Professional Documents
Culture Documents
MAT216-Lecture 13 (Saba Fatema)
MAT216-Lecture 13 (Saba Fatema)
Contents:
It is possible for a system of linear equations to have exactly one solution, or an infinite
number of solutions, or no solution.
A system of linear equations is called consistent if it has at least one solution and
inconsistent if it has no solution.
Depending upon the number of variables and the number of equations in a system of linear
equations, there are two more criteria.
A system of linear equations is said to be underdetermined if there are more variables than
equations.
Example:
𝑥 2𝑥 3𝑥 4
2𝑥 𝑥 4𝑥 3
On the other hand, a system of linear equations is said to be is overdetermined if there are
more equations than variables.
Example:
𝑥 𝑥 2
2𝑥 3𝑥 3
𝑥 2𝑥 1
𝑎 𝑥 𝑎 𝑥 𝑎 𝑥 ⋯ 𝑎 𝑥 0
𝑎 𝑥 𝑎 𝑥 𝑎 𝑥 ⋯ 𝑎 𝑥 0
𝑎 𝑥 𝑎 𝑥 𝑎 𝑥 ⋯ 𝑎 𝑥 0
𝑎 𝑥 𝑎 𝑥 𝑎 𝑥 ⋯ 𝑎 𝑥 0
𝐴𝐱 𝟎 2
𝑎 𝑎 ⋯ 𝑎 𝑥 0
𝑎 𝑎 ⋯ 𝑎 𝑥 0 ∈ℝ
where, 𝐴 ⋮ ⋮ ⋮ ⋮ , 𝐱 ⋮ ∈ ℝ , and 𝟎
⋮
𝑎 𝑎 ⋯ 𝑎 𝑥 0
𝑥 2𝑥 2𝑥 𝑥 0
3𝑥 6𝑥 5𝑥 4𝑥 0 1
𝑥 2𝑥 3𝑥 0
𝑥
1 2 2 1 𝑥 0
3 6 5 4 𝑥 0
1 2 0 3 𝑥 0
𝐴𝐱 𝟎 2
𝑥
1 2 2 1 𝑥 0
where, 𝐴 3 6 5 4 , 𝐱 𝑥 , and 𝟎 0 .
1 2 0 3 𝑥 0
It is easy to see that a homogeneous system must have at least one solution. If we set 𝑥 0,
0
0
𝑥 0, 𝑥 0, 𝑥 0, i.e., 𝐱 𝟎 in the above system, we find that each of the
0
0
equations is satisfied. Therefore, 𝐱 𝟎 is a solution to this system. This solution is called
trivial solution.
5
For example, 𝐱 1 is a nontrivial solution of (1) because
1
1
1 2 2 1 5 0
3 6 5 4 1 0
1
1 2 0 3 0
1
𝑁 𝐴 𝐱 ∈ ℝ : 𝐴𝐱 𝟎∈ℝ
is called the nullspace of 𝐴. In other words, the nullspace of a matrix 𝐴 is the set of all
solutions of the homogeneous system of linear equations 𝐴𝐱 𝟎.
1 2 2 1
𝐴 3 6 5 4
1 2 0 3
Consider
0 5 1
𝐱 0 , 𝐱 1 , 𝐱 2 .
0 1 2
0 1 1
We see that
0
𝐴𝐱 𝟎 𝐴𝐱 , but 𝐴𝐱 1 𝟎.
2
Saba Fatema Page | 5
Lecturer
Department of Mathematics and Natural Sciences
Brac University
The Nullspace of a Matrix
Theorem 1. If 𝐴 be an 𝑚 𝑛 matrix, then 𝑁 𝐴 is a subspace of ℝ .
𝐴 𝐱 𝐱 𝐴𝐱 𝐴𝐱 𝟎 𝟎 𝟎.
𝐴 𝑐𝐱 𝑐𝐴𝐱 𝑐𝟎 𝟎.
Consider,
1 2 2 1
𝐴 3 6 5 4
1 2 0 3
𝑥
𝑥
𝐱 𝑥 .
𝑥
1 2 2 1
𝐴 3 6 5 4
1 2 0 3
→
→
1 2 2 1 → 1 2 2 1 → 1 2 0 3
⎯⎯⎯⎯⎯⎯⎯⎯ 0 0 1 1 ⎯⎯⎯⎯⎯⎯⎯⎯ 0 0 1 1 ⎯⎯⎯⎯⎯⎯⎯ 0 0 1 1
0 0 2 2 0 0 0 0 0 0 0 0
𝑥 2𝑥 3𝑥 0
𝑥 𝑥 0
𝑥 2𝑠 3𝑡, 𝑥 𝑠, 𝑥 𝑡, 𝑥 𝑡.
𝑥 2𝑠 3𝑡 2 3
𝑥 𝑠 1 0
𝐱 𝑥 𝑠 𝑡
𝑡 0 1
𝑥 𝑡 0 1
2 3
which gives the nullspace 𝑁 𝐴 𝑠 1 𝑡 0 .
0 1
0 1
Note. All the solutions of the above homogeneous system 𝐴𝐱 𝟎 are linear combination of
2 3
two vectors 1 and 0 , i.e. these two vectors span the solution space. When a
0 1
0 1
homogeneous system is solved from the row-echelon form or, from the reduced row-echelon
form, the spanning set is always independent.
2 3
𝐵 1 , 0
0 1
0 1
The number of linearly independent vectors in the row space of a matrix 𝐴 is known as
row rank.
In other words, the dimension of the row space of a matrix 𝐴 is called the row rank of 𝐴.
Similarly, the number of linearly independent vectors in column space of a matrix 𝐴 is known
as column rank, i.e. the dimension of the column space of matrix 𝐴.
Theorem 2 If 𝐴 is an 𝑚 𝑛 matrix, then the row space and column space of 𝐴 have the same
dimension. ∎
Since we generally perform the row reduction to achieve the row echelon form, so the
number of nonzero rows in row echelon form of the matrix 𝐴 gives the rank of 𝐴.
Alternatively, we can say, rank 𝐴 is obtained by the number of pivots in row echelon form
of 𝐴.
1 2 2 1
Observe that, the reduced row echelon form of 𝐴 3 6 5 4 is
1 2 0 3
Pivots 1 2 0 3
0 0 1 1
0 0 0 0
which has two nonzero rows, so the rank of 𝐴 is 2. Alternatively, we can say that there are 2
pivots in the reduced row echelon form of 𝐴, so rank 𝐴 is 2. Here, 𝑥 , 𝑥 are pivot variables.
On the other hand, 𝑥 , 𝑥 are free variables, so the dimension of the nullspace i.e. nullity is
also 2. So, we have the following observation:
rank 𝐴 nullity 𝐴 𝑛.
1 0 2 1 0
𝐴 0 1 3 1 3
2 1 1 1 3
0 3 9 0 12
1 0 2 1 0 1 0 2 0 1
𝐴 0 1 3 1 3 → 0 1 3 0 4
2 1 1 1 3 0 0 0 1 1
0 3 9 0 12 0 0 0 0 0
Since the reduced row echelon form has three nonzero rows, so
rank 𝐴 , 𝑟 3.
nullity 𝐴 𝑛 𝑟
5 3
Theorem 4 Homogeneous linear system with 𝑛 𝑚 has at least one free variable.∎
Saba Fatema Page | 9
Lecturer
Department of Mathematics and Natural Sciences
Brac University
Worked Out Examples
Example 1 Find the nullspace, rank and nullity of the following matrices:
1 2 𝐴
i 𝐴 , ii 𝐵 , iii 𝐶 𝐴 2𝐴
3 8 2𝐴
Solution
1 2
i 𝐴
3 8
𝑥
𝐱 𝑥 .
→
1 2 → 1 2 1 2 → 1 0
𝐴 ⎯⎯⎯⎯⎯⎯⎯⎯ ⎯⎯⎯⎯⎯ ⎯⎯⎯⎯⎯⎯⎯⎯
3 8 0 2 0 1 0 1
1 0 𝑥 0
0 1 𝑥 0
𝑥 0
or, 𝑥 , i.e. 𝑥 0, 𝑥 0.
0
0
Hence, 𝑁 𝐴 .
0
Rank 𝐴 , 𝑟 2.
Nullity 𝐴 𝑛 𝑟 2 2 0. ∎
Note. For the above matrix 𝐴, the nullspace contains only one solution, which is trivial
solution.
𝑥
𝐱 𝑥 .
→
1 2 → 1 2 1 2 1 2 1 0
→
3 8 → 0 2 0 1 → 0 1 → 0 1
𝐵 ⎯⎯⎯⎯⎯⎯⎯⎯ ⎯⎯⎯⎯⎯ ⎯⎯⎯⎯⎯⎯⎯⎯ ⎯⎯⎯⎯⎯⎯⎯⎯
2 4 0 0 0 0 0 0 0 0
6 16 0 4 0 4 0 0 0 0
1 0 0
0 1 𝑥 0
0 0 𝑥 0
0 0 0
𝑥 0
or, 𝑥 , i.e. 𝑥 0, 𝑥 0.
0
0
Hence, 𝑁 𝐵 .
0
Rank 𝐵 , 𝑟 2.
Nullity 𝐵 𝑛 𝑟 2 2 0. ∎
1 2 2 4
iii 𝐶 𝐴 2𝐴
3 8 6 16
𝑥
𝑥
𝐱 𝑥 .
𝑥
1 2 2 4
𝐶
3 8 6 16
→
→ 1 2 2 4 1 2 2 4 → 1 0 2 0
⎯⎯⎯⎯⎯⎯⎯⎯ ⎯⎯⎯⎯⎯ ⎯⎯⎯⎯⎯⎯⎯⎯
0 2 0 4 0 1 0 2 0 1 0 2
𝑥 0
1 0 2 0 𝑥 0
0 1 0 2 𝑥 0
𝑥 0
i.e.
𝑥 2𝑥 0
𝑥 2𝑥 0.
𝑥 2𝑠, 𝑥 2𝑡, 𝑥 𝑠, 𝑥 𝑡.
2𝑠 2 0
2𝑡 0 2
Hence, 𝑁 𝐶 𝑠 𝑡 .
𝑠 1 0
𝑡 0 1
Rank 𝐶 , 𝑟 2.
Nullity 𝐶 𝑛 𝑟 4 2 2. ∎
3 6 21
𝐴 2 4 14
1 2 7
Solution
3 6 21
𝐴 2 4 14
1 2 7
Here, 𝑁 𝐴 𝐱 ∈ ℝ : 𝐴𝐱 𝟎∈ℝ
Now we apply elimination to 𝐴 up to the reduced row-echelon form, and obtain the following
→
3 6 21 ↔ 1 2 7 →
1 2 7
𝐴 2 4 14 ⎯⎯⎯ 2 4 14 ⎯⎯⎯⎯⎯⎯⎯⎯ 0 0 0
1 2 7 3 6 21 0 0 0
1 2 7 𝑥 0
0 0 0 𝑥 0
0 0 0 𝑥 0
Which implies 𝑥 2𝑥 7𝑥 0.
𝑥 2𝑠 7𝑡, 𝑥 𝑠, 𝑥 𝑡.
2𝑠 7𝑡 2 7
𝐱 𝑠 𝑠 1 𝑡 0
𝑡 0 1
2 7
1 , 0
0 1
Example 3 Find (a) a basis for, and (b) the dimension of, the solution space of the following
homogeneous system of linear equations.
4𝑥 𝑦 2𝑧 0 2𝑥 2𝑥 4𝑥 2𝑥 0
i 2𝑥 3𝑦 2𝑧 0 ii 𝑥 2𝑥 𝑥 2𝑥 0
3𝑥 𝑦 𝑧 0 𝑥 𝑥 4𝑥 2𝑥 0
Solution
4𝑥 𝑦 2𝑧 0
i 2𝑥 3𝑦 2𝑧 0
3𝑥 𝑦 𝑧 0
4 1 2 𝑥 0
𝐴 2 3 2 , 𝐱 𝑦 and 𝟎 0
3 1 1 𝑧 0
Here, 𝑁 𝐴 𝐱 ∈ ℝ : 𝐴𝐱 𝟎∈ℝ
4 1 2 1 0 0
𝐴 2 3 2 → 0 1 0
3 1 1 0 0 1
1 0 0 𝑥 0
0 1 0 𝑦 0
0 0 1 𝑧 0
𝑥 0
𝐱 𝑦 0
𝑧 0
(a) Since, all the unknowns 𝑥, 𝑦, 𝑧 are pivot variables, and there is no free variable, so there
is no linearly independent vector in the solution space of the given system 𝐴𝐱 𝟎. Hence, A
basis for the solution space of 𝐴𝐱 𝟎 is given by the empty set ∅ .
2𝑥 2𝑥 4𝑥 2𝑥 0
ii 𝑥 2𝑥 𝑥 2𝑥 0
𝑥 𝑥 4𝑥 2𝑥 0
𝑥
2 2 4 2 𝑥 0
𝐴 1 2 1 2 , 𝐱 𝑥 and 𝟎 0
1 1 4 2 𝑥 0
Here, 𝑁 𝐴 𝐱 ∈ ℝ : 𝐴𝐱 𝟎∈ℝ
2 2 4 2 1 0 0 5/8
𝐴 1 2 1 2 → 0 1 0 15/8
1 1 4 2 0 0 1 9/8
Which implies
5
𝑥 𝑥 0
8
15
𝑥 𝑥 0
8
9
𝑥 𝑥 0
8
5 15 9
𝑥 𝑡, 𝑥 𝑡, 𝑥 𝑡, 𝑥 𝑡
8 8 8
5 5
⎡ 𝑡⎤ ⎡ ⎤
⎢ 8 ⎥ ⎢ 8 ⎥
⎢ 15 𝑡⎥ ⎢ 15⎥
𝐱 ⎢ 8 ⎥ 𝑡⎢ 8 ⎥
⎢ 9 ⎥ ⎢ 9⎥
⎢ 8𝑡 ⎥ ⎢ 8⎥
⎣ 𝑡 ⎦ ⎣ 1⎦
(a) A basis for the solution space of 𝐴𝐱 𝟎 is given by the following set
5
⎧⎡ ⎤⎫
⎪⎢ 8⎥⎪
⎪ 15 ⎪
⎢ ⎥
⎢ 8⎥
⎨⎢ 9 ⎥ ⎬
⎪
⎪⎢ 8 ⎥ ⎪
⎪
⎩⎣ 1 ⎦ ⎭
2. For the following matrices, find (a) a basis for, and (b) the dimension of, the solution
space of 𝐴𝐱 𝟎.
1 3 2 4
2 1 1 2 3
i 𝐴 ii A iii 𝐴 0 1 1 2
1 3 0 1 0
2 6 4 8
3. Find (a) a basis for, and (b) the dimension of, the solution space of the following
homogeneous system of linear equations.
𝑥 𝑦 𝑧 0 3𝑥 3𝑥 15𝑥 11𝑥 0
i 3𝑥 𝑦 0 ii 𝑥 3𝑥 𝑥 𝑥 0
2𝑥 4𝑦 5𝑧 0 2𝑥 3𝑥 11𝑥 8𝑥 0