Gaussian Elimination
# Statement
Let $F$ be a Field and let $m,n \in \mathbb{N}$. There exists an Algorithm that maps any $A \in F^{m \times n}$ to a Row Reduced Echelon Matrix only through the use of Elementary Row Operations. This algorithm is known as Gaussian Elimination.
# Proof
Consider the following Algorithm
Name: RREF Input: $F$ Field, $m, n \in \mathbb{N}$, $A \in F^{m \times n}$ Output: $R \in F^{m \times n}$ so that $A \sim_{R} R$ and $R$ is a Row Reduced Echelon Matrix.
- Copy $A$ to $R$
- For $i=1$ to $m$:
\\ create a row-reduced matrix
2. if $R_{i \cdot } = \mathbf{0}$: 2. Continue; 3. $j = \min\limits {r \in [n] : R_{ir} \neq 0}$; 4. $R_{i \cdot} \leftarrow R_{ij}^{-1} R_{i \cdot}$; 7. For $i’=1$ to $m$:
- $R_{i’ \cdot} \leftarrow R_{i’ \cdot} + (-R_{rj} R_{i \cdot})$;
- $S \leftarrow {i \in [m] : R_{i \cdot} \neq \mathbf{0}}$;
- For $i \in S$: 1. $k_{i} \leftarrow \min\limits {r \in [n] : R_{ir} \neq 0}$;
- $r = 1$;
- While $S \neq \emptyset$:
\\ Make it row-reduced echelon
1. $i \leftarrow \text{argmin}{k_{s} : s \in S }$; 2. Swap $R_{r \cdot}$ with $R_{i \cdot}$; 3. $S \leftarrow S \setminus {i}$; 4. $r \leftarrow r + 1$;- Return $R$;
To show this algorithm is correct we need to show
- it terminates,
- It only transforms $R$ by Elementary Row Operations
- It returns a Row Reduced Echelon Matrix.
To that end: