Scribd
Upload
15 views

Lecture 05

1. The lecture discusses methods to determine if a square matrix is invertible and how to find its inverse. 2. A matrix is invertible if it can be transformed into the identity matrix using elementary row operations. 3. The Gauss-Jordan method can be used to simultaneously solve the systems Ax=e1, Ax=e2, etc. to find the inverse, by transforming the augmented matrix [A|I] into [I|X] where X is the inverse.

Uploaded by

Kushal Ramakrishna
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
15 views

Lecture 05

1. The lecture discusses methods to determine if a square matrix is invertible and how to find its inverse. 2. A matrix is invertible if it can be transformed into the identity matrix using elementary row operations. 3. The Gauss-Jordan method can be used to simultaneously solve the systems Ax=e1, Ax=e2, etc. to find the inverse, by transforming the augmented matrix [A|I] into [I|X] where X is the inverse.

Uploaded by

Kushal Ramakrishna
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 12

Linear Algebra

Lecture 05

Prof. Shreedevi K. Masuti


IIT Dharwad

14 Jan 2020

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 05


Summary of Lecture 4

• Row Canonical Form (RCF) or Reduced Row Echelon Form (RREF)


• Elementary Matrices
• Elementary matrices are invertible and inverse is of the same type
• Elementary row operation of type I , II and III on A correspond to
pre-multiplying A by an elementary matrix of type I , II and III ,
respectively

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 05


Learning Outcome

• Various criterion for invertibility of a matrix


• Finding the inverse of a matrix using Gauss-Jordan Method

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 05


Inverse of a square matrix

Definition
Let A be an n × n matrix. If there is an n × n matrix B such that
AB = In×n = BA then we say A is invertible and B is the inverse of A.
The inverse of A is denoted by A−1 .

Proposition
For any invertible square matrix A, (At )−1 = (A−1 )t .

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 05


A Criterion for invertible Matrices

Theorem
Let A be an n × n square matrix. Then the following statements are
equivalent:
(1) A is invertible.
(2) The system Ax = 0 has only the zero solution x = 0.
(3) A can be transformed to In×n by a sequence of elementary row
operations.
(4) The RCF of A is In×n .
(5) A is a product of elementary matrices.

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 05


Proof: (1) ⇒ (2) : Let B be the inverse of A. Then

Ax = 0 =⇒ B(Ax) = B0 =⇒ (BA)x = 0 =⇒ Ix = x = 0.

(2) ⇒ (3) : Let A0 be its REF. Then A has n nonzero rows as Ax = 0


has only the trivial solution. Hence using ERO’s again A can be
transformed to RCF with n nonzero rows and hence is In×n .
(3) ⇒ (4) : Clear.
(4) ⇒ (5) : Since RCF of A is In×n , A can be transformed to In×n by
elementary row operations. Since each ERO corresponds to
pre-multiplying by elementary matrix, we have

E1 . . . Ek A = In×n

for some elementary matrices E1 , . . . , Ek .

A = E−1 −1
k . . . E1 .

Since the inverse of an elementary matrix is again elementary, the result


follows.

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 05


(5) ⇒ (1) : Let
A = E1 . . . Ek
for some elementary matrices E1 . . . Ek . Then by tutorial 1, we know A is
invertible and
A−1 = E−1 −1
k . . . E1 .

Corollary
Let A ∈ Mn,n . If there is B ∈ Mn,n such that either BA = I or AB = I,
then A is invertible, and A−1 = B.

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 05


Note: The above result is a definite improvement over requiring the
existence of a matrix B satisfying both BA = I and AB = I for the
invertibility of a square matrix A.

Proposition
Let A and B be square matrices. Then AB is invertible if and only if A
and B are invertible, and then (AB)−1 = B−1 A−1 .

Remark: We see that Ek · · · E1 A = I implies A−1 = Ek · · · E1 I. Thus if


we apply the row operations that take A to I on I, we get A−1 .

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 05


Finding the inverse of an invertible Matrix

1 Remark: Suppose an n × n square matrix A is invertible.


2 Now Ax = b ⇐⇒ Ix = c for x ∈ Mn,1 .
3 Thus c is the unique solution of Ax = b.
4 This observation is the basis of an important method to find the
inverse of a square matrix.
5 Gauss-Jordan Method for Finding the Inverse of a Matrix:
6 Let A ∈ Rn×n be an invertible matrix.
7 Consider the basic column vectors e1 , . . . , en ∈ Mn,1 .
 
8 Then e1 · · · en = In×n .

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 05


Gauss-Jordan Method

1 Let x1 , . . . , xn be the unique elements of Rn×1 be such that


Ax1 = e1 , . . . , Axn = en , and define X := x1 · · · xn .
2 Then
     
AX = A x1 ··· xn = Ax1 ··· Axn = e1 ··· en = I.

3 By an earlier result, it follows that X = A−1 .


4 Hence to find A−1 , we may solve the n linear systems
Ax1 = e1 , . . . , Axn = en simultaneously by considering the n × 2n
augmented matrix
[A|e1 · · · en ] = [A | I]
and transform A to its RCF, namely to I, by EROs.

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 05


Gauss-Jordan Method

1 Thus if [A | I] is transformed to [I | X], then X is the inverse of A.


2 Remark: To carry out the above process, we need not know
beforehand that the matrix A is invertible.
3 This follows by noting that A can be transformed to the identity
matrix by EROs if and only if A is invertible.
4 Hence the process itself reveals whether A is invertible or not.
5 Example: Let  
−1 1 2
A :=  3 −1 1 .
−1 3 4

6 We use EROs to transform [A | I] to [I | X], where X ∈ R3×3 .

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 05


   
−1 1 2 1 0 0 −1 1 2 1 0 0
3 −1 1 0 1 0 →  0 2 7 3 1 0 →
−1 3 4 0 0 1 0 2 2 −1 0 1

  
−1 1 2 1 0 0

1 −1 −2 −1 0 0
 0 2 7 3 1 0→0 1 3.5 1.5 0.5 0 
 

0 0 −5 −4 −1 1 0 0 1 0.8 0.2 −0.2


 
1 −1 0 0.6 0.4 −0.4
−→ 0 1 0 −1.3 −0.2 0.7  −→
0 0 1 0.8 0.2 −0.2
 
1 0 0 −0.7 0.2 0.3
0 1 0 −1.3 −0.2 0.7  = [I | X].
0 0 1 0.8 0.2 −0.2
Thus A is invertible and
 
−7 2 3
1 
A−1 =X= −13 −2 7 .
10
8 2 −2

Prof. Shreedevi K. Masuti, IIT Dharwad Linear Algebra: Lecture 05

You might also like