Fourier Series and Sturm-Liouville Eigenvalue Problems: Y. K. Goh

Fourier Series and Sturm-Liouville
Eigenvalue Problems
Y. K. Goh
2009
Y. K. Goh
Fourier Series and Sturm-Liouville Eigenvalue Problems
Outline
Functions
Fourier Series Representation
Half-range Expansion
Convergence of Fourier Series
Parsevals Theorem and Mean Square Error
Complex Form of Fourier Series
Inner Products
Orthogonal Functions
Self-adjoint Operators
Sturm-Liouville Eigenvalue Probelms

Y. K. Goh
Periodic Functions
Denition (Periodic Function)
A 2L-periodic function f : R R {} is a function
such that there exists a constant L > 0 such that
f(x) = f(x + 2L), x R. (1)
Here 2L is called the fundamental period or just period.
For example, f(x) = sin(x) is a 2-periodic funtion with
period 2, since sin(x + 2) = sin(x).
Y. K. Goh
Even and Odd Functions
Denition (Even and Odd Functions)
A function f is even if and only if f(x) = f(x), x.
A function f is odd if and only if f(x) = f(x), x.

For example
sin x is an odd function since sin(x) = sin x.
cos x is an even function since cos(x) = cos x.

Note that even function is symmetric about the y-axis. On the
other hand, odd function is symmetric about the origin.
Y. K. Goh
Examples of Even and Odd Functions
Y. K. Goh
Piecewise Continuous Functions
Denition (Piecewise Continuous Functions)
A function f is said to be piecewise continuous on the interval
[a, b] if
1. f(a+) and f(b) exist, and
2. f is dened and continous on (a, b) except at a nite
number of points in (a, b) where the left and right limits
exits
Y. K. Goh
Piecewise Smooth Functions
Denition (Piecewise Smooth Functions)
A function f, dened on the interval [a, b], is said to be
piecewise smooth if f and f
are piecewise continous on [a, b].

Thus f is piecewise smooth if
1. f is piecewise continous on [a, b],
2. f
exists and is continous in (a, b) except possibly at

nitely many points c where the one-sided limits
lim
xc
f
(x) and lim

xc
+ f
(x) exist. Furthermore,

lim
xa
+ f
(x) and lim

xb
f
(x) exist.
Y. K. Goh
Examples of Piecewise Functions
Figure: A piecewise smooth
function.
Figure: Another piecewise
smooth function.
Y. K. Goh
Some useful integrations
If f is even. Then, for any a R

_
a
a
f(x) dx = 2
_
a
0
f(x) dx.
If f is odd. Then, for any a R

_
a
a
f(x) dx = 0.
If f is piecewise continuous and 2L-periodic. Then, for

any a R
_
2L
0
f(x) dx =
_
a+2L
a
f(x) dx.
Y. K. Goh
Denition (Orthogonal Functions)
Two functions f and g are said to be orthogonal in the
interval [a, b] if
_
b
a
f(x)g(x) dx = 0. (2)
We will come back to orthogonal functions again later.
Y. K. Goh
Orthogonal Properties of Trigonometric Functions
The orthogonal properties of sine and cosine functions are
summarised as follow:
_

cos mx sin nx dx = 0, (3)

_

cos mx cos nx dx =
mn
, (4)
_

sin mx sin nx dx =
mn
(5)
where
mn
is the Kroneckers delta.
Y. K. Goh
Kroneckers Delta
Denition (Kroneckers Delta)
mn
=
_
1, m = n
0, m = n.
(6)
Y. K. Goh
Fourier Series
Theorem (Fourier Series Representation)
Suppose f is a 2L-periodic piecewise smooth function, then
Fourier series of f is given by
f(x) =
a
0
2
+
n=1
(a
n
cos nx +b
n
sin nx) (7)
and the Fourier series converges to f(x) if f is continuous at
x and to
1
2
[f(x+) +f(x)] otherwise.
Here =
2
2L
is called the fundamental frequency, while the
amplitudes a
0
, a
n
, and b
n
are called Fourier coecients of f
and they are given by the Euler formula.
Y. K. Goh
Euler Formula
Denition (Euler Formula)
The Fourier coecients for a 2L-periodic function f are given
by
a
0
=
1
L
_
L
L
f(x) dx,
a
n
=
1
L
_
L
L
f(x) cos nx dx =
1
L
_
L
L
f(x) cos
nx
L
dx,
b
n
=
1
L
_
L
L
f(x) sin nx dx =
1
L
_
L
L
f(x) sin
nx
L
dx.
for n = 1, 2, . . . .
Y. K. Goh
Collorary
If f is even and 2L-periodic, then the Fourier series

representation is
f(x) =
a
0
2
+
n=1
a
n
cos nx.
If f is odd and 2L-periodic, then the Fourier series

representation is
f(x) =
n=1
b
n
sin nx.
Here, a
0
, a
n
, and b
n
are given by the Euler formula.
Y. K. Goh
Examples of Fourier Series
(Odd function, digital impulses) Find the Fourier

representation of the periodic function f(x) with period
2, where
f(x) =
_
1, < x < 0,
1, 0 < x < .
[Answer:]
f(x) =
n odd
4 sin nx
n
=
k=1
4 sin(2k 1)x
(2k 1)
.
Y. K. Goh
Graphs for Digital Pulse Train and its Fourier Series
Y. K. Goh
(Even function) Find the Fourier series for f(x) = |x| if

1 < x < 1 and f(x + 2) = f(x).
[Answer:]
f(x) =
1
2

n=1
4 cos(2n 1)x
(2n 1)
2
2
.
Y. K. Goh
Graphs for f(x) = |x|, f(x + 2) = f(x)
Y. K. Goh
Find the Fourier series of the 2-periodic function

f(x) = x
3
+ if 1 < x < 1.
[Answer:]
f(x) = +
n=1
(1)
n
12 2n
2
2
n
3
3
sin nx.
Y. K. Goh
Graphs for f(x) = x
3
+ , f(x + 2) = f(x)
Y. K. Goh
Full-range Extensions
Consider a function f that is only dened in the interval [0, p).
We could always extension the function outside the range to
produce a new function. Of course, we have innite many
ways to extend the function, but here we will focus only on
three specic extensions.
Denition (Full-range Periodic Extension)
The full-range periodic extension g of a function f dened in
[0, p) is a p-periodic function given by
g(x) = f(x) if 0 x < p,
g(x) = g(x +p) if x < 0,
g(x) = g(x p) if x p.
Y. K. Goh
Half-range Periodic Extensions
Denition (Half-range Even Periodic Extension)
The half-range even periodic extension f
e
of a function f
dened in [0, p) is a 2p-periodic even function given by
f
e
(x) =
_
f(x) 0 x < p,
f(x) p x < 0.
Y. K. Goh
Half-range Periodic Extensions
Denition (Half-range Odd Periodic Extension)
The half-range odd periodic extension f
o
of a function f
dened in [0, p) is a 2p-periodic odd function given by
f
e
(x) =
_
f(x) 0 x < p,
f(x) p x < 0.
Y. K. Goh
Examples Periodic Extensions
Y. K. Goh
Full-range Fourier Series for f dened on [0, p)
Theorem
If f(x) is a piecewise smooth function dened on an interval
[0, p), then f has a full-range Fourier series expansion
f(x) =
a
0
2
+
n=1
(a
n
cos nx +b
n
sin nx) , 0 x < p, (8)
where =
2
p
and the Fourier coecients
a
0
=
1
p/2
_
p
0
f(x) dx, a
n
=
1
p/2
_
p
0
f(x) cos nx dx, and
b
n
=
1
p/2
_
p
0
f(x) sin nx dx.
Y. K. Goh
Fourier Cosine Series for f dened on [0, p)
Theorem
[0, p), then f has a half-range Fourier cosine series expansion
f(x) =
a
0
2
+
n=1
a
n
cos nx, 0 x < p, (9)
where =

p
and the Fourier coecients a
0
=
2
p
_
p
0
f(x) dx,
and a
n
=
2
p
_
p
0
f(x) cos nx dx.
Y. K. Goh
Fourier Since Series for f dened on [0, p)
Theorem
[0, p), then f has a half-range Fourier sine series expansion
f(x) =
n=1
b
n
sin nx, 0 x < p, (10)
where =

p
and the Fourier coecients
b
n
=
2
p
_
p
0
f(x) sin nx dx.
Y. K. Goh
Examples
Consider a signal f(t) = t measured from an experiment over
the duration given by 0 <= t < 4.
Sketch the full-range periodic extension of f(t). Find the

corresponding Fourier expansion of f(t).
Sketch the half-range even extension of f(t) and nd the

corresponding Fourier cosine expansion of f(t).
Sketch the half-range odd extension of f(t) and nd the

corresponding Fourier sine expansion of f(t).
Y. K. Goh
Convergence of Fourier Series
In the Fourier Series Representation Theorem, we were saying
that for every 2L-periodic piecewise smooth function f, we
could construct a partial sum
s
N
(x) =
a
2
+
N
n=1
(a
n
cos nx +b
n
sin nx) .
And, when N , the partial sum s
N
(x) converges to
f(x), if f(x) is continuous for all x;
1
2
[f(x+) +f(x)], at the discontinuous points, or jumps.
Y. K. Goh
Figure: s
N
(x) converges to f(x),
except at the jumps.
Figure: s
N
(x) converges
uniformly on the interval [-1, 1]
Y. K. Goh
Pointwise Convergence
Denition (Pointwise Convergence)
A sequence of functions {s
n
} is said to converge pointwise to
the function f on the set E, if the sequence of numbers
{s
n
(x)} converges to the number f(x), for each x in E.
Or,
if x E, s
n
(x) f(x), then {s
n
} converge pointwise to f.
Y. K. Goh
Uniform Convergence
Denition (Uniform Convergence)
We say that s
n
converges to f uniformly on a set E, and we
write s
n
f uniformly on E if, given > 0, we can nd a
positive integer N such that for all n N
|s
n
(x) f(x)| < , x E.
Denition (Uniform Convergence Series)
A series s(x) =
k=0
u
k
(x) is said to converge uniformly to
f(x) on a set E if the sequence of partial sums
s
n
(x) =
n
k=0
u
k
(x) converges uniformly to f(x).
Y. K. Goh
Note that if a sequence of partial sums s
n
converges uniformly
to f, then s
n
is also pointwise convergence. However, the
converse is not always true.
In order to determine is a s
n
is uniformly convergence, we use
the WeierstraM-test.
Y. K. Goh
WeierstraM-Test
Theorem (WeierstraM-Test)
Let {u
k
}
k=0
be a sequence of real- or complex-valued functions
on E. If there exists a sequence {M
k
}
k=0
of nonnegative real
numbers such that the following two conditions hold:
|u
k
(x)| M
k
, x E, and
k=0
M
k
< .
Then
k=0
u
k
(x) converges uniformly on E.
Y. K. Goh
Gibbs Phenomena
Here is an example of non-uniform convergence. The peaks
remain same height but the width of the peaks changes.
-1.5
-1
-0.5
0
0.5
1
1.5
-1 -0.5 0 0.5 1
N=10
sqr(x)
S10(x)
-1.5
-1
-0.5
0
0.5
1
1.5
-1 -0.5 0 0.5 1
N=30
sqr(x)
S30(x)
-1.5
-1
-0.5
0
0.5
1
1.5
-1 -0.5 0 0.5 1
N=50
sqr(x)
S50(x)
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
-1 -0.5 0 0.5 1
N=10
err(x)
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
-1 -0.5 0 0.5 1
N=30
err(x)
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
-1 -0.5 0 0.5 1
N=50
err(x)
Y. K. Goh
Mean Square Error
Since s
N
converges to f only when N , for most
practical purposes, we need N to be large but nite. Thus, we
are approximating f with s
N
, and it is important for us to
keep track of the error of the approximation.
Denition (Mean Square Error)
The mean square error of the partial sum s
N
relative to f is
E
N
=
1
2L
_
L
L
[f(x) s
N
(x)]
2
dx
=
1
2L
_
L
L
[f(x)]
2
dx
1
4
a
2
0
1
2
N
n=1
_
a
2
n
+b
2
n
_
.
Y. K. Goh
Mean Square Approximation
Theorem (Mean Square Approximation)
Suppose that f is square integrable, i.e.
_
L
L
|f(x)|
2
dx on
[L, L]. Then s
N
, the Nth partial sum of the Fourier series of
f, approximates f in the mean square sense with an error E
N
that decreases to zero as N .
lim
N
E
N
=
1
2L
_
L
L
[f(x)]
2
dx
1
4
a
2
0
1
2
N
n=1
_
a
2
n
+b
2
n
_
= 0.
Y. K. Goh
Bessels Inequality and Parsevals Identity
Since E
N
> 0, from the denition of E
N
, we get
Denition (Bessels Inequality)
1
4
a
2
0
+
1
2
N
n=1
(a
2
n
+b
2
n
)
1
2L
_
L
L
[f(x)]
2
dx.
A stronger result is when taking the limit N
Denition (Parsevals Identity)
1
4
a
2
0
+
1
2
n=1
(a
2
n
+b
2
n
) =
1
2L
_
L
L
[f(x)]
2
dx.
Y. K. Goh
Multiplication Theorem
A generalisation of the Parsevals identity is the multiplication
theorem.
Theorem (Multiplication (Inner Product) Theorem)
If f and g are two 2L-periodic piecewise smooth functions
1
2L
_
L
L
f(x)g(x) dx =
n=
c
n
d
n
where c
n
and d
n
are the Fourier coecients for the complex
Fourier series of f and g respectively.
Y. K. Goh
Complex Form of Fourier Series
Theorem (Complex Form of Fourier Series)
Let f be a 2L-periodic piecewise smooth function. The
complex form of the Fourier series of f is
n=
c
n
e
inx
,
where the frequency = 2/2L and the Fourier coecients
c
n
=
1
2L
_
L
L
f(x)e
inx
dx, n = 0, 1, 2, . . . .
For all x, the complex Fourier series converges to f(x) if f is
continuous at x, and to
1
2
[f(x+) +f(x)] otherwise.
Y. K. Goh
Relations of Complex and Real Fourier Coecients
c
n
= c
n
;
c
0
=
1
2
a
0
c
n
=
1
2
(a
n
ib
n
);
c
n
=
1
2
(a
n
+ib
n
).
a
0
= 2c
0
;
a
n
= c
n
+c
n
;
b
n
= i(c
n
c
n
).
Y. K. Goh
Complex Form of Parsevals Identity
Theorem (Complex Form of Parsevals Identity)
Suppose f is a square integrable 2L-periodic piecewise smooth
function on [L, L]. Then
1
2L
_
L
L
[f(x)]
2
dx =
n=
c
n
c
n
=
n=
|c
n
|
2
where c
n
is the complex Fourier coecients of f.
Y. K. Goh
Example
Find the complex Fourier series for the 2-periodic function
f(x) = e
x
dened in (, ).
Y. K. Goh
Frequency Spectra
The distribution of the magnitude of complex Fourier
coecients |c
n
| in frequency domain is called the amplitude
spectrum of f.
The distribution of p
0
= |c
0
|
2
and p
n
= |c
n
|
2
in frequency
domain is called the power spectrum of f.
Y. K. Goh
Inner Products
Denition (Inner Products)
Let and be (possibly complex) functions of x on the
interval (a, b). Then the inner product of and is
| =
_
b
a
(x)(x) dx.
Note that the notation of inner product in some books is
(, ) =
_
b
a
(x)(x) dx.
Please take note on the order of and in the brackets.
Y. K. Goh
Norm
Denition (Norm)
Let f be (possibly complex) function of x on the interval
(a, b). Then the norm of f is
|||| =
_
f|f =
__
b
a
|f(x)|
2
dx
_
1/2
.
Y. K. Goh
Denition (Orthogonal Functions)
The functions f and g are called orthogonal on the interval
(a, b) if their inner product is zero,
f|g =
_
b
a
f
(x)g(x) dx = 0.
Denition (Orthogonal Set of Functions)
A set of functions {F
1
, F
2
, F
3
, . . . } dened on the interval
(a, b) is called an orthogonal set if
||F
n
|| = 0 for all n; and
F
m
|F
n
= 0, for m = n.
Y. K. Goh
Normalisation and Orthonormal Set
Denition (Normalisation)
A normalised function f
n
for a function F
n
, with ||F
n
|| = 0, is
dened as
f
n
(x) =
F
n
(x)
||F
n
||
.
Denition (Orthonormal Set of Functions)
A set of functions {f
1
, f
2
, f
3
, . . . } dened on the interval
(a, b) is called an orthonormal set if
||f
n
|| = 1 for all n; and
f
m
|f
n
= 0, for m = n;
or simply, f
m
|f
n
=
mn
, where
mn
is the Kroneckers delta.
Y. K. Goh
Generalized Fourier Series
Theorem (Generalized Fourier Series)
If {f
1
, f
2
, f
3
, . . . } is a complete set of orthogonal functions on
(a, b) and if f can be represented as a linear combination of
f
n
, then the generalised Fourier series of f is given by
f(x) =
n=1
a
n
f
n
(x) =
n=1
f
n
|f
||f
n
||
2
f
n
(x),
where a
n
=
f
n
|f
||f
n
||
2
is the generalised Fourier coecient.
Y. K. Goh
Parsevals Identity
Theorem (Generalized Parsevals Identity)
If {f
1
, f
2
, f
3
, . . . } is a complete set of orthogonal functions on
(a, b) and let f be such that ||f|| is nite. Then
_
b
a
|f(x)|
2
dx =
n=1
|f
n
|f|
2
||f
n
||
2
.
Y. K. Goh
Orthogonality with respect to a Weight, w(x)
(Inner product)
f|g =
_
b
a
f
(x)g(x)w(x) dx
(Orthogonality)
f
m
|f
n
=
_
b
a
f
m
(x)f
n
(x)w(x) dx = ||f
m
||
2
mn
(Generalised Fourier Series)

f(x) =
n=1
f
n
|f
||f
n
||
2
f
n
(x)
(Generalised Parsevals Identity)

_
b
a
|f(x)|
2
w(x) dx =
n=1
|f
n
|f|
2
||f
n
||
2
Y. K. Goh
Adjoint and Self-adjoint Operators
Denition (Adjoints of Dierential Operators)
Suppose u and v are (possible complex) functions of x in (a, b)
and let L be a linear dierential operator. Then, the formal
adjoint M of L is another operator such that for all u and v
f|L[g] = M[f]|g.
Denition (Self-adjoint Operators)
Suppose M is the formal adjoint operator for a linear operator
L in space S. If M = L, then the operator L is said to be
formally self-adjoint or formally Hermitian.
Y. K. Goh
Example: Adjoint for L[] p(x)d/dx
Suppose L[] p(x)d/dx, then
u|L[v] =
_
b
a
u
_
p
d
dx
_
v dx = [u
pv]
b
a
_
b
a
v
_
d
dx
(p
u)
_
dx
= M[u]|v
In the last step, we set the boundary term to zero. The the
adjoint for the operator L consists of
Formal adjoint M[]

d
dx
(p
); and
Boundary conditions [u
pv]
b
a
= 0.
Furhermore, if p(x) is a pure imaginary constant, then
M = L. i.e. L is self-adjoint.
Y. K. Goh
Adjoint for 2
nd
Order Linear Dierential Operators
Suppose L[] a
0
(x)
+a
1
(x)
+a
2
(x). Then,
u|L[v] = [u
a
0
v
+u
a
1
v v(a
0
u
]
b
a
+
_
b
a
v [(a
0
u
(a
1
u
+a
2
u
] dx = M[u]|v with
appropriate choice of boundary conditions. Note that
M[] a
0
+ (2a
0
a
1
)
+ (a
2
a
1
+a
0
).
M can be made self-adjoint if a

1
= a
0
.
The self-adjoint operator S is

S[] =
d
dx
_
a
0
(x)
d
dx
_
+a
2
(x).
The neccessary boundary condition is

[a
0
u
a
0
v(u
]
b
a
= 0.
Y. K. Goh
Eigenvalue Problems & Sturm-Liouville Equation
Denition (Eigenvalue Problem)
The eigenvalue problem associated to a dierential operator L
is the equation Ly +y = 0, where is called the eigenvalue,
and y is called the eigenfunction.
It is possible to nd a weight factor w(x) > 0 for L such that
S[y] w(x)L[y] is self-adjoint. The resulting eigenvalue
equation is called the Sturm-Liouville Equation
Denition (Sturm-Liouville Equation)
[S +w(x)] y =
d
dx
_
a
0
(x)
dy
dx
_
+a
2
(x)y +w(x)y = 0 for
a < x < b.
Y. K. Goh
Regular Sturm-Liouville Problems
Denition (Regular Sturm-Liouville Problem)
A regular SL problem is a boundary value problem on a closed
nite interval [a, b] of the form
d
dx
_
a
0
(x)
dy
dx
_
+a
2
(x)y +w(x)y = 0, a < x < b,
satisfying regularity conditions and boundary conditions
c
1
y(a) +c
2
y
(a) = 0, d
1
y(b) +d
2
y
(b) = 0,
where at least one of c
1
and c
2
and at least one of d
1
and d
2
are non-zero.
Y. K. Goh
Singular Sturm-Liouville Problems
Denition (Regularity Conditions)
The regularity conditions of a regular SL problem are
a
0
(x), a
0
(x), a
2
(x) and w(x) are continuous in [a, b];
a
0
(x) > 0 and w(x) > 0.
Denition (Singular Sturm-Liouville Problem)
A singular SL problem is a boundary value problem consists of
Sturm-Liouville equation, but either
fails the regularity conditions; or
innite boundary conditions; or
one or more of the coecients become singular.

Y. K. Goh
Solutions of SL Problems
A trivial (not useful) solution to the SL problem is y = 0.
Other non-trivial solutions would be the eigenfunctions

y
m
, and for each of these eigenfunctions there is a
corresponding eigenvalue
m
.
There are innite many of these eigenfunctions, and the

set of eigenfunctions {y
1
, y
2
, . . . , y
m
, . . . } forms a
complete orthogonal set of functions that span the
innite dimensional Hilbert space.
Any function f in the Hilbert space can be expressed as a

linear combination of the eigenfunctions,
f(x) =
n=1
a
n
y
n
(x).
Y. K. Goh
Eigenvalues of Sturm-Liouville Problems
Theorem (Sturm-Liouville Problem)
The eigenvalues and eigenfunctions of a SL problem has the
properties of
All eigenvalues are real and compose a countably innite

collections satisfying
1
<
2
<
3
< . . . where
j

as j .
To each eigenvalue
j
there corresponds only to one
independent eigenfunction y
j
(x).
The eigenfunctions y
j
(x), j = 1, 2, . . . , compose a
complete orthogonal set with appropriate to the weight
functions w(x) in doubly-integrable functions space
L
2
(a, b).
Y. K. Goh
Eigenfunction Expansions
Theorem (Eigenfunction Expansions)
If f L
2
(a, b) then eigenfunction expansion of f on
{y
1
, y
2
, . . . } is
f(x) =
n=1
A
n
y
n
, a < x < b,
where
A
n
=
y
n
|f
||y
n
||
2
=
_
b
a
y
n
(x)f(x)w(x) dx/
_
b
a
|y
n
(x)|
2
w(x) dx.
Y. K. Goh
Example SL Problem: Harmonic Equation
An example of SL equation is the Harmonic Equation with
boundary condition.
ODE : y
+y = 0, 0 < x < L.
Dirichlet Boundary condition: y(0) = y(L) = 0.
Eigenvalues:
n
= k
2
n
=
n
2
2
L
2
, n = 1, 2, . . . ;
Eigenfunctions: y
n
(x) = sin
_
n
L
x
_
for n = 1, 2, . . . ;
Eigenfunction expansion: f(x) =
n=1
b
n
sin
n
L
x.
Y. K. Goh
Again, the Harmonic Equation with another boundary
condition.
ODE : y
+y = 0, 0 < x < L.
Neumann Boundary condition: y
(0) = y
(L) = 0.
Eigenvalues:
n
= k
2
n
=
n
2
2
L
2
;
Eigenfunctions: y
n
(x) = cos
_
n
L
x
_
for n = 0, 1, 2, . . . ;

a
0
2
+
n=1
a
n
cos
n
L
x.
Y. K. Goh
The Harmonic Equation with periodic boundary condition.
ODE: y
+y = 0, 0 < x < 2.
Periodic boundary condition: y(0) = y(2).
Eigenvalues:
n
= k
2
n
= n
2
;
Eigenfunctions: y
n
(x) = e
inx
for n = 0, 1, 2, . . . ;
n=
c
n
e
inx
.
Y. K. Goh
Example SL Problem: (Parametric) Bessel
Equation
Bessel Equation. x
2
y
+xy
+ (
2
x
2
2
)y = 0 or
[x
2
y
+ (
2
x

2
x
)y = 0 in the interval 0 < x < L.
Since a
0
(x) = x
2
is zero at x = 0, = singular SL
problem.
ODE: x
2
y
+xy
+ (
2
x
2
2
)y = 0;
Boundary conditions: y(x) is bounded, and y(L) = 0;
Eigenvalues:
2
=
2
n
=

n
L
, n = 1, 2, . . . , where
n
is
the nth-root of J
, i.e. J
(
n
) = 0;
Eigenfunctions: y
n
(x) = J
(
n
x), n = 1, 2, . . . ;
Weight: w(x) = x;
n=1
a
n
J
(
n
x).
Y. K. Goh
Example SL Problem: Spherical Bessel Equation
Bessel Equation. x
2
y
+xy
+ (
2
x
2
n(n + 1))y = 0 in the
interval 0 < x < .
Since a
0
(x) = x
2
is zero at x = 0, = singular SL
problem.
ODE: x
2
y
+xy
+ (
2
x
2
( + 1))y = 0;
Boundary conditions: y(x) is bounded, and y(L) = 0;
Eigenvalues:
2
=
2
n
=

n
L
, n = 1, 2, . . . , where
n
is
the nth-root of j
, i.e. j
(
n
) = 0;
Eigenfunctions: y
n
(x) = j
(
n
x), n = 1, 2, . . . ;
Weight: w(x) = x;
n=1
a
n
j
(
n
x).
Y. K. Goh
Example SL Problem: Legendre Equation
Legendre Equation. (1 x
2
)y
2xy
+n(n + 1)y = 0 or
[(1 x
2
)y
+n(n + 1)y = 0, in the interval 1 < x < 1.
a
0
(x) = (1 x
2
) and a
0
(1) = 0 = singular SL
problem.
ODE: (1 x
2
)y
2xy
+y = 0;
Boundary conditions: y(x) is bounded at x = 1;
Eigenvalues: =
n
= n(n + 1), n = 0, 1, 2, . . . ;
Eigenfunctions: y
n
(x) = P
n
(x), n = 0, 1, 2, . . . ;
n=0
a
n
P
n
(x).
Y. K. Goh

Fourier Series and Sturm-Liouville Eigenvalue Problems: Y. K. Goh

Uploaded by

Copyright:

Available Formats

Fourier Series and Sturm-Liouville Eigenvalue Problems: Y. K. Goh

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fourier Series and Sturm-Liouville Eigenvalue Problems: Y. K. Goh

Uploaded by

Copyright:

Available Formats

Fourier Series and Sturm-Liouville

Fourier Series Representation

Convergence of Fourier Series

Parsevals Theorem and Mean Square Error

Complex Form of Fourier Series

Sturm-Liouville Eigenvalue Probelms

A function f is even if and only if f(x) = f(x), x.

A function f is odd if and only if f(x) = f(x), x.

sin x is an odd function since sin(x) = sin x.

cos x is an even function since cos(x) = cos x.

are piecewise continous on [a, b].

exists and is continous in (a, b) except possibly at

(x) and lim

(x) exist. Furthermore,

(x) and lim

If f is even. Then, for any a R

If f is odd. Then, for any a R

If f is piecewise continuous and 2L-periodic. Then, for

cos mx sin nx dx = 0, (3)

If f is even and 2L-periodic, then the Fourier series

If f is odd and 2L-periodic, then the Fourier series

(Odd function, digital impulses) Find the Fourier

(Even function) Find the Fourier series for f(x) = |x| if

Find the Fourier series of the 2-periodic function

Sketch the full-range periodic extension of f(t). Find the

Sketch the half-range even extension of f(t) and nd the

Sketch the half-range odd extension of f(t) and nd the

f(x), if f(x) is continuous for all x;

(Generalised Fourier Series)

(Generalised Parsevals Identity)

Formal adjoint M[]

M can be made self-adjoint if a

The self-adjoint operator S is

The neccessary boundary condition is

fails the regularity conditions; or

innite boundary conditions; or

one or more of the coecients become singular.

A trivial (not useful) solution to the SL problem is y = 0.

Other non-trivial solutions would be the eigenfunctions

There are innite many of these eigenfunctions, and the

Any function f in the Hilbert space can be expressed as a

All eigenvalues are real and compose a countably innite

Dirichlet Boundary condition: y(0) = y(L) = 0.

Eigenfunction expansion: f(x) =

Neumann Boundary condition: y

Eigenfunction expansion: f(x) =

Periodic boundary condition: y(0) = y(2).

Eigenfunction expansion: f(x) =

Boundary conditions: y(x) is bounded, and y(L) = 0;

Eigenfunction expansion: f(x) =

Boundary conditions: y(x) is bounded, and y(L) = 0;

Eigenfunction expansion: f(x) =

+n(n + 1)y = 0, in the interval 1 < x < 1.

Boundary conditions: y(x) is bounded at x = 1;

Eigenfunction expansion: f(x) =

You might also like