"Anatomy" of Dimensions
\( A^{m \times n} \)
\[ \begin{bmatrix} a_{11} & a_{12} & \cdots & a_{1n} \\ a_{21} & a_{22} & \cdots & a_{2n} \\ \vdots & & & \vdots \\ a_{m1} & a_{m2} & \cdots & a_{mn} \end{bmatrix} \]\( m \) rows
\( n \) columns
"Anatomy" of Dimensions
\( A^{m \times n} \longrightarrow \dashrightarrow R \) ((R)REF of \( A \))
Figure 1
"Anatomy" of Dimensions
\( A^{m \times n} \longrightarrow \dashrightarrow R \) ((R)REF of \( A \))
Figure 2
pivot variables
free variables
"Anatomy" of Dimensions
\[ \begin{gathered} \text{Rank}(A) = \#\text{ of pivot variables} \\[0.5em] \text{Nullity}(A) = \#\text{ of free variables} \end{gathered} \]"Anatomy" of Dimensions
- \[ \begin{gathered} \text{Col}(A) = \text{Column space of } A \\[0.5em] = \text{Span}\{\text{columns of } A\} \subseteq \mathbb{R}^m \\[0.5em] \dim(\text{Col}(A)) = \#\text{ of pivots} = \text{Rank} \end{gathered} \]
- \[ \begin{gathered} \text{Null}(A) = \{X : AX = \vec{0}\} \subseteq \mathbb{R}^n \\[0.5em] \dim(\text{Null}(A)) = \#\text{ of free var.} = \text{Nullity} \end{gathered} \]
"Anatomy" of Dimensions
- \[ \begin{gathered} \text{Row}(A) = \text{Row Space of } A \\[0.5em] = \text{Span}\{\text{rows of } A\} \subseteq \mathbb{R}^n \end{gathered} \]
If \( A \xrightarrow{\ \text{elementary row operations}\ } \dashrightarrow B \) then \( \text{Row}(A) = \text{Row}(B) \)
\[ A \longrightarrow \longrightarrow \longrightarrow \dashrightarrow R \quad \text{((R)REF of } A\text{)} \]We have
\[ \text{Row}(A) = \text{Row}(R) \]"Anatomy" of Dimensions
- \[ \begin{gathered} \text{Row}(A) = \text{Row Space of } A \\[0.5em] = \text{Span}\{\text{rows of } A\} \\[0.5em] = \text{Span}\{\text{rows of } R\} \subseteq \mathbb{R}^n \\[0.5em] \dim(\text{Row}(A)) = \#\text{ of non-zero rows of } R \\[0.5em] = \#\text{ of pivots} \\[0.5em] = \text{Rank}(A) \quad (= \dim(\text{Col}(A))) \end{gathered} \]
Relationship to Lin Ind.
Given \( A^{m \times n} \),
-
Columns of \( A \) are lin. ind.
\[ \begin{gathered} \Longleftrightarrow \text{No free variables} \\[0.5em] \Longleftrightarrow \text{Rank}(A) = n \end{gathered} \] -
Rows of \( A \) are lin. ind
\[ \begin{gathered} \Longleftrightarrow \text{No zero rows in } R \\[0.5em] \Longleftrightarrow \text{Rank}(A) = m \end{gathered} \]
Relationship to Solving AX=B
Given \( A^{m \times n} \),
-
\( AX = B \) is solvable for any \( B \)
\[ \begin{gathered} \Longleftrightarrow \text{no zeros row in } R \\[0.5em] \Longleftrightarrow \text{Rank}(A) = m \end{gathered} \] -
\( AX = B \) has unique solution (if exists)
\[ \begin{gathered} \Longleftrightarrow \text{no free variables, i.e. only pivots} \\[0.5em] \Longleftrightarrow \text{Rank}(A) = n \end{gathered} \]
Non-Singular Matrix
\( A^{m \times n} \) is called a non-singular matrix if for any \( B \), \( AX = B \) has a unique solution
\[ \begin{gathered} \Longrightarrow \quad \text{Rank}(A) = m \text{ and } \text{Rank}(A) = n \\[0.5em] \Longrightarrow \quad \underline{m = n}, \text{ i.e. } \boxed{A \text{ is a square matrix}} \end{gathered} \]Non-Singular Matrix
\( A^{m \times n} \) is called a non-singular matrix if for any \( B \), \( AX = B \) has a unique solution
\[ \begin{gathered} \Longrightarrow \quad \text{Rank}(A) = m \text{ and } \text{Rank}(A) = n \\[0.5em] \Longrightarrow \quad \underline{m = n}, \text{ i.e. } \boxed{A \text{ must be a square matrix.}} \end{gathered} \]Non-Singular Matrix (Given \( A^{n \times n} \))
The following statements are equivalent:
- \( A \) is non-singular;
- \( \text{Null}(A) = \{0\} \), ie. no free variables;
- \( AX = B \) has at most one solution;
- \( AX = B \) is solvable for any \( B \);
- \( A \) can be row reduced to \( I \)
Non-Singular Matrix (Given \( A^{n \times n} \))
The following statements are equivalent:
- \( \text{Rank}(A) = n \); \( (\dim(\text{Col}(A)) = n) \)
- Columns of \( A \) are lin ind;
- Columns of \( A \) span \( \mathbb{R}^n \)
- \( \text{Col}(A) = \mathbb{R}^n \)
- Columns of \( A \) form a basis of \( \mathbb{R}^n \)
Non-Singular Matrix (Given \( A^{n \times n} \))
The following statements are equivalent:
- Rows of \( A \) are lin ind;
- Rows of \( A \) span \( \mathbb{R}^n \)
- \( \text{Row}(A) = \mathbb{R}^n \)
- Rows of \( A \) form a basis of \( \mathbb{R}^n \)