. Consistent and Independent: The lines have different slopes (m_1 ≠ m_2) and intersect at a unique point (x_0, y_0).
2. Consistent and Dependent: The lines have identical slopes and identical y-intercepts (m_1 = m_2, b_1 = b_2). The lines are coincident, yielding infinitely many solutions along the line.
3. Inconsistent: The lines have identical slopes but different y-intercepts (m_1 = m_2, b_1 ≠ b_2). The lines are parallel and distinct, resulting in no solution (empty intersection ∅).

Systems in Three Variables (3 x 3)

In ℝ³, each equation ax + by + cz = d represents a flat two-dimensional plane:

• Unique Solution: The three normal vectors are linearly independent; the planes intersect at a single point.
• Infinitely Many Solutions (Dimension 1): The planes intersect along a single common line (rank of coefficient matrix is 2).
• Infinitely Many Solutions (Dimension 2): All three equations represent the identical plane (rank is 1).
• No Solution (Inconsistent): The planes fail to share a common intersection. This occurs when two or more planes are parallel, or when the three planes intersect in pairs along three parallel lines, forming a triangular prism.

---

2. Augmented Matrices & Gaussian Elimination

A linear system can be encoded compactly as an augmented matrix [A | B]. For a system of m equations in n variables, the coefficient matrix A is augmented with the constant column vector B.

Elementary Row Operations (EROs)

Gaussian elimination applies three row operations that transform the augmented matrix into an upper triangular Row Echelon Form (REF) or Reduced Row Echelon Form (RREF) without changing the solution set:

1. Row Swap (R_i <-> R_j): Interchange the positions of two rows.
2. Row Scaling (c · R_i -> R_i, c ≠ 0): Multiply all entries in a row by a non-zero scalar.
3. Row Addition / Replacement (R_i + c · R_j -> R_i): Add a scalar multiple of row j to row i.

Diagnosing Solution Types from REF / RREF

• Inconsistency Signal: Any row of the form [0 0 ... 0 | c] where c ≠ 0 translates to 0x_1 + 0x_2 + ... + 0x_n = c => 0 = c, an immediate contradiction confirming no solution.
• Unique Solution: Every column corresponding to a variable contains a leading pivot, with no contradictory rows.
• Infinite Solutions: If the system is consistent and the number of leading pivots r = rank(A) is strictly less than the number of variables n, there are n - r free variables that parameterize the solution space.

---

3. Matrix Operations, Dimension Rules & Non-Commutativity

Let A and B be matrices over ℝ:

• Addition & Subtraction: Defined if and only if A and B share identical dimensions (m x n). Operations occur component-wise.
• Scalar Multiplication: For k in ℝ, (kA){ij} = k(A{ij}).
• Matrix Multiplication: The product AB is defined if and only if the number of columns in A equals the number of rows in B. If A is m x k and B is k x n, then C = AB is m x n, where each entry is the dot product of row i of A and column j of B.

The Failure of Commutativity (AB ≠ BA)

Matrix multiplication is associative (A(BC) = (AB)C) and distributive over addition (A(B + C) = AB + AC), but generally non-commutative:

$$AB \neq BA$$

Even when both A and B are square n x n matrices so that both AB and BA are defined, AB rarely equals BA.

Algebraic Expansion Warning: Because matrix multiplication does not commute, standard polynomial identities cannot be applied without preserving order: (A + B)² = (A + B)(A + B) = A² + AB + BA + B² ≠ A² + 2AB + B² (A - B)(A + B) = A² + AB - BA - B² ≠ A² - B² Equality holds if and only if A and B commute (AB = BA).

---

4. Determinants, Invertibility & Algebraic Properties

For a square n x n matrix A, the determinant det(A) (or |A|) is an invariant scalar value that characterizes the matrix's algebraic and geometric properties.

2 x 2 and 3 x 3 Determinant Formulas

• 2 x 2 Matrix: det([[a, b], [c, d]]) = ad - bc
• 3 x 3 Matrix (Expansion by Minors along Row 1): det([[a1, b1, c1], [a2, b2, c2], [a3, b3, c3]]) = a1 · det([[b2, c2], [b3, c3]]) - b1 · det([[a2, c2], [a3, c3]]) + c1 · det([[a2, b2], [a3, b3]])

Fundamental Properties of the Determinant

1. Multiplicative Property: For any two n x n matrices, det(AB) = det(A) · det(B).
2. Transpose Invariance: det(Aᵀ) = det(A).
3. Inverses: If A is invertible, det(A⁻¹) = 1 / det(A).
4. Scalar Scaling: For an n x n matrix A and scalar k in ℝ: det(kA) = kⁿ det(A) (Because multiplying matrix A by scalar k scales every one of its n individual rows by k).
5. Invertibility Criterion: A square matrix A is invertible (A⁻¹ exists) if and only if det(A) ≠ 0.

---

5. Matrix Inverses & Solving Matrix Equations

For an invertible 2 x 2 matrix A = [[a, b], [c, d]] with det(A) = ad - bc ≠ 0, the unique multiplicative inverse A⁻¹ is:

$$A^{-1} = \frac{1}{ad - bc} \begin{bmatrix} d & -b \\ -c & a \end{bmatrix}$$

Solving the System Matrix Equation AX = B

Given the linear system encoded as AX = B:

1. Verify det(A) ≠ 0.
2. Multiply both sides on the left by A⁻¹: A⁻¹ (AX) = A⁻¹ B => (A⁻¹ A) X = A⁻¹ B => I X = A⁻¹ B => X = A⁻¹ B

Instructional Caution: Because matrix multiplication does not commute, writing X = B A⁻¹ or X = B / A is completely invalid. Matrix division does not exist, and right-multiplying by A⁻¹ yields AX A⁻¹, which does not simplify to X.

---

6. Geometric Interpretation: Linear Transformations & Area Scaling

A 2 x 2 matrix T = [[a, c], [b, d]] defines a linear transformation T: ℝ² -> ℝ² sending vectors x -> Tx.

The columns of T represent the landing positions (images) of the standard basis vectors e_1 = [1, 0]ᵀ and e_2 = [0, 1]ᵀ:

T(e_1) = [a, b]ᵀ, T(e_2) = [c, d]ᵀ

• The unit square spanned by e_1 and e_2 (having area 1) is mapped to the parallelogram spanned by [a, b]ᵀ and [c, d]ᵀ.
• The area of this transformed parallelogram is precisely |det(T)| = |ad - bc|.
• General Area Distortion: For any 2D region S with area A_0, the area of the transformed image T(S) is: Area(T(S)) = |det(T)| · Area(S)
• Geometric Orientation:
  • If det(T) > 0, the transformation preserves orientation (counterclockwise loops remain counterclockwise).
  • If det(T) < 0, the transformation reverses orientation (involves an odd number of reflections).
  • If det(T) = 0, the transformation collapses the 2D plane onto a 1D line or 0D point, resulting in an area of zero and confirming non-invertibility.