Chapter 12: Orthogonal Functions and Fourier Series

Orthogonal Functions
1, 2 1 2
( ) ( ) ( ) ( ) 0
b
a
f f f x f x w x dx = =
}
:
1, 2
f f

[ , ] a b
Orthogonal

ex)
2 3
1 2
( ) , ( ) f x x f x x = =
interval
[ 1,1]

1
1
1 2 3 6
1, 2
6
1 1
( ) 0 f f x x dx x

= = =
}
Orthogonal with w(x)=1
Orthogonal Sets
( , ) ( ) ( ) ( ) 0,
b
m n m n
a
x x w x dx m n | | | | = = =
}
, where
( ) w x
: weight function

2
2
( ) ( ) ( )
b
n n
a
x w x x dx | | =
}
and
2
( ) ( ) ( ) .
b
n n
a
x w x x dx | | =
}

If
{ ( )}
n
x |
,
2
( ) 1, 0, 1, 2,
n
x n | = =
Orthonormal Set
Example 1 Orthogonal Set of Functions
{ } { }
| |
0 0
cos
( , ) 1 ( ) ( ) cos
1
sin
1
sin sin( ) 0, 0
n
n n
nx
x x dx nxdx
nx
n
n n n
n
t t
t t
t
t
|
| | | |
t t

=
= =
=
= = =
} }


| |
cos
( , ) ( ) ( ) 1 cos cos
1

2
1 sin( ) sin( )
0,
( ) cos( ) trig ident
2
ity
m n m n
m
x x dx mx nxd
n x m n x
x
dx
m n x m n x
m n
m n m n
t t
t t
t
t
t
t
| | | |

= =
=
+
(
= + = =
(
+

+ +
} }
}

orthogonal [, ] with = 1
Norms
2
0
( ) 1 2 x dx
t
t
| t

= =
}

0
( ) 2 . x | t =

| |
2
2
1
( ) cos 1 cos 2
2
n
x nxdx nx dx
t t
t t
| t

= = + =
} }
( ) .
n
x | t =

1 cos cos 2
, , ,
2
x x
t t t

`
)
Orthonormal set
Why do we need this orthogonal set? Series expansion!
Weve learnt power series;
=
0
+
1
+
2

2
+ =
=0

()
!

Analogously, we can expand , , , an orthogonal set

{
0
,
1
,
2
, }
=
0

0
+
1

1
+ +

+ =
=0

()
In many cases, these expansions are found to be very useful!
Taylor series!
power
Orthogonal Series Expansion
0 0 1 1
0
( ) ( ) ( ) ( ) ( ) ?
n n n n
n
f x c x c x c x c x | | | |

=
= = + + + +

0 0 1 1
0 0 1 1
( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( )
( ) ( ) ( )
( , ) ( , ) ( , ) .
b b b
m m m
a a a
b
n n m
a
m m n n m
f x x w x dx c x x w x dx c x x w x dx
c x x w x dx
c c c
| | | | |
| |
| | | | | |
= + +
+ +
= + + + +
} } }
}


2
( ) ( ) ( ) ( ) ( ) .
b b
n n n
a a
f x x w x dx c x w x dx | | =
} }


2
2
( ) ( ) ( ) ( ) ( ) ( )
( )
(
, 0
) ( )
,1, 2, .
b b
n n
a a
n b
n
n
a
f x x w x dx f x x w x dx
c
x
x w x dx
n
| |
|
|
= = =
} }
}

2
0 0
( , )
( ) ( ) ( )
( )
n
n n n
n n
n
f
f x c x x
x
|
| |
|

= =
= =


Fourier Series
Introduction
0 1 2
{ ( ), ( ), ( ), } x x x | | |
Orthogonal Set
0 0 1 1
0
( ) ( ) ( ) ( ) ( )
n n n n
n
f x c x c x c x c x | | | |

=
= = + + + +

Trigonometric Series
2 3
1, cos , cos , cos ,
2 3
sin , sin , sin ,
x x x
p p p
x x x
p p p
t t t
t t t

`
)

`
)
Orthogonal with
( ) 1 w x =
on
[ , ] p p
(1)

0
1
( ) cos sin .
2
n n
n
a n n
f x a x b x
p p
t t

=
| |
= + +
|
\ .

(2)
0
1
( ) cos sin .
2
p p p p
n n
p p p p
n
a n n
f x dx dx a xdx b xdx
p p
t t

=
| |
= + +
|
\ .

} } } }
= 0 = 0
0 0
0
( )
2 2
p
p p
p p
p
a a
f x dx dx x pa

= = =
} }

0
1
( ) .
p
p
a f x dx
p

=
}
cos( / ) m x p t

0
1
( ) cos cos
2
cos cos cos sin .
p p
p p
p p
n n
p p
n
a m m
f x xdx xdx
p p
m n m n
a x xdx b x xdx
p p p p
t t
t t t t

=
=
| |
+ +
|
\ .
} }

} }

cos 0, 0, cos sin 0
p p
p p
m m n
xdx m x xdx
p p p
t t t

= > =
} }
0,
cos cos
, .
p
p
m n
m n
x xdx
p m n p p
t t

}
( ) cos
p
n
p
n
f x xdx a p
p
t

=
}
cos
2
=
1+cos 2
2

1
( ) cos .
p
n
p
n
a f x xdx
p p
t

=
}
sin( / ) m x p t

sin 0, 0, sin cos 0
p p
p p
m m n
xdx m x xdx
p p p
t t t

= > =
} }
0,
sin sin
, ,
p
p
m n
m n
x xdx
p m n p p
t t

}

1
( )sin .
p
n
p
n
b f x xdx
p p
t

=
}
Fourier Series
0
1
( ) cos sin ,
2
n n
n
a n n
f x a x b x
p p
t t

=
| |
= + +
|
\ .

0
1
( )
p
p
a f x dx
p

=
}
1
( ) cos
p
n
p
n
a f x xdx
p p
t

=
}
1
( )sin .
p
n
p
n
b f x xdx
p p
t

=
}
Example 1 Expansion in a Fourier Series
0, - 0
( ) , ( ) 1
, 0
x
f x p w x
x x
t
t
t t
< <

= = =

< <

2
0
0
0
0
0
0
0
0
0
2
1 1 1
( ) 0 ( )
2 2
1 1
( ) cos 0 ( ) cos
1 sin 1
( ) sin
1 cos

cos 1
cos ( 1)
n
n
x
a f x dx dx x dx x
a f x nxdx dx x nxdx
nx
x nxdx
n n
nx
n n
n
n
n
t
t t
t t
t t
t t
t
t
t
t
t
t t
t t t
t
t t
t
t
t
t
t


(
(
= = + = =
(
(


(
= = +
(

(
=

+
(
(

=

=
+
=
} } }
} } }
}
2
1 ( 1)
.
n
n t

=

0
1 1
( )sin .
n
b x nxdx
n
t
t
t
= =
}
=

4
+
=1

1 (1)

cos +
1

sin
Periodic Extension
2 3 2 3
1, cos , cos , cos , , sin , sin , sin , x x x x x x
p p p p p p
t t t t t t
`
)
(1)
( ) ( ) f x T f x + =

(0 ) (0 )
2 2
f f t + +
=
and
( ) ( )
0,
2
f f t t + +
=





Sequence of Partial Sums
1
( )
4
S x
t
=
2
2
( ) cos sin
4
S x x x
t
t
= + +
3
2 1
( ) cos sin sin 2
4 2
S x x x x
t
t
= + + +













Fourier Cosine and Sine Series
Review
Even Function:
( ) ( ) f x f x =

Odd Function:
( ) ( ) f x f x =

Even and Odd Functions
even integer
2
( ) f x x
+
=
:
2 2
( ) ( ) ( ) f x x x f x = = =

odd integer
3
( ) f x x
+
=
:
3 3
( ) ( ) ( ). f x x x f x = = =


cos( ) cos x x =
: Even Function
sin( ) sin x x =
: Odd Function
Properties
Properties of Even/Odd Functions
(a) Two even function even
(b) Two odd function even
(c) An even function and an odd function odd
(d) Sum (difference) of two even function even
(e) Sum (difference) of two odd function odd
(f)
( ) f x
is even
0
( ) 2 ( )
a a
a
f x dx f x dx

=
} }

(g)
( ) f x
is odd
( ) 0
a
a
f x dx

=
}


Proof of (b)
( ), ( ) f x g x
is odd
( ) ( ), ( ) ( ) f x f x g x g x = =

( ) ( ) ( ) F x f x g x

( ) ( ) ( ) ( ( ))( ( )) ( ) ( ) ( ). F x f x g x f x g x f x g x F x = = = =

Cosine and Sine Series
( ) f x
is
[ , ] p p
even
0
0
1 2
( ) ( )
p p
p
a f x dx f x dx
p p

= =
} }
0
1 2
( ) cos ( ) co
even
s
p p
n
p
n n
a f x x dx f x xdx
p p p p
t t

= =
} }

odd
1
( )sin 0
p
n
p
n
b f x x dx
p p
t

= =
}


( ) f x
is
[ , ] p p
odd
0
2
0, 0,1, 2, , ( )sin .
p
n n
n
a n b f x xdx
p p
t
= = =
}
Example Expansion in a Sine Series
1, 0
( )
1, 0
x
f x
x
t
t
< <

s <

( , ) t t
odd function Sine Series
0
2 2 1 ( 1)
(1)sin ,
n
n
b nxdx
n
t
t t

= =
}

1
2 1 ( 1)
( ) sin .
n
n
f x nx
n t

=

=

(7)
Sturm-Liouville Problem: An extension of orthogonality
| |
( ) ( ( ) ( )) 0
d
r x y q x p x y
dx
' + + =
in the interval | | , a b
(3)
| | ( ) & ( ) 0 in the interval , Assume r x p x a b >

1 1
( ) ( ) 0 A y a B y a ' + =
(4)
2 2
( ) ( ) 0 A y b B y b ' + =
(5) (Homogeneous Problem)

Example 1 Eigenvalues and Eigenfunctions
0, (0) 0, ( ) 0 y y y y L '' ' ' + = = =
(2)
0: sin cos , (0) 0 A 0 y A x B x y ' > = + = =

'( ) sin 0 y L B L L n t = = =


2 2 2
1
/ , 0,1, 2, 3, , cos( / )
n
n L n y c n x L t t = = =


B.Cs:
eigenvalue
( ) ( ) 1,
( ) 0
p x r x
q x
= =
=
) = 0 =
1
+
2
,
1
= 0,
2
= 0
From

0 =

= 0 trivial!
) < 0 : = cosh + sin A=B=0
From

0 =

= 0 trivial!
Similarly, in Sturm-Liouville problem,
we can set

, a eigenvalue set!
for each

, only one eigenfunction,

exists:
ex)

= cos

for

=

2

2
Each eigenfunction is linearly independent
This set of

eigenfunctions are orthogonal with = ()!

Properties of the Regular Sturm-Liouville Problem
for two eigenvalue
, ,
m n m n
y y
: eigenfunctions
| |
( ) ( ( ) ( )) 0
m m m
d
r x y q x p x y
dx
' + + =
(6)
| |
( ) ( ( ) ( )) 0
n n n
d
r x y q x p x y
dx
' + + =
(7)
(6) (7)
n m
y y

| | | |
( ) ( ) ( ) ( ) .
m n m n m n n m
d d
p x y y y r x y y r x y
dx dx
' ' =

Integrating x a = to x b =
| |
( ) ( )
= ( )
( ) ( ) ( ) ( ) ( ) (
b b b
b b
m n m n m n m n n m n m
a a
a a a
b
m n n m
a
m n n m
p x y y dx y r y ry y dx y r y ry y dx
r y y y y
r b y b y b y b y b r a

( (
' ' ' ' ' ' =
( (

' '
' ' =
} } }
| | ) ( ) ( ) ( ) ( )
m n n m
y a y a y a y a ' '
(8)
From B.Cs

+
1

= 0 (1)

+
1

= 0 (2)

1
=
0
0
if (1) & (2) linearly independent,
1
=
1
= 0 trivial
(= )
1
&
1
,
det

= 0

= 0

()

()

()

() = 0
Similarly,

+
2

= 0 (1)

+
2

= 0 (2)

= 0
(

= 0, if orthogonal
Example 2 A Regular Sturm-Liouville Problem
0, (0) 0, (1) (1) 0. y y y y y '' ' + = = + =
(10)
( ) 1, ( ) 1, ( ) 0 r x p x q x = = =

Solution
1) 0 =
Trivial Solution
2)
2
0 o = <
Trivial Solution
3)
2
0 o = >
2
0 y y o '' + =


1 2
cos sin y c x c x o o = +

1
0 c =
,
2 2
sin cos 0. c c o o o + =

tano o =
Then,

= sin

, where tan

This set is orthogonal!

Example 3 Bessel Equation: in cylindrical coordinate!
| |
2 2
2 2
'' ' 0 or ' 0,
n d n
xy y x y xy x y
x dx x
o o
| | | |
+ + = + =
| |
\ . \ .

1 2
( ) ( )
n n
y c J x c Y x o o = +


2
2
( ) , ( ) , , ( )
n
r x x p x x q x
x
o = = = =

(0) 0 (0) , only ( ) is bounded at 0!!
n n
r Y J x x o = =

A set { ( )} is orthogonal with respect to ( ) ( ) on an interval [0, ]!
n i
J x w x p x b o =

0
( ) ( ) 0, ,
b
n i n j i j
xJ x J x dx o o = =
}

'
2 2
( ) ( ) 0. * B.C.
n n
A J b B J b Homogeneous o o o + =

We do not need homogeneous B.C. at 0, because (0) 0! x r = =

Legendres Equation: frequently found in spherical coordinate
2
(1 ) " 2 ( 1) 0 x y xy n n y + + =


2
(1 ) ' ( 1) 0.
d
x y n n y
dx
( + + =

(21)
2
2
( ) 1 , ( ) 1, ( 1), ( ) 0
( ) 1 ( 1) (1) 0!
r x x p x n n q x
r x x r r
= = = + =
= = =


On the interval [-1,1],
1
1
( ) ( ) 0, .
m n
P x P x dx m n

= =
}
orthogonal with w(x)=1
Regardless of B.C.s at
1 & 1 x x = =

Fourier-Bessel Series
For an interval [0,b], B.C. at x=b,
'
2 2
( ) ( ) 0.
n n
A J b B J b o o o + =

1
( ) ( ),
i n i
i
f x c J x o

=
=

0
2
( ) ( )
.
( )
b
n i
i
n i
xJ x f x dx
c
J x
o
o
=
}

2
2
0
( ) ( ) .
b
n i n i
J x xJ x dx o o =
}


Square Norm
| | | | ( )
2
2 2 2 2
' 0 2 ' ' 2 ' 0
d n d
xy x y xy xy yy x n
dx x dx
o o
| |
+ = + =
|
\ .
by
2 ' xy

| | ( ) | |
2 2
2 2 2
' 0.
d d
xy x n y
dx dx
o + =

| | ( )
( )
2
2 2 2 2 2 2
0
0
2 '
b
b
xy dx xy x n y o o = +
}

( )| |
2
2
2 2 2 2 ' 2 2 2
0
2 ( ) ( ) ( ) ,
b
n n n
xJ x dx b J b b n J b o o o o o o ( = +
}

Now, check B.C. at = :
2
( ) +
2

( ) =
2

( ) +
2

( ) = 0
)
2
= 0,

( ) = 0

) = 0,

2
=

2
Since,

( ) =

( )
+1
( )

( ) =

( )
+1
( )
2
2

2
( ) = [

( )]
2
+ (
2

2

2
)[

( )]
2

Known value
=0
=0

+1
( )

2
(

) =

2
2

+1
2
(

)
)
2
= ,
2
= ,

( ) +

( ) = 0

( ) =

( )
2
2

0

2
( ) = [

( )]
2
+ (
2

2

2
)[

( )]
2

0

2
(

) =
1
2

2
(

2

2
+
2
)

2
(

) i=1,2,3
No 0!
=

( )
Ex) = , 0 < < 3 =
=1

1
(

),
1
( 3) = 0 . .

=
2
3
2

2
2
(3

)

0
3

) let t =

=
2
3
2

2
2
(3

)
1

3

0
3

1
( ) =
2
3
2

2
2
(3

)

2

2
( )
3

0
=
2

2
(3

1
( ) =

( )) , = 2
2

1
( ) =

(
2

2
( ))
w(x)
f(x)
[
2
= 1,
2
= 0, = 1, = 3]
=
=1

2
1
(

2
(3

)
Summary: Fourier-Bessel Series
1
( ) ( )
i n i
i
f x c J x o

=
=

2 2
0
1
2
( ) ( ) ,
( )
b
i n i
n i
c xJ x f x dx
b J b
o
o
+
=
}
1
( ) ( )
i n i
i
f x c J x o

=
=

2
2 2 2 2 2
0
2
( ) ( ) ,
( ) ( )
b
i
i n i
i n i
c xJ x f x dx
b n h J b
o
o
o o
=
+
}
1 0
2
( ) ( )
i i
i
f x c c J x o

=
= +

1 0
2 2 2
0 0
1
2 2
( ) , ( ) ( ) ,
( )
b b
i i
i
c xf x dx c xJ x f x dx
b b J b
o
o
= =
} }
:

( ) = 0
:

( ) +

( ) = 0
:
0
( ) = 0
Fourier-Legendre Series: on the interval [-1,1], w(x)=1
1
2
2
1
2
( ) ( ) .
2 1
n n
P x P x dx
n

= =
+
}

0
( ) ( ),
n n
n
f x c P x

=
=

1
1
2 1
( ) ( ) .
2
n n
n
c f x P x dx

+
=
}


Alternative Form of Series
In many cases in spherical coordinate, it is useful to let
0
( ) (cos )
n n
n
F c P u u

=
=

0
2 1
( ) (cos )sin ,
2
n n
n
c F P d
t
u u u u
+
=
}


Proof of this is quite tricky. We just accept this here.
= cos
= cos
= sin & 1,1 , 0