Learn Without Walls
← Lesson 2 Lesson 3 of 4 Next Lesson →

Lesson 3: Gauss-Jordan and Reduced Row Echelon Form

Estimated time: 30-40 minutes

Learning Objectives

By the end of this lesson, you will be able to:

From REF to RREF

In Lesson 2, Gaussian elimination produced row echelon form (REF), and we used back-substitution to finish. Gauss-Jordan elimination takes it further: we continue using row operations to also eliminate entries above each pivot, and we scale each pivot to 1. The result is reduced row echelon form (RREF).

Reduced Row Echelon Form (RREF): A matrix is in RREF if it satisfies all REF conditions plus:

  1. Every pivot is equal to 1.
  2. Every pivot is the only nonzero entry in its column (zeros above and below).

Example: REF vs. RREF

REF (not reduced):

[ 1   2   3  |  10 ]
[ 0   1   1  |   4 ]
[ 0   0   2  |   6 ]

Pivots are 1, 1, and 2. The last pivot is not 1, and there are nonzero entries above pivots.

RREF:

[ 1   0   0  |  1 ]
[ 0   1   0  |  1 ]
[ 0   0   1  |  3 ]

Every pivot is 1, and each pivot column has zeros everywhere else. You can read off: x_1 = 1, x_2 = 1, x_3 = 3.

Gauss-Jordan Elimination Algorithm

Steps (continuing from REF)

  1. Scale each pivot to 1: Divide each pivot row by its pivot value.
  2. Eliminate above each pivot: Working from the rightmost pivot to the left, use row replacement to create zeros above each pivot.

Worked Example: Full Gauss-Jordan Elimination

Take the REF from above:

[ 1   2   3  |  10 ]
[ 0   1   1  |   4 ]
[ 0   0   2  |   6 ]

Step 1: Scale row 3. (1/2)R_3 → R_3:

[ 1   2   3  |  10 ]
[ 0   1   1  |   4 ]
[ 0   0   1  |   3 ]

Step 2: Eliminate above pivot in column 3.

R_2 - R_3 → R_2: [0, 1, 1-1 | 4-3] = [0, 1, 0 | 1]

R_1 - 3R_3 → R_1: [1, 2, 3-3 | 10-9] = [1, 2, 0 | 1]

[ 1   2   0  |  1 ]
[ 0   1   0  |  1 ]
[ 0   0   1  |  3 ]

Step 3: Eliminate above pivot in column 2.

R_1 - 2R_2 → R_1: [1, 2-2, 0 | 1-2] = [1, 0, 0 | -1]

[ 1   0   0  |  -1 ]
[ 0   1   0  |   1 ]
[ 0   0   1  |   3 ]

Read the solution: x_1 = -1, x_2 = 1, x_3 = 3.

Uniqueness of RREF

Uniqueness Theorem: Every matrix has exactly one reduced row echelon form. Different sequences of row operations may be used, but they all lead to the same RREF.

This is important: while there are many possible REFs for a matrix (depending on which operations you choose), the RREF is unique. This makes RREF the canonical or standard form for any matrix.

Pivot Columns and Free Variables

In RREF, each pivot occupies a column. We call these pivot columns. Columns without a pivot correspond to free variables -- variables that can take any value. This is the key to understanding infinite solution sets, which we explore fully in Lesson 4.

Example: Identifying Pivot and Free Variables

RREF:

[ 1   0   3  |  5 ]
[ 0   1  -2  |  1 ]
[ 0   0   0  |  0 ]

Pivot columns: Columns 1 and 2 (where the 1s are).

Free variable: x_3 (column 3 has no pivot).

Reading the solution: x_1 + 3x_3 = 5 and x_2 - 2x_3 = 1. Let x_3 = t (any real number). Then x_1 = 5 - 3t, x_2 = 1 + 2t, x_3 = t.

Check Your Understanding

1. Is this matrix in RREF? [ 1 0 2 | 3 ] / [ 0 2 1 | 4 ]

Answer: No. The pivot in row 2 is 2, but in RREF every pivot must equal 1. You would need to scale row 2 by 1/2.

2. Reduce to RREF: [ 1 3 | 7 ] / [ 0 2 | 4 ]

Answer: Scale R_2: (1/2)R_2 gives [0 1 | 2]. Then R_1 - 3R_2 → R_1: [1, 0 | 7-6] = [1, 0 | 1]. RREF: [ 1 0 | 1 ] / [ 0 1 | 2 ]. Solution: x = 1, y = 2.

3. What is the advantage of RREF over REF?

Answer: In RREF, you can read the solutions directly without needing back-substitution. Each pivot variable equals the constant in its row (for unique solutions), or the relationship with free variables is explicit.

4. In the RREF [ 1 0 -1 | 2 ] / [ 0 1 3 | -1 ] / [ 0 0 0 | 0 ], identify the pivot variables and free variables.

Answer: Pivot variables: x_1 (column 1) and x_2 (column 2). Free variable: x_3 (column 3 has no pivot). The solution is x_1 = 2 + x_3, x_2 = -1 - 3x_3, x_3 is free.

Key Takeaways

Next Lesson

Lesson 4 explores solution sets in depth -- unique, infinitely many, and no solution -- and how to express them in parametric vector form.

Start Lesson 4

Module Home

Module 1 Home