Scribd
Upload
15 views

Lec17 - 06 10 2023

This document discusses subspaces associated with matrices, including row space, column space, and their properties. It defines row space and column space, and proves that row-equivalent matrices have the same row space. It discusses how to find bases for the row and column spaces. Examples are provided to find row space, column rank, column space, and bases. The relationship between row and column ranks is proved. Applications of the rank-nullity theorem to homogeneous and non-homogeneous systems are described.

Uploaded by

aryan.pandey23b
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
15 views

Lec17 - 06 10 2023

This document discusses subspaces associated with matrices, including row space, column space, and their properties. It defines row space and column space, and proves that row-equivalent matrices have the same row space. It discusses how to find bases for the row and column spaces. Examples are provided to find row space, column rank, column space, and bases. The relationship between row and column ranks is proved. Applications of the rank-nullity theorem to homogeneous and non-homogeneous systems are described.

Uploaded by

aryan.pandey23b
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 10

Lec 17

Subspaces associated with matrices

Let 𝐴 = 𝑎 be a 𝑚 × 𝑛 matrix.
×

𝑅
𝑅
𝐴 = [𝐶 , 𝐶 , … , 𝐶 ] =
⋮
𝑅

Definition: The row space of 𝐴 is the subspace of ℝ spanned by the row


vectors 𝑅 , 𝑅 , … , 𝑅 of 𝐴.

𝑟𝑜𝑤 𝑠𝑝𝑎𝑐𝑒(𝐴) = 𝑠𝑝𝑎𝑛{𝑅 , 𝑅 , … , 𝑅 }


Notation: 𝑟𝑜𝑤(𝐴).

Row rank (A)= dim(row space of 𝐴)

Example 1:

Let

What is row space of 𝐴 and row rank of A?

Theorem 3.20: Let 𝐵 be any matrix that is row equivalent to a matrix 𝐴. Then the
row spaces of 𝐴 and 𝐵 are equal ie. 𝑟𝑜𝑤𝑠𝑝𝑎𝑐𝑒(𝐵) = 𝑟𝑜𝑤𝑠𝑝𝑎𝑐𝑒(𝐴).

Proof:
Since 𝐴 and 𝐵 are row equivalent, the rows of 𝐵 are obtained by a finite no. of
ERO's on 𝐴.
Thus each row of B is a linear combination of rows of A.
Let 𝒙 be an element in the row space of B.
Then 𝑥 is a linear combination of the rows of B, which are linear combinations of
the rows of A. Therefore, 𝑥 is a linear combination of the rows of A.
Therefore, 𝑥 is in the row space of 𝐴.
Hence the row space of 𝐵 ⊆ row space of 𝐴------(1)

Similarly, by reverse ERO's A can be obtained from B.


So the row space (A) ⊆ row space(B) -------(2).
Hence from (1) and (2), 𝑟𝑜𝑤𝑠𝑝𝑎𝑐𝑒(𝐵) = 𝑟𝑜𝑤𝑠𝑝𝑎𝑐𝑒(𝐴).

Question: How to find basis for Row space of A?

Example 2:

Find the row space and row rank of 𝐴.

Solution:

Row space = 𝑠𝑝𝑎𝑛{𝑅 , 𝑅 , 𝑅 } =

To find row rank- reduce A to RREF form.

Row rank=

Definition: The column space of 𝐴 is the subspace of ℝ spanned by the


column vectors 𝐶 , 𝐶 , … , 𝐶 of 𝐴.
Column space(𝐴) = 𝑠𝑝𝑎𝑛{𝐶 , 𝐶 , … , 𝐶 }
Notation: Col(𝐴)
Column rank(A)= dim(column space of 𝐴).

Note:
Elementary row operations do not affect the row space and null space of a matrix.
Whereas, the ERO's do affect the column space.

Question: How to find the basis for the column space?

Example 3: Let

Find the row space, row rank, column space, column rank and the
basis for the row space and column space of A.

Solution:

The RREF of 𝐴 is

Row space of 𝐴 = {𝛼 𝑅 + 𝛼 𝑅 + 𝛼 𝑅 : 𝛼 , 𝛼 , 𝛼 ∈ ℝ}
= 𝑠𝑝𝑎𝑛{(1, 1,2), (1, 2,5), (5,3,4)}
Two non-zero L. I rows in R and hence row rank = 2.
Basis for Row space of A= 𝑠𝑝𝑎𝑛{(1,1,2), (1, 2, 5)}
since (5,3,4) = 7(1,1,2) − 2(1,2,5).

Column space of 𝐴 = {α 𝐶 + α 𝐶 + α 𝐶 : α , α , α ∈ ℝ}
= 𝑠𝑝𝑎𝑛{(1,1,5), (1,2,3), (2, 5, 4)}
𝐶 and 𝐶 of R are L. I and hence column rank = 2.
Basis for column space of 𝐴 = 𝑠𝑝𝑎𝑛{(1,1,5), (1, 2,3)}
since (2,5,4) = −1(1,1,5) + 3(1,2,3).

Note: Here Row rank(A) = Column rank(A) =2.

Recall:
Linear Transformation associated with matrix
T : ℝ → ℝ given by

𝑇 (𝒙) = 𝐴𝒙

Range(𝑇 ) = 𝑠𝑝𝑎𝑛{𝑇 (𝑒 ), 𝑇 (𝑒 ), … , 𝑇 (𝑒 )}
= 𝑠𝑝𝑎𝑛{𝐶 , 𝐶 , … , 𝐶 }
=column space of 𝐴.

• rank(𝑇 ) = dim𝑅(𝑇 )= column rank of A = no. of L. I columns of 𝐴.


• rank(𝑇 ) ≤ 𝑚

Similarly, define 𝐿 : ℝ → ℝ as

𝐿 (𝒙 ) = 𝐴 𝒙

Range(𝐿 ) = 𝑠𝑝𝑎𝑛{𝐿 (𝑒 ), 𝐿 (𝑒 ), … , 𝐿 (𝑒 )}
= 𝑠𝑝𝑎𝑛{𝑅 , 𝑅 , … , 𝑅 }
= Row space of 𝐴.

• rank(𝐿 ) = dim𝑅(𝐿 )= row rank of 𝐴 = no. of L. I rows of 𝐴.


• rank(𝐿 ) ≤ 𝑛.

Note:
Let 𝐴 be an 𝑚 × 𝑛 matrix.

The column space of 𝐴 is equal to the Row space of 𝐴 .


The Row space of 𝐴 is equal to the Column space of 𝐴 .

Theorem 3.24: For any 𝑚 × 𝑛 matrix 𝐴, the row rank and column rank are the same.

Proof:
Definition: For a 𝑚 × 𝑛 matrix 𝐴, rank of 𝐴 is defined as the row rank or
the column rank of 𝐴.
Notation: rank(A) or r(A).

• Note that 𝑟𝑎𝑛𝑘 (𝐴) = 𝑟𝑎𝑛𝑘(𝐴 ).

Some applications of the Rank-Nullity Theorem:

Result 1 : Let 𝐴 be an 𝑚 × 𝑛 matrix and consider the homogeneous system 𝐴𝒙 = 𝟎.


Then
(1) If 𝑟𝑎𝑛𝑘 (𝐴) = 𝑛, then 𝐴𝒙 = 𝟎 has only the trivial solution, so nullspace of 𝐴 = {𝟎}.
(2) If 𝑟𝑎𝑛𝑘 (𝐴) = 𝑟 < 𝑛, then 𝐴𝑥 = 0 has infinite number of solutions , all of which
can be obtained from
𝑥 = 𝑐 𝑥 + 𝑐 𝑥 + ⋯+ 𝑐 𝑥
where {𝑥 , 𝑥 , … , 𝑥 } is any linearly independent set of 𝑛 − 𝑟 solutions to 𝐴𝒙 = 𝟎.
Result 2: Let 𝐴 be an 𝑚 × 𝑛 matrix. For 𝐴𝒙 = 𝒃, the following holds:

(1) 𝐴𝐱 = 𝐛 has a solution iff rank(𝐴|𝑏) = 𝑟𝑎𝑛𝑘(𝐴).


(2) 𝐴𝐱 = 𝐛 has a unique solution for all 𝐛 ∈ ℝ iff 𝑟𝑎𝑛𝑘 (𝐴|𝑏) = 𝑟𝑎𝑛𝑘 (𝐴) = 𝑛.
(3) 𝐴𝐱 = 𝐛 has infinitely many solutions for all b ∈ ℝ iff 𝑟𝑎𝑛𝑘 (𝐴|𝑏) = 𝑟𝑎𝑛𝑘(𝐴) < 𝑛.

Another interpretation of the above result:

The Rank - Nullity Theorem:


Let 𝑉 and 𝑊 be two finite dimensional real vector spaces. If 𝑇: 𝑉 → 𝑊 is a L.T., then
𝑟𝑎𝑛𝑘 (𝑇) + 𝑛𝑢𝑙𝑙𝑖𝑡𝑦(𝑇) = dim(𝑉).

Proof:

You might also like