Back Substitution
What You'll Learn
You'll learn how to systematically solve for all variables in a linear system once it's been transformed into upper triangular or row echelon form.
The Core Idea
After using Gaussian elimination to transform your augmented matrix into row echelon form, you have a staircase-like structure with pivot positions (leading entries) that make the system much easier to solve. Back substitution is the final step: you start from the bottom row and work your way up, solving for one variable at a time.
Think of it like climbing down a ladder. The bottom rung is simplest—often just one variable equals a number. Once you know that value, you can substitute it into the equation above, which then becomes simple to solve. You keep climbing up, substituting known values as you go.
The Process
- Start at the bottom equation: This usually has just one variable, so solve it directly.
- Move up one row: Substitute the value(s) you just found into this equation, then solve for the next variable.
- Repeat upward: Continue substituting and solving until you reach the top row.
Example
Suppose forward elimination gave you:
x + 2y - z = 3
y + z = 5
z = 2
Back substitution:
- Bottom: z = 2 ✓
- Middle: y + 2 = 5 → y = 3 ✓
- Top: x + 2(3) - 2 = 3 → x + 4 = 3 → x = -1 ✓
Solution: x = -1, y = 3, z = 2
Key Takeaway: Back substitution solves a system by working from the simplest (bottom) equation upward, substituting known values into each equation above to systematically find all variables.