(Official) Definition of det

\[ A^{n\times n} \longrightarrow \det A \in \mathbb{R} \]

\( \det(\cdot) \) is a multi-linear, alternating function on square matrices such that \( \det I_n = 1 \)

(1) Multi-linear (linear on each row)

\[ \det\begin{pmatrix} -\;r_1\;- \\ \vdots \\ -\;\alpha u + \beta v\;- \\ \vdots \\ -\;r_n\;- \end{pmatrix} = \alpha \det\begin{pmatrix} -\;r_1\;- \\ \vdots \\ -\;u\;- \\ \vdots \\ -\;r_n\;- \end{pmatrix} + \beta \det\begin{pmatrix} -\;r_1\;- \\ \vdots \\ -\;v\;- \\ \vdots \\ -\;r_n\;- \end{pmatrix} \]

(2) Alternating

\[ \det\begin{pmatrix} -\;r_i\;- \\ -\;r_j\;- \end{pmatrix} = -\det\begin{pmatrix} -\;r_j\;- \\ -\;r_i\;- \end{pmatrix} \]

(1)1

\[ \det\begin{pmatrix} -\;r_1\;- \\ \vdots \\ -\;u+v\;- \\ \vdots \\ -\;r_n\;- \end{pmatrix} = \det\begin{pmatrix} -\;r_1\;- \\ \vdots \\ -\;u\;- \\ \vdots \\ -\;r_n\;- \end{pmatrix} + \det\begin{pmatrix} -\;r_1\;- \\ \vdots \\ -\;v\;- \\ \vdots \\ -\;r_n\;- \end{pmatrix} \]

(1)2

\[ \det\begin{pmatrix} -\;r_1\;- \\ \vdots \\ -\;cu\;- \\ \vdots \\ -\;r_n\;- \end{pmatrix} = c \det\begin{pmatrix} -\;r_1\;- \\ \vdots \\ -\;u\;- \\ \vdots \\ -\;r_n\;- \end{pmatrix} \]

Tricks of Computing det / Properties of det

  1. \( \det(\text{upper/lower triangular matrix}) = \) product of the diagonal entries
\[ \begin{gathered} \begin{vmatrix} a_{11} & & & \\ & a_{22} & & \\ & & \ddots & \\ \text{0's} & & & a_{nn} \end{vmatrix} = a_{11}\,a_{22}\cdots a_{nn} \\[0.5em] \begin{vmatrix} a_{11} & & & \\ & a_{22} & & \text{0's} \\ & & \ddots & \\ & & & a_{nn} \end{vmatrix} = a_{11}\,a_{22}\cdots a_{nn} \\[0.5em] \begin{vmatrix} a & b & c \\ 0 & d & e \\ 0 & 0 & f \end{vmatrix} = adf = \begin{vmatrix} a & 0 & 0 \\ b & c & 0 \\ d & e & f \end{vmatrix} \end{gathered} \]

\( \det(\text{upper/lower triangular block matrix}) = \) product of dets of diagonal blocks.

\[ \begin{gathered} \begin{vmatrix} A_1 & & & \\ & A_2 & & * \\ & & \ddots & \\ \text{0's} & & & A_k \end{vmatrix} = (\det A_1)(\det A_2)\cdots(\det A_k) \\[0.5em] \begin{vmatrix} A_1 & & & \\ & A_2 & & \text{0's} \\ * & & \ddots & \\ & & & A_k \end{vmatrix} = (\det A_1)(\det A_2)\cdots(\det A_k) \end{gathered} \]

eg

\[ \begin{gathered} \begin{vmatrix} 1 & 0 & 0 & 0 \\ * & 2 & 0 & 0 \\ * & * & 3 & 0 \\ * & * & * & 4 \end{vmatrix} = \begin{vmatrix} 1 & * & * & * \\ 0 & 2 & * & * \\ 0 & 0 & 3 & * \\ 0 & 0 & 0 & 4 \end{vmatrix} \\[0.5em] = (1)(2)(3)(4) = 24 \end{gathered} \]

eg

\[ \begin{gathered} \begin{vmatrix} 1 & 2 & * & * & * \\ 3 & 4 & * & * & * \\ 0 & 0 & 5 & * & * \\ 0 & 0 & 0 & 6 & 7 \\ 0 & 0 & 0 & 8 & 9 \end{vmatrix} = \begin{vmatrix} 1 & 2 & 0 & 0 & 0 \\ 3 & 4 & 0 & 0 & 0 \\ * & * & 5 & 0 & 0 \\ * & * & * & 6 & 7 \\ * & * & * & 8 & 9 \end{vmatrix} \\[0.5em] = \begin{vmatrix} 1 & 2 \\ 3 & 4 \end{vmatrix} (5) \begin{vmatrix} 6 & 7 \\ 8 & 9 \end{vmatrix} \\[0.5em] = (4-6)(5)(54-56) \\[0.5em] = (-2)(5)(-2) \\[0.5em] = 20 \end{gathered} \]
  1. If any row (or column) is zero, then \( \det = 0 \).
  2. If 2 rows (or 2 columns) are the same, then \( \det = 0 \)
  3. If the rows (or columns) are lin dep, then \( \det = 0 \)
  4. If the rows (or columns) are lin ind, then \( \det \ne 0 \)

(4) + (5)

Note

\( \det \ne 0 \iff \) rows/cols are lin ind.

\( \det = 0 \iff \) rows/cols are lin dep.

  1. \( A^{-1} \) exists \( \iff \det A \ne 0 \)

    ( There is a formula of \( A^{-1} \) using \( \det A \))

  2. \[ \det (AB) = (\det A)(\det B) \]

    \( \Longrightarrow \det (A^{-1}) = \dfrac{1}{\det A} \)

  3. \[ \det (A^{T}) = \det A \]

When is \( A^{n \times n} \) invertible / non-singular?

(The following statements are equivalent.)

  1. \( A \) is onto (Rank \( = n \))

    ( \( \iff \) Nullity \( = 0 \))

    • \( AX = b \) is solvable for any \( b \)
  2. \( A \) is one-to-one ( Nullity \( = 0 \))

    ( \( \iff \) Rank \( = 0 \) )

    • \( AX = AY \Longrightarrow X = Y \)
    • \( AX = 0 \Longrightarrow X = 0 \)
    • \( \mathrm{Null}(A) = \{ \vec{0} \} \)
  3. Columns of \( A \) are lin ind
  4. Rows of \( A \) are lin ind.
  5. \( \det A \ne 0 \)

Applications of det

  1. Area of parallelogram

    Shaded parallelogram spanned by vectors u equals (a, b) and v equals (c, d).
    Figure 1
    A blue shaded parallelogram drawn with two arrows emanating from its bottom-left corner. One arrow points up and to the right along the left edge and is labelled \( \vec{v} = \begin{pmatrix} c \\ d \end{pmatrix} \). The other arrow points to the right along the bottom edge and is labelled \( \vec{u} = \begin{pmatrix} a \\ b \end{pmatrix} \). The remaining two sides are drawn as dotted lines, completing the parallelogram and showing the region whose area is being measured.
    \[ \begin{gathered} \text{Area}(\square) = \left| \det \begin{pmatrix} \vec{u} & \vec{v} \end{pmatrix} \right| \\[0.5em] = \left| \det \begin{pmatrix} a & c \\ b & d \end{pmatrix} \right| = \left| ad - bc \right| \end{gathered} \]
  2. Volumn of parallelepiped

    Sketch of a parallelepiped spanned by vectors u, v, and w with dashed hidden edges.
    Figure 2
    A blue freehand sketch of a three-dimensional parallelepiped (a slanted box) drawn in perspective. Three arrows start from the near bottom-left vertex: one pointing up and to the right labelled v, one pointing to the right along the bottom labelled u, and one dashed arrow going back into the page labelled w. Hidden edges are drawn with dashed lines. A curved black arrow points from the interior of the solid to the word Volume written to the right of the figure.
    \[ \text{Volume} = \det \begin{pmatrix} | & | & | \\ \vec{u} & \vec{v} & \vec{w} \\ | & | & | \end{pmatrix} \]
  1. Cross-product

    \[ \begin{gathered} u = \begin{pmatrix} a \\ b \\ c \end{pmatrix}, \quad v = \begin{pmatrix} d \\ e \\ f \end{pmatrix} \\[0.5em] u \times v = \det \begin{pmatrix} i & j & k \\ a & b & c \\ d & e & f \end{pmatrix} =(bf-ec)i - (ac-dc)j + (ae-bd)k\end{gathered} \]
  1. Cramer's Rule for solving square linear system

    \[ \begin{gathered} AX = b \quad ( A^{n \times n}, \ \det A \ne 0 \iff A^{-1} \text{ exists} ) \\[0.5em] A = \begin{pmatrix} u_1 & u_2 & \cdots & u_n \end{pmatrix}, \quad X = \begin{pmatrix} x_1 \\ x_2 \\ \vdots \\ x_n \end{pmatrix} \\[0.5em] x_1 = \frac{\det \begin{pmatrix} b & u_2 & u_3 & \cdots & u_n \end{pmatrix}}{\det A} \\[0.5em] x_2 = \frac{\det \begin{pmatrix} u_1 & b & u_3 & \cdots & u_n \end{pmatrix}}{\det A} \\[0.5em] \vdots \\[0.5em] x_n = \frac{\det \begin{pmatrix} u_1 & u_2 & \cdots & u_{n-1} & b \end{pmatrix}}{\det A} \end{gathered} \]

eg \( 2 \times 2 \)

\[ \begin{gathered} \begin{cases} ax + by = p \\ cx + dy = q \end{cases} \iff \begin{pmatrix} a & b \\ c & d \end{pmatrix} \begin{pmatrix} x \\ y \end{pmatrix} = \begin{pmatrix} p \\ q \end{pmatrix} \\[0.5em] x = \frac{\det \begin{pmatrix} p & b \\ q & d \end{pmatrix}}{\det \begin{pmatrix} a & b \\ c & d \end{pmatrix}} = \frac{pd - bq}{ad - bc} \\[0.5em] y = \frac{\det \begin{pmatrix} a & p \\ c & q \end{pmatrix}}{\det \begin{pmatrix} a & b \\ c & d \end{pmatrix}} = \frac{qa - pc}{ad - bc} \end{gathered} \]

Note :

The above solution formula only works for invertible square matrices

(5) Formula for \( A^{-1} \) ( \( \det A \ne 0 \) )

\[ \begin{gathered} A^{-1} = \frac{\left[ (-1)^{ij} \det M_{ij} \right]^{T}}{\det A} \\[0.5em] (A^{-1})_{ij} = \frac{(-1)^{ij} \det M_{ji}}{\det A} \\[0.5em] \begin{bmatrix} a & b \\ c & d \end{bmatrix}^{-1} = \frac{\begin{bmatrix} +d & -c \\ -b & +a \end{bmatrix}^{T}}{\det \begin{pmatrix} a & b \\ c & d \end{pmatrix}} = \frac{\begin{bmatrix} d & -b \\ -c & a \end{bmatrix}}{ad - bc} \\[0.5em] \begin{bmatrix} a & b & c \\ d & e & f \\ g & h & i \end{bmatrix}^{-1} = \frac{\begin{bmatrix} +\begin{vmatrix} e & f \\ h & i \end{vmatrix} & -\begin{vmatrix} d & f \\ g & i \end{vmatrix} & +\begin{vmatrix} d & e \\ g & h \end{vmatrix} \\[6pt] -\begin{vmatrix} b & c \\ h & i \end{vmatrix} & +\begin{vmatrix} a & c \\ g & i \end{vmatrix} & -\begin{vmatrix} a & b \\ g & h \end{vmatrix} \\[6pt] +\begin{vmatrix} b & c \\ e & f \end{vmatrix} & -\begin{vmatrix} a & c \\ d & f \end{vmatrix} & +\begin{vmatrix} a & b \\ d & e \end{vmatrix} \end{bmatrix}^{T}}{\det A} = \frac{\begin{bmatrix} +\begin{vmatrix} e & f \\ h & i \end{vmatrix} & -\begin{vmatrix} b & c \\ h & i \end{vmatrix} & +\begin{vmatrix} b & c \\ e & f \end{vmatrix} \\[6pt] -\begin{vmatrix} d & f \\ g & i \end{vmatrix} & +\begin{vmatrix} a & c \\ g & i \end{vmatrix} & -\begin{vmatrix} a & c \\ d & f \end{vmatrix} \\[6pt] +\begin{vmatrix} d & e \\ g & h \end{vmatrix} & -\begin{vmatrix} a & b \\ g & h \end{vmatrix} & +\begin{vmatrix} a & b \\ d & e \end{vmatrix} \end{bmatrix}}{\det A} \end{gathered} \]