Scribd
Upload
67 views

The World of Eigenvalues-Eigenfunctions: Eigenfunction Eigenvalue

Eigenvalues and eigenfunctions are fundamental concepts in linear algebra. An eigenfunction of a linear operator A is a non-zero function that when operated on by A simply scales the function by a constant factor called the eigenvalue. The eigenvalues of A are the roots of its characteristic polynomial. Eigenvalues and eigenfunctions provide useful information about the properties of linear operators and matrices.

Uploaded by

Raj Kumar
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
67 views

The World of Eigenvalues-Eigenfunctions: Eigenfunction Eigenvalue

Eigenvalues and eigenfunctions are fundamental concepts in linear algebra. An eigenfunction of a linear operator A is a non-zero function that when operated on by A simply scales the function by a constant factor called the eigenvalue. The eigenvalues of A are the roots of its characteristic polynomial. Eigenvalues and eigenfunctions provide useful information about the properties of linear operators and matrices.

Uploaded by

Raj Kumar
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
You are on page 1/ 9

The world of Eigenvalues-eigenfunctions An operator function.

operates on a function and produces a

For every operator, there is a set of functions which when operated by the operator produces the same function modified only multiplied by a constant factor. Such a function is called the eigenfunction of the operator, and the constant modifier is called its corresponding eigenvalue. An eigenvalue is just a number: Real or complex. A typical eigenvalue equation would look like
Ax = x

Here, the matrix or the operator A operates on a vector (or a function) x producing an amplified or reduced vector x . Here the eigenvalue belongs to eigenfunction x . Suppose the operator is A = ( x dx ) . d n n n x n produces Ax = x x = nx . dx
d

operating on

Therefore, the operator A has an eigenvalue n corresponding to eigenfunction x n . 1. Eigenfunctions are not unique. Suppose Ax = x . Define, another vector z = cx , where c is a constant. Now, Az = Acx = cAx = c x = cx = z Therefore, z is also an e-function (eigenfunction) of A.
2.

If Ax = x is an eigenvalue equation (and we assume that x is not a zero vector), then


Ax = x (A - I)x = 0 det(A - I) = 0

This leads to a characteristic polynomial in :

3.

is an e-value of

p A = det( A I )

only if
A

pA = 0.

Spectrum of an operator e-values.

is ( A ) : set of all its

4.

Spectral radius of an operator ( A ) = max | | | i | = 1max ( A ) i n

is

5. Computation of spectrum and spectral radius:

be the matrix and we want to Let A = 2 5 compute its eigenvalues and eigenfunctions. Its characteristic equation (CE) is: 2 det 2 1 = 0 (2 - )(5 - ) + 2 = 0 5

This gives 2 7 + 12 = 0 Therefore,


A

( 3 )( 4 ) = 0

has two eigenvalues: 3 and 4.


x x = 1 x2

Let the eigenfunction be the vector corresponding to e-value 3. Then


2 1 x1 x1 3x1 2 5 x = 3 x = 3 x 2 2 2

Therefore, we have 2 x1 x2 = 3x1 yielding x1 = x2 . Also, we get 2 x1 + 5 x2 = 3 x2 which gives us no new result. Therefore, we can arbitrarily take the 1 corresponding to e-value 3 following solution: e1 = 1 for the matrix A.

Similarly, for e-value of 4, the eigenfunction appears 1 . to be e2 = 2 6. Faddeev-Leverrier Method to get characteristic polynomial. Define a sequence of matrices
1 P2 = A[ P 1 p1 I ] , p 2 = trace( P2 ) 2 1 P3 = A[ P2 p2 I ] , p3 = trace( P3 ) 3

P 1 = A, p1 = trace( P 1)

Then the characteristic polynomial


P( ) = ( 1 )n n p1n 1 p2 n 2 ... pn
6 12 A = 6 16 6 2 6 2 16

1 Pn = A[ Pn 1 pn 1I ] , pn = trace( Pn ) n

P( )

is

e.g.

Define
6 12 6 16 6 2 312 = 108 108

P 1 = A, p1 = trace( A ) = 12 + 16 + 16 = 44

P2 = A( P 1 p1 I ) =
6 32 2 6 16 6 108 408 60 6 28 2 6 2 28

108 60 , p 2 = 564 408

And one proceeds this way to get

p3 = 1728

The CA polynomial = ( 1 )3 [3 442 + 564 1728] The eigenvalues are next found solving

[3 442 + 564 1728] = 0

7. More facts about eigenvalues. Assume Ax = x . Therefore, A with eigenvector x .

is the eigenvalue of

a. A1 has the same eigenvector as A and the corresponding eigenvalue is 1 . b. An has the same eigenvector as eigenvalue n .
A

with the
A

c. ( A + I ) has the same eigenvector as eigenvalue ( + ) . d. If


A

with the

is symmetric, all its eigenvalues are real.


P 1 AP

e. If P is an invertible matrix then same eigenvalues as A . Proof of e.

has the

Suppose, the eigenfunction of eigenvalue k . Then,


1 P APy = ky

P 1 AP

is

with

Therefore, Py = x and k must be equal to . Therefore the eigenvalues of A and P 1 AP are identical and the eigenvector of one is a linear mapping of the other one. If the eigenvalues of A , 1 ,2 ,...,n are all distinct then there exists a similarity transformation such that
1 0 0 2 1 P AP = D = 0 0 .. .. 0 0 0 0 3 .. 0 .. 0 .. 0 .. 0 .. 0 .. n
x ( 1 ) , x ( 2 ) ,..., x ( i ) ,...x ( n )

APy = Pky = kPy

Let the eigenvectors of A be such that we have Ax( i ) = i x( i )

Then the matrix P = [ x( 1 ) , x( 2 ) ,..., x( n ) ] Then AP = [ Ax( 1 ) , Ax( 2 ) ,..., Ax( n ) ]


= 1 x( 1 ) ,2 x( 2 ) ,..., n x( n )

= PD

= x( 1 ) , x( 2 ) ,..., x( n ) 1e( 1 ) , 2 e( 2 ) ,..., n e( n )

][

Therefore,

P 1 AP = D

Also, note the following. If

is symmetric, then

. So, we can normalize each x eigenvector and obtain u = x so that the


( x( i ) )t x(
j)

= 0 , i j

(i)

(i) (i)

matrix Q = [u ( 1 ) ,u ( 2 ) ,...,u( n ) ] would be an orthogonal matrix. i.e. Q AQ = D


t

Matrix-norm. Computationally, the determined as


l 2 -norm l 2 -norm

of a matrix is

of
0 1 2 1 2 1

A =|| A ||2 = ( At A )

e.g.

1 A = 1 1

1 2 1

]1 / 2

Then

1 A A = 1 0
t

1 1 1 1 2 1

1 2 1

0 3 1 = 2 2 1

2 6 4

1 4 5

The eigenvalues are:


1 = 0, 2 = 7 + 7 , 3 = 7 7

Therefore,
1 A =1 1

A2 =

( At A ) = 7 + 7 3.106

The l norm is defined as e.g.


1 2 1 0 1 4

A = max aij 1 i n
j

j =1

a1 j = 1 + 1 + 0 = 2 ,
3

j =1

a2 j = 1 + 2 + 1 = 4

j =1

a3 j = 6

Therefore,

A = max( 2 ,4 ,6 ) = 6

In computational matrix algebra, we would often be interested about situations when A k becomes small (all the entrees become almost zero). In that case, A is considered convergent. i.e.
A

is convergent if Is
1 A = 2 1 4

lim A k

( )ij = 0

Example.
1 0 A2 = 4 1 1 , 4 4

0 1 2

convergent?

1 1 0 0 A3 = 8 A 4 = 16 3 1 , 1 1 , 16 8 8 16

It appears that
1 0 k Ak = 2 k 1 2k + 1 2k

In the limit matrix.

k ,

1 2k

. Therefore,

is a convergent

Note the following equivalent results: a. A is a convergent matrix Ak = 0 b1. klim 2


A =0 b2. klim c. ( A ) < 1 A k x = 0 x d. klim
k

Condition number is computed as


K ( A ) = A . A -1

K( A )

of a non-singular matrix

A matrix is well-behaved if its condition number is close to 1. When K ( A ) of a matrix A is significantly larger than 1, we call it an ill-behaved matrix.

You might also like