This section applies the theory in preceding sections to obtain important theoretical formulas and a geometric interpretation of the determinant.

(I) Cramer's Rule

Let \( A_{n \times n} = \begin{bmatrix} \vec{a}_1 & \dots & \vec{a}_n \end{bmatrix} \). For any \( \vec{b} \in \mathbb{R}^n \), consider the system \( A\vec{x} = \vec{b} \) where \( \vec{x} = \begin{bmatrix} x_1 \\ \vdots \\ x_n \end{bmatrix} \).

Define \( A_i(\vec{b}) \) as the matrix obtained from \( A \) by replacing Column \( i \) by the vector \( \vec{b} \):

\[ A_i(\vec{b}) = \begin{bmatrix} \vec{a}_1 & \dots & \vec{b} & \dots & \vec{a}_n \end{bmatrix} \]

Theorem 7: Cramer's Rule

Let \(A\) be an invertible \(n \times n\) matrix. For any \( \vec{b} \) in \( \mathbb{R}^n \), the unique solution \(\vec{x}\) of \( A\vec{x} = \vec{b} \) has entries given by:

\[ x_i = \frac{\det(A_i(\vec{b}))}{\det A} \]

Note: Cramer's Rule is needed in a variety of theoretical calculations. For example, it can be used to study how the solution of \( A\vec{x} = \vec{b} \) is affected by changes in the entries of \( \vec{b} \). We can also use Cramer's Rule to derive a formula for \(A^{-1}\). However, the formula is generally inefficient for hand calculations on large systems.

Example 1

Determine the values of the parameter \( s \) for which the system has a unique solution, and describe the solution.

\[ \begin{cases} 3sx_1 + 5x_2 = 3 \\ 12x_1 + 5sx_2 = 2 \end{cases} \]

Solution:

By Theorem 5 in Section 2.2, \( A\vec{x} = \vec{b} \) has a unique solution if \(A\) is invertible, which means \(\det A \neq 0\).

\[ A = \begin{bmatrix} 3s & 5 \\ 12 & 5s \end{bmatrix}, \quad \vec{b} = \begin{bmatrix} 3 \\ 2 \end{bmatrix} \]

Find the determinant of \(A\):

\[ \det A = \begin{vmatrix} 3s & 5 \\ 12 & 5s \end{vmatrix} = 15s^2 - 60 = 15(s^2 - 4) = 15(s + 2)(s - 2) \]

The system has a unique solution when \( \det A \neq 0 \), meaning \( s \neq \pm 2 \).

Now apply Cramer's Rule to describe the solution:

\[ \det A_1(\vec{b}) = \begin{vmatrix} 3 & 5 \\ 2 & 5s \end{vmatrix} = 15s - 10 = 5(3s - 2) \] \[ \det A_2(\vec{b}) = \begin{vmatrix} 3s & 3 \\ 12 & 2 \end{vmatrix} = 6s - 36 = 6(s - 6) \]

Thus, the entries for the solution are:

\[ x_1 = \frac{\det A_1(\vec{b})}{\det A} = \frac{15s - 10}{15(s+2)(s-2)} = \frac{3s - 2}{3(s+2)(s-2)} \] \[ x_2 = \frac{\det A_2(\vec{b})}{\det A} = \frac{6s - 36}{15(s+2)(s-2)} = \frac{2s - 12}{5(s+2)(s-2)} \]

(II) A Formula for \(A^{-1}\)

Let \( A_{n \times n} = \begin{bmatrix} \vec{a}_1 & \vec{a}_2 & \dots & \vec{a}_n \end{bmatrix} \).

Let its inverse be denoted by columns: \( A^{-1} = \begin{bmatrix} \vec{v}_1 & \vec{v}_2 & \dots & \vec{v}_n \end{bmatrix} \).

Since \( AA^{-1} = I_n \), we have:

\[ A \begin{bmatrix} \vec{v}_1 & \dots & \vec{v}_n \end{bmatrix} = \begin{bmatrix} A\vec{v}_1 & \dots & A\vec{v}_n \end{bmatrix} = \begin{bmatrix} \vec{e}_1 & \dots & \vec{e}_n \end{bmatrix} \]

This implies that \( A\vec{v}_j = \vec{e}_j \). The \(j\)-th column of \(A^{-1}\) is a vector that satisfies \( A\vec{x} = \vec{e}_j \).

By Cramer's Rule, the \((i, j)\)-entry of \(A^{-1}\) is \(x_i\), given by:

\[ x_i = \frac{\det A_i(\vec{e}_j)}{\det A} \]

Notice that \( A_i(\vec{e}_j) = \begin{bmatrix} \vec{a}_1 & \dots & \vec{e}_j & \dots & \vec{a}_n \end{bmatrix} \), where \(\vec{e}_j\) is in the \(i\)-th column. Expanding the determinant of \( A_i(\vec{e}_j) \) down the \(i\)-th column (where \(\vec{e}_j\) has a 1 in the \(j\)-th row and 0s elsewhere) yields the \((j, i)\)-cofactor of \(A\):

\[ \det A_i(\vec{e}_j) = (-1)^{j+i} \det A_{ji} = C_{ji} \]

Therefore, substituting this back gives:

\[ A^{-1} = \frac{1}{\det A} \begin{bmatrix} C_{11} & C_{21} & \dots & C_{n1} \\ C_{12} & C_{22} & \dots & C_{n2} \\ \vdots & \vdots & \ddots & \vdots \\ C_{1n} & C_{2n} & \dots & C_{nn} \end{bmatrix} \]

Notice that the cofactors are transposed! Normally we do rows-cols, but here the indices are swapped. This matrix of cofactors is called the adjugate (or adjoint) of \(A\), denoted as \( \text{adj } A \).

Theorem 8 (An Inverse Formula)

Let \(A\) be an invertible \(n \times n\) matrix. Then:

\[ A^{-1} = \frac{1}{\det A} \text{adj } A \]

Fact: \( (\text{adj } A)A = (\det A)I_n \)

Example 2

Find the inverse of the matrix \( A = \begin{bmatrix} 2 & 1 & 3 \\ 1 & -1 & 1 \\ 1 & 4 & -2 \end{bmatrix} \).

Solution: First, compute the nine cofactors:

\[ C_{11} = -2, \quad C_{12} = 3, \quad C_{13} = 5 \] \[ C_{21} = 14, \quad C_{22} = -7, \quad C_{23} = -7 \] \[ C_{31} = 4, \quad C_{32} = 1, \quad C_{33} = -3 \]

Construct the adjugate matrix by transposing the matrix of cofactors:

\[ \text{adj } A = \begin{bmatrix} -2 & 14 & 4 \\ 3 & -7 & 1 \\ 5 & -7 & -3 \end{bmatrix} \]

We can compute \(\det A\) directly, or use the fact that \( (\text{adj } A)A = (\det A)I_n \):

\[ (\text{adj } A)A = \begin{bmatrix} -2 & 14 & 4 \\ 3 & -7 & 1 \\ 5 & -7 & -3 \end{bmatrix} \begin{bmatrix} 2 & 1 & 3 \\ 1 & -1 & 1 \\ 1 & 4 & -2 \end{bmatrix} = \begin{bmatrix} 14 & 0 & 0 \\ 0 & 14 & 0 \\ 0 & 0 & 14 \end{bmatrix} = 14 I_3 \]

So, \(\det A = 14\). Finally, multiply by \( \frac{1}{\det A} \):

\[ A^{-1} = \frac{1}{\det A} \text{adj } A = \frac{1}{14} \begin{bmatrix} -2 & 14 & 4 \\ 3 & -7 & 1 \\ 5 & -7 & -3 \end{bmatrix} = \begin{bmatrix} -1/7 & 1 & 2/7 \\ 3/14 & -1/2 & 1/14 \\ 5/14 & -1/2 & -3/14 \end{bmatrix} \]

(Reminder: The algorithm introduced in Section 2.2 often gives a better/more efficient way to compute \(A^{-1}\)).

(III) Determinants as Area or Volume

We will revisit the geometric interpretation of determinants. (Recall Example 3 in Section 3.1).

Theorem 9

If \(A\) is a \(2 \times 2\) matrix, the area of the parallelogram determined by the columns of \(A\) is \(|\det A|\).

If \(A\) is a \(3 \times 3\) matrix, the volume of the parallelepiped determined by the columns of \(A\) is \(|\det A|\).

Main idea of the proof for the \(2 \times 2\) case:

This is visually obvious for a diagonal matrix \( A = \begin{bmatrix} a & 0 \\ 0 & d \end{bmatrix} \) where the area is simply base \(\times\) height (\(|a| \cdot |d|\)). For a general matrix \( A = \begin{bmatrix} \vec{a}_1 & \vec{a}_2 \end{bmatrix} \), it can be transformed into a diagonal matrix using row operations in a way that changes neither the area of the associated parallelogram nor \(|\det A|\).

Example 3

Calculate the area of the parallelogram determined by the points.

Two side-by-side 2D Cartesian coordinate plots illustrating the translation of a parallelogram. The left plot shows the original red parallelogram with vertices at (-1, 4), (2, 5), (3, 9), and (0, 8). The right plot shows the translated blue parallelogram where the bottom-left vertex has been shifted to the origin, resulting in new vertices at (0, 0), (3, 1), (4, 5), and (1, 4). Blue vector arrows originate from the origin to the points (3, 1) and (1, 4), representing the columns of the matrix used to calculate the determinant and area.
Figure: Translating a Parallelogram to the Origin to Compute Area

Solution: First, translate the parallelogram to one having the origin as a vertex. Subtract the base vertex from all other vertices. This yields standard vectors radiating from the origin.

Let the translated vectors forming the sides of the parallelogram be \( \begin{bmatrix} 2 \\ -1 \end{bmatrix} \) and \( \begin{bmatrix} 4 \\ 4 \end{bmatrix} \).

Construct the matrix \(A\):

\[ A = \begin{bmatrix} 2 & 4 \\ -1 & 4 \end{bmatrix} \]

Find the determinant:

\[ |\det A| = |(2)(4) - (-1)(4)| = |8 + 4| = 12 \]

The area of the parallelogram is \( 12 \).

(IV) Linear Transformations

Theorem 10

Let \(T: \mathbb{R}^2 \rightarrow \mathbb{R}^2\) be the linear transformation determined by a \(2 \times 2\) matrix \(A\). If \(S\) is a parallelogram in \(\mathbb{R}^2\), then:

\[ \{ \text{Area of } T(S) \} = |\det A| \cdot \{ \text{Area of } S \} \]

If \(T: \mathbb{R}^3 \rightarrow \mathbb{R}^3\) is determined by a \(3 \times 3\) matrix \(A\), and \(S\) is a parallelepiped in \(\mathbb{R}^3\), then:

\[ \{ \text{Volume of } T(S) \} = |\det A| \cdot \{ \text{Volume of } S \} \]

It holds whenever \(S\) is a region with a finite area / finite volume.

Example 4

Let \(S\) be the parallelogram determined by \( \vec{b}_1 = \begin{bmatrix} -2 \\ 3 \end{bmatrix} \) and \( \vec{b}_2 = \begin{bmatrix} -2 \\ 5 \end{bmatrix} \), and let \( A = \begin{bmatrix} 6 & -3 \\ -3 & 2 \end{bmatrix} \).

Compute the area of the image of \(S\) under the mapping \( \vec{x} \mapsto A\vec{x} \).

Solution:

First, find the area of the original parallelogram \(S\):

\[ \text{Area of } S = \left| \det \begin{bmatrix} -2 & -2 \\ 3 & 5 \end{bmatrix} \right| = |(-2)(5) - (-2)(3)| = |-10 + 6| = |-4| = 4 \]

Next, find the determinant of the transformation matrix \(A\):

\[ \det A = (6)(2) - (-3)(-3) = 12 - 9 = 3 \]

Apply the theorem to find the area of the transformed region \(T(S)\):

\[ \text{Area of } T(S) = |\det A| \cdot (\text{Area of } S) = 3 \cdot 4 = 12 \]