
Solutions of a Linear System
Matrix Form of a Linear System
Given a linear system of equations, we have seen how to write the augmented matrix [ A ∣ b ].
We may also write the SLE as a matrix equation, using the same coefficient matrix A and constant vector b:
Ax=b
Here, x=x1x2⋮xn is the unknown solution vector.
Procedure to Solve Ax=b
- Write the linear system as an augmented matrix
- Turn the matrix into RREF using EROs
- Find all solutions
- Use rank to determine the number of solutions
- Turn the augmented matrix back into a system of linear equations (assign parameters if needed)
- Solve for any missing variables
Example 1
Find the general solution to the system represented by the augmented matrix:
100010001abc
⟹⎩⎨⎧x1=ax2=bx3=c⟹x=abc
Example 2
Find the general solution to the system represented by the augmented matrix:
[100123ab]
n−rank(A) = 3−2 = 1 free variable.
Note thatx1 and x2 have leading 1s in their columns, but x3 does not.
⟹x3 is a free variable, and we assign it a parameter: let x3=t.
[100123ab]⟹⎩⎨⎧x1+2x3=ax2+3x3=bx3=t⟹⎩⎨⎧x1+2t=ax2+3t=bx3=t⟹⎩⎨⎧x1=a−2tx2=b−3tx3=t⟹x=a−2tb−3tt⟹x=ab0+t−2−31 is the 1-parameter family of solutions

Homogeneous Linear Systems
A system of linear equations is said to be homogeneous if the augmented column is all 0s:
As a matrix equation, a homogeneous system is written:
Ax=0
Notes
- A homogeneous system of equations is always consistent!
- unique solution → trivial solution (x=0)
- infinitely many solutions
- The general solution is a linear combination of basic solutions (vectors that appear next to parameters)
Example
[1010011200] represents a system with general solution:
x=sbasic solution−1100+tbasic solution−10−21
The vectors being multiplied by parameters are non-unique basic solutions.

Example: Solving a Linear System (Unique Solution)
Solve the system of linear equations:
2xx++2y2z−−2z73y===20−7−2z
Steps
- Let's first write this in augmented matrix form:
2x+2y−2z1x+0y+2z0x+3y+2z===27−7→210203−22227−7
- Turn the matrix into RREF using EROs:
⟶⟶⟶⟶⟶⟶210203−22227−7R1↔R21200232−2272−7R2−2R11000232−627−12−721R21000132−327−6−7R3−3R21000102−3117−611111R31000102−317−61R1−2R3R2+3R31000100015−31
- Find all solutions to the SLE:
The coefficient matrix has 3 leading 1s and 3 columns: rank(A)=n ⟹ there is a unique solution.
Rewrite the augmented matrix back into a system of linear equations:
xyz===5−31
Therefore, the unique solution is
x=5−31

Example: Solving a Linear System (Infinitely Many Solutions)
Solve the following system of linear equations:
x+2yy+zz+w===431
The standard form of the linear system is 1x+2y+0z+0w0x+1y+1z+0w0x+0y+1z+1w===431 and the augmented matrix is 100210011001431.
(Note that the order of the variables does not matter, but be consistent!)
We need to reduce the matrix to RREF:
⟶⟶100210011001431R1−2R2100010−211001−231R1+2R3R2−R31000100012−11021
There are 3 leading 1s and 4 columns in the RREF, so rank(A)=3 < n=4, so there are infinitely many solutions.
n−rank(A) = 4−3 = 1 so there is one free variable (1-parameter family of solutions).
Let's assign a parameter to the free variable (the column with a missing leading 1, w) and rewrite the SLE using the RREF.
Let w=t. (Remember to include this equation in the list of equations we obtain from the RREF. We need a solution for all variables!)
⎩⎨⎧x+2wy−wz+ww====021t⟹⎩⎨⎧x+2ty−tz+tw====021t⟹⎩⎨⎧xyzw====−2t2+t1−tt
Therefore, the 1-parameter family of solutions is x=0210+t−21−11.
Solve the following homogeneous system of linear equations. What type of solution is it?
2x1−x2+2x42x2+4x3−x4=0=0