Week 7 (Lecture 14)

Contents

 Introduction to Nonhomogeneous system of linear equations

 Particular solution
 Theorem 1: Difference between particular solutions of a nonhomogeneous system is
a solution of the associated homogeneous system
 Theorem 2: Construction of general solution of a nonhomogeneous system
 Dependency of the solution of 𝐴𝐱 𝐛 on the rank of the matrix 𝐴
 Worked out examples
 Exercises

Saba Fatema Page | 1

Lecturer
Department of Mathematics and Natural Sciences
Brac University 
 
 Solutions of Nonhomogeneous Linear Systems
If 𝐴 be an 𝑚 𝑛 matrix and 𝐱 ∈ ℝ and 𝐛 ∈ ℝ , the matrix equation

𝐴𝐱 𝐛 1

where 𝐛 𝟎 is called a nonhomogeneous system of linear equations.

The matrix equation

𝐴𝐱 𝟎 2

is called the associated homogeneous system of linear equations to (1).

For example, the following is a nonhomogeneous system of linear equations

𝑥 𝑏𝑥 2𝑥 𝑥 5

3𝑥 𝑥 5𝑥 𝑥 8

𝑥 2𝑥 2𝑥 5𝑥 9

The associated homogeneous system to the above is

𝑥 𝑏𝑥 2𝑥 𝑥 0

3𝑥 𝑥 5𝑥 𝑥 0

𝑥 2𝑥 2𝑥 5𝑥 0

In the previous lecture, we saw that the set of all solution vectors of the homogeneous linear
system 𝐴𝐱 𝟎 is a subspace (nullspace or solution space).

Is this true also of the set of all solution vectors of the nonhomogeneous linear system
𝐴𝐱 𝐛?

The answer is NO. Because the zero vector is never a solution of a nonhomogeneous linear
system.

Saba Fatema Page | 2

Lecturer
Department of Mathematics and Natural Sciences
Brac University 
 
 Solutions of Nonhomogeneous Linear Systems
Particular Solution

The vector 𝐱 𝐱 is called a particular solution of (1) if the vector satisfies 𝐴𝐱 𝐛.

For example,

5
𝐱 7
0
0

is a particular solution of
𝑥 𝑏𝑥 2𝑥 𝑥 5

3𝑥 𝑥 5𝑥 𝑥 8

𝑥 2𝑥 2𝑥 5𝑥 9

To verify this—

1 0 2 1 5
𝐴 3 1 5 0 and 𝐛 8
1 2 0 5 9

1 0 2 1 5 5
𝐴𝐱 3 1 5 0 7 𝐛
8
0
1 2 0 5 9
0

Saba Fatema Page | 3

Lecturer
Department of Mathematics and Natural Sciences
Brac University 
 
 Solutions of Nonhomogeneous Linear Systems
There is a relationship between the sets of solutions of the two systems 𝐴𝐱 𝟎 and 𝐴𝐱 𝐛.

Theorem 1 If 𝐱 and 𝐱 be particular solutions of a nonhomogeneous system 𝐴𝐱 𝐛, then

𝐱 𝐱 is a solution of the associated homogeneous system 𝐴𝐱 𝟎.

Proof. Define 𝐯 𝐱 𝐱 . Now

𝐴𝐯 𝐴 𝐱 𝐱 𝐴𝐱 𝐴𝐱

𝐛 𝐛

This shows that 𝐱 𝐱 is a solution of the associated homogeneous system 𝐴𝐱 𝟎. ∎

Theorem 2 Let 𝐴 be an 𝑚 𝑛 matrix. Also let 𝐵 𝐯 ,𝐯 ,…𝐯 is a basis of the nullspace of

𝐴. If 𝐱 is any particular solution and 𝐮 is an arbitrary solution of the nonhomogeneous
system 𝐴𝐱 𝐛, then there exist scalars 𝑐 , 𝑐 , … 𝑐 such that

𝐮 𝐱 𝑐 𝐯 𝑐 𝐯 ⋯ 𝑐 𝐯

Proof. Since 𝐮 and 𝐱 are solutions of 𝐴𝐱 𝐛, then by the Theorem 1, 𝐮 𝐱 is a solution of

𝐴𝐱 𝟎. Therefore, 𝐮 𝐱 ∈ 𝑁 𝐴 . Since 𝐵 is a basis of 𝑁 𝐴 , there exist scalars 𝑐 , 𝑐 , … 𝑐
such that

𝐮 𝐱 𝑐 𝐯 𝑐 𝐯 ⋯ 𝑐 𝐯

𝐮 𝐱 𝑐 𝐯 𝑐 𝐯 ⋯ 𝑐 𝐯

This completes the proof of the theorem. ∎

Naturally, the expression 𝐱 𝑐 𝐯 𝑐 𝐯 ⋯ 𝑐 𝐯 is called the general/complete

solution of 𝐴𝐱 𝐛.
Saba Fatema Page | 4
Lecturer
Department of Mathematics and Natural Sciences
Brac University 
 
 Solutions of Nonhomogeneous Linear Systems
For example, consider the following system.

𝑥 𝑏𝑥 2𝑥 𝑥 5

3𝑥 𝑥 5𝑥 𝑥 8

𝑥 2𝑥 2𝑥 5𝑥 9

If 𝐱 be a particular solution and 𝐵 be a basis of the nullspace of the associated homogeneous

system, where

5 2 1
𝐱 7 , 𝐵 1 , 3
0 1 0
0 0 1

then the general solution is given by

5 2 1
𝐱 7 𝑐 1 𝑐 3 .
0 1 0
0 0 1

Example Solve the following system

𝑥 𝑏𝑥 2𝑥 𝑥 5

3𝑥 𝑥 5𝑥 𝑥 8

𝑥 2𝑥 2𝑥 5𝑥 9

Solution Let’s write the system in the form 𝐴𝐱 𝐛, where

𝑥
1 0 2 1 𝑥 5
𝐴 3 1 5 0 , 𝐱 𝑥 , 𝐛 8
1 2 0 5 𝑥 9

By applying the Gauss–Jordan elimination to the augmented matrix we obtain

Saba Fatema Page | 5

Lecturer
Department of Mathematics and Natural Sciences
Brac University 
 
 Solutions of Nonhomogeneous Linear Systems
→
1 0 2 1 5 →
1 0 2 1 5
3 1 5 0 8 ⎯⎯⎯⎯⎯⎯⎯⎯ 0 1 1 3 7
1 2 0 5 9 0 2 2 6 14

→ 1 0 2 1 5
⎯⎯⎯⎯⎯⎯⎯⎯ 0 1 1 3 7
0 0 0 0 0

The above matrix gives

𝑥 𝑥 2𝑥 𝑥 5

𝑥 𝑥 3𝑥 7

Since 𝑥 and 𝑥 are free variables, set 𝑥 𝑠 and 𝑥 𝑡 where 𝑠, 𝑡 ∈ ℝ. Then

𝑥 5 2𝑠 𝑡, 𝑥 7 𝑠 3𝑡, 𝑥 𝑠, 𝑥 𝑡

Therefore, the general solution is given by

5 2𝑠 𝑡 5 2 1
𝐱 7 𝑠 3𝑡 7 𝑠 1 𝑡 3 .
𝑠 0 1 0
𝑡 0 0 1

In the next example, we will establish a condition, for which a nonhomogeneous system will
be consistent.

Example

𝑥 𝑥 𝑏
𝑥 2𝑥 𝑏
2𝑥 3𝑥 𝑏

Let’s reduce the augmented matrix,

1 1 𝑏1 1 0 2𝑏1 𝑏2
1 2 𝑏2 → 0 1 𝑏2 𝑏1
2 3 𝑏3 0 0 𝑏1 𝑏2 𝑏3
Saba Fatema Page | 6
Lecturer
Department of Mathematics and Natural Sciences
Brac University 
 
 Solutions of Nonhomogeneous Linear Systems
The above system will be consistent if

𝑏 𝑏 𝑏 0

Considering the above condition, the given system provides the following solution

𝑥 2𝑏 𝑏
𝐱 𝑥 .
𝑏 𝑏

Next, we are going to discuss on the dependency of the solution of 𝐴𝐱 𝐛 on rank 𝑟.

Case I. 𝐴 is a square invertible matrix

In 𝐴𝐱 𝐛, if 𝐴 is a square invertible matrix, i.e., 𝑚 𝑛 𝑟, then the particular solution is the

one and only solution 𝐱 𝐴 𝐛. There will exist no free variables and hence no special
solution.

For example, consider the system

𝑥 2𝑥 1
2𝑥 3𝑥 3

Here,

1 2 1 3 2
𝐴 , 𝐴
2 3 2 1

3 2 1 3
𝐱 𝐴 𝐛
2 1 3 1

Since the nullspace 𝑁 𝐴 consists only the trivial solution, the complete solution to 𝐴𝐱 𝐛
𝑥 3
is given by 𝐱 𝑥 .
1

Saba Fatema Page | 7

Lecturer
Department of Mathematics and Natural Sciences
Brac University 
 
 Solutions of Nonhomogeneous Linear Systems
Case II. 𝐴 has full column rank

For an 𝑚 𝑛 matrix 𝐴 (with 𝑚 𝑛), if the column rank 𝐴 𝑛, then 𝐴 is called a matrix
with 𝐟𝐮𝐥𝐥 𝐜𝐨𝐥𝐮𝐦𝐧 𝐫𝐚𝐧𝐤 𝑟 𝑛 . For 𝐴𝐱 𝐛, if 𝐴 has full column rank, then the system has
the following properties:

 All columns of 𝐴 are pivot columns.

 There are no free variables, and hence no special solutions.
 The nullspace 𝑁 𝐴 contains only the zero vector 𝐱 𝟎.
 𝐴𝐱 𝐛 may have only one solution, or none.

Let’s have an example.

Consider the following system,

𝑥 𝑥 2
2𝑥 3𝑥 3
𝑥 2𝑥 1

It can be written as,

𝐴𝐱 𝐛

1 1 𝑥 2
where, 𝐴 2 3 , 𝐱 𝑥 , 𝐛 3
1 2 1

Applying row reduction to augmented matrix, we obtain

1 1 2 1 0 3
2 3 3 → 0 1 1
1 2 1 0 0 0
𝑥 3
The general solution will be 𝐱 𝑥 .∎
1

3 0
Note. Since, 𝐱 𝐱 𝐱 , this implies that, the associated homogeneous
1 0
system 𝐴𝐱 𝟎 has the trivial solution only, i.e. 𝑁 𝐴 contains the zero vector only.
Saba Fatema Page | 8
Lecturer
Department of Mathematics and Natural Sciences
Brac University 
 
 Solutions of Nonhomogeneous Linear Systems
Consider another system,

𝑥 𝑥 2
2𝑥 3𝑥 3
𝑥 2𝑥 5

After reduction, augmented matrix becomes

1 1 2 1 0 0
2 3 3 → 0 1 0
1 2 5 0 0 1

Since the last row provides a false statement, the system is inconsistent.

Case III. 𝐴 has full row rank

For an 𝑚 𝑛 matrix 𝐴 (with 𝑚 𝑛), if the row rank 𝐴 𝑚, then 𝐴 is called a matrix with
𝐟𝐮𝐥𝐥 𝐫𝐨𝐰 𝐫𝐚𝐧𝐤 𝑟 𝑚 . In such case 𝐴𝐱 𝐛 has the following properties:

 All rows have pivots, i.e., the reduced row echelon form, 𝑅 has no zero rows.
 There will exist free variable s , and hence special solution s .
 The column space is the whole space ℝ .
 There are 𝑛 𝑟 𝑛 𝑚 special solutions in the nullspace of 𝐴.
 𝐴𝐱 𝐛 has infinitely many solutions.

Let’s see an example.

Consider the following system

𝑥 𝑥 𝑥 3
𝑥 2𝑥 𝑥 4

Applying elimination in the augmented matrix

𝐴|𝐛
1 1 1 3
→ 1 0 3 2
1 2 1 4 0 1 2 1

Saba Fatema Page | 9

Lecturer
Department of Mathematics and Natural Sciences
Brac University 
 
 Solutions of Nonhomogeneous Linear Systems
we get,

𝑥 3𝑥 2
𝑥 2𝑥 1

since 𝑥 is free variable, set 𝑥 𝑠, where 𝑠 ∈ ℝ. Then

𝑥 2 3𝑠, 𝑥 1 2𝑠, 𝑥 𝑠

So, the complete solution is given by

𝑥 2 3𝑠 2 3
𝐱 𝑥 1 2𝑠 1 𝑠 2 .
𝑥 𝑠 0 1

Case IV. 𝐴 is not full rank i.e. 𝑟 𝑚, 𝑟 𝑛

In such case, 𝐴𝐱 𝐛 has no solution, or has infinitely many solutions.

Let’s have examples.

Consider

𝑥 2𝑥 3𝑥 7
2𝑥 𝑥 3𝑥 8𝑥 17
3𝑥 𝑥 5𝑥 11𝑥 24

Augmented matrix becomes

1 0 2 3 7 1 0 2 3 7
2 1 3 8 17 → 0 1 1 2 3
3 1 5 11 24 0 0 0 0 0

Pivot variables are 𝑥 , 𝑥 ; and free variables are 𝑥 , 𝑥 .

So, set 𝑥 𝑠, 𝑥 𝑡, where 𝑠, 𝑡 ∈ ℝ. Then

𝑥 7 2𝑠 3𝑡, 𝑥 3 𝑠 2𝑡, 𝑥 𝑠, 𝑥 𝑡

Saba Fatema Page | 10

Lecturer
Department of Mathematics and Natural Sciences
Brac University 
 
 Solutions of Nonhomogeneous Linear Systems
7 2𝑠 3𝑡 7 2 3
3 𝑠 2𝑡 3 1 2
𝐱 𝑠 𝑡 .
𝑠 0 1 0
𝑡 0 0 1

Consider another example,

𝑥 2𝑥 3𝑥 7
2𝑥 𝑥 3𝑥 8𝑥 8
3𝑥 𝑥 5𝑥 11𝑥 24

Augmented matrix becomes

1 0 2 3 7 1 0 2 3 0
2 1 3 8 17 → 0 1 1 2 0
3 1 5 11 24 0 0 0 0 1

Last row of the above matrix gives a false statement. Hence, the system is inconsistent. ∎

Dependency of 𝑨𝐱 𝐛 on rank 𝑟 :

Relation among 𝒓, 𝒎, 𝒏 Type/shape of matrix 𝑨 No. of solution to 𝑨𝐱 𝐛

𝑟 𝑚 𝑛 Square, Invertible One solution

𝑟 𝑛 and 𝑚 𝑛 Tall, Thin No solution, or One solution

𝑟 𝑚 and 𝑚 𝑛 Short, Wide Infinitely many solutions

𝑟 𝑚 and 𝑟 𝑛 Not full rank No solution or, Infinitely

many solutions

Saba Fatema Page | 11

Lecturer
Department of Mathematics and Natural Sciences
Brac University 
 
 Solutions of Nonhomogeneous Linear Systems
Exercise

1. a Determine whether the nonhomogeneous system 𝑨𝐱 𝐛 is consistent, and b if

the system is consistent, then find the complete solution in the form 𝐱 𝐱 𝐱 ,
where 𝐱 is a particular solution of 𝑨𝐱 𝐛 and 𝐱 is a solution of 𝑨𝐱 𝟎 and.

𝑥 3𝑦 10𝑧 18
3𝑥 2𝑥 16𝑥 2𝑥 7
2𝑥 7𝑦 32𝑧 29
i ii 𝑥 5𝑥 14𝑥 18𝑥 29
𝑥 3𝑦 14𝑧 12
3𝑥 𝑥 14𝑥 2𝑥 1
𝑥 𝑦 2𝑧 8

2. a Under what condition on 𝑏 , 𝑏 and 𝑏 , the following system will be solvable or,
consistent ? b Find 𝐱 in that case.
𝑥 3𝑦 𝑧 𝑏
3𝑥 8𝑦 2𝑧 𝑏
2𝑥 4𝑦 𝑏

Saba Fatema Page | 12

Lecturer
Department of Mathematics and Natural Sciences
Brac University