Scribd
Upload
532 views6 pages

Rank and Nullity Theorem

Rank and Nullity theorem states that for any linear map f: Rn → Rm, the dimension of the image space (im(f)) plus the dimension of the null space (null(f)) equals n. This theorem can be proved using properties of matrices. Elementary row and column operations preserve the rank and nullity of a matrix. Any matrix A can be reduced, using elementary operations, to a block matrix with an identity matrix of size equal to the rank of A. This rank is equal to the column and row rank of A.

Uploaded by

sdfsdf
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
532 views6 pages

Rank and Nullity Theorem

Rank and Nullity theorem states that for any linear map f: Rn → Rm, the dimension of the image space (im(f)) plus the dimension of the null space (null(f)) equals n. This theorem can be proved using properties of matrices. Elementary row and column operations preserve the rank and nullity of a matrix. Any matrix A can be reduced, using elementary operations, to a block matrix with an identity matrix of size equal to the rank of A. This rank is equal to the column and row rank of A.

Uploaded by

sdfsdf
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/ 6

Rank and Nullity theorem

Rank and Nullity of a matrix


I

Given a linear map f : Rn Rm , its image space or range


space is defined as im(f ) = {f (x) : x Rn } and its kernel
or null space is defined by null(f ) = {x : f (x) = 0}.

Similarly, the same notions are defined for a matrix


A Rmn which represents f .

Dually, we have similar definitions Im(AT ) and null(AT ) for


the transpose AT Rnm

im(f ) and null(f ) are both subspaces (i.e., closed under


vector addition and scalar multiplication).

The celebrated Rank-Nullity theorem says that for any


linear map f : Rn Rm we have:
dim(im(f )) + dim(null(f )) = n.

This theorem is an immediate consequence of the


following results on matrices.

Rank and nullity of matrices


I

Let A Rmn , we show that

column rank(A) = row rank(A) := rank(A).

column rank(A) + nullity(A) = n.

column rank(A) + nullity(AT ) = m.

These properties are consequence of the following facts:

(i) An elementary row or column operation does not change


the column rank or the row rank of A.
(ii) Using these elementary operations A can be reduced to
the following block matrix of an identity matrix Ir r of
dimension r and three zero matrices 0r (nr ) Rr (nr ) ,
0(mr )r R(mr )r and 0(mr )(nr ) R(mr )(nr ) :


Ir r
0r (nr )
0(mr )r 0(mr )(nr )

Elementary operations preserve rank and nullity I


I
I

I
I
I

I
I
I

Claim 1. An elementary row operation on A Rmn does


not change its column rank.
Proof. Since the column rank of A is the maximum number
of columns of A that are linearly independent, it is sufficient
(by an exercise in Tutorial Sheet 4) to show that an
elementary row operation does not change the linear
independence of the columns of A.
Recall that columns of A are linearly independent iff
Ax = 0 implies x = 0.
Let A0 be the result of an elementary row operation on A.
We show that Ax = 0 iff A0 x = 0 which proves that
columns of A and those of A0 are either both linearly
independent or both linearly dependent (why?).
We show this for the elementary row operation of
subtracting times the second row a2 from the first row a1 :
Ax = 0 iff a1 x = a2 x = = am x = 0 iff
(a1 a2 ) x = a2 x = = am x = 0 iff A0 x = 0.
The proof for the other elementary operations is similar.

Elementary operations preserve rank and nullity II


I
I

I
I

I
I

Claim 2. An elementary row operation does not change


the linear independence of the rows of a matrix.
Proof. By taking the transpose of the matrix, we can
equivalently show that an elementary column operation
does not change the linear independence of the columns
of a matrix A Rmn .
Again the case of swapping two rows or multiplying a row
by a non-zero number are trivial.
We show it for A00 = [a1 a2 , a2 , , an ] obtained by the
elementary column operation of subtracting times the
second column a2Pfrom the first column a1 of A.
Recall that Ax = ni=1 xi ai .
If a1 , a2 , , an are linearly independent and if A00 x = 0,
then x1 (a1 a2 ) + x2 a2 + xn an = 0, i.e.,
x1 a1 + (x2 x1 )a2 + xn an = 0, which by the linear
independence of ai s implies x1 = x2 x1 = . . . = xn = 0,
i.e., x1 = x2 = . . . = xn = 0. Thus, the columns of A00 are
linearly independent.

Elementary operations preserve rank and nullity III


I

I
I
I

Now for the converse, assume that the columns of


A00 = [a1 a2 , a2 , , an ] are linearly independent and
that Ax = 0.
Then x1 a1 + x2 a2 + x3 a1 + + xn an = 0 and thus
x1 (a1 a2 ) + (x2 + x1 )a2 + + xn an = 0 which gives
x1 = x2 + x1 = x3 = = xn = 0,
which implies xi = 0 for all i.
Thus, the columns of A are linearly independent, which
completes the proof of that elementary row operations do
not change the column or row rank of a matrix. By
considering AT , the same is true for elementary column
operations.
Finally, we know that using Gauss-Jordan technique we
can reduce any m n matrix by elementary row and
column operations to a diagonal matrix where all its
non-zero entries are on the diagonal.
This completes the proof of the rank-nullity theorem.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy