Proof: Row Operations Preserve Solutions
We prove that elementary row operations preserve the solution set of a linear system. This fundamental result justifies the entire methodology of Gaussian elimination.
Theorem: If an elementary row operation is applied to the augmented matrix to obtain , then the systems and have the same solution set.
Proof: We consider each type of elementary row operation separately. Let the original system have equations and let be any solution.
Case 1: Row Switching ()
Interchanging rows and simply reorders the equations. If satisfies all equations in the original order, it clearly satisfies them in the new order. Conversely, any solution to the reordered system satisfies the original system. Thus the solution sets are identical.
Case 2: Row Multiplication ( where )
Suppose equation is . After the operation, equation becomes:
If satisfies the original equation, then:
Multiplying both sides by yields:
so satisfies the new equation. Conversely, if satisfies the new equation, dividing by (which is nonzero) shows it satisfies the original equation. All other equations remain unchanged.
Case 3: Row Addition ( where )
Let equation be and equation be .
The new equation is:
If satisfies both and , then:
Adding times the second equation to the first:
Thus satisfies . Since equation is unchanged, satisfies the new system.
Conversely, suppose satisfies the new system with and . Then satisfies:
Subtracting times the second from the first:
so satisfies the original equation and hence the original system.
Since each row operation is reversible by an operation of the same type, and each preserves solutions, the solution sets must be identical. ∎
This proof demonstrates that row operations establish a two-way correspondence between solutions. The reversibility of each operation is crucial: it ensures that not only do solutions of the original system solve the modified system, but also that solutions of the modified system solve the original system.