Ref:- Denis G.

Zill-Differential Equations with boundary value problems

Orthogonal functions

A function can be considered to be a generalization of a vector. Two

vector concepts of inner product and orthogonality can be extended to
functions.

Inner product
The inner product of two functions f1 and f2 on an interval [a, b]
is the value
Z b
(f1 , f2 ) = f1 (x)f2 (x)dx.
a
Orthogonal functions
Two functions f1 and f2 are sand to be orthogonal on an interval [a, b] if their
inner product Z b
(f1 , f2 ) = f1 (x)f2 (x)dx = 0
a

is a zero. There’s however no geometric significance (like "⊥ er") on the

condition above.
For example: The functions 1, x, 3x2 /2 − 1/2 are orthogonal on the interval
[−1, 1] since the inner products

1 1
x2
Z
1 2
1 − (−1)2 = 0

(f1 , f2 ) = (1)(x)dx = =
−1 2 −1 2
Z 1 1
1 1
 2  Z
3x 1 1 3
3x2 − 1 dx =



(f1 , f3 ) = (1) − dx = x −x
−1 2 2 2 −1 2 −1
1
= (1 − 1 − (−1 + 1)) = 0
2

1 1
3x2 1 1 1 1 3 4 x2
Z   Z 
3


(f2 , f3 ) = (x) − dx = 3x − x · dx = x −
−1 2 2 2 −1 2 4 2 −1
  
1 3 1 3 1
= − − − =0
2 4 2 4 2

1
 Orthogonal Sets
The norm (or length) of a vector p v is expressed (in the form of the inner
product (v, v) = ∥v∥2 or ∥v∥ = (v, v).p The norm, or generalised length of
a function ϕn is defined as ∥ϕn (x)∥ = (ϕn , ϕn ) which is :

s
Z b
∥ϕn (x)∥ = ϕ2n (x)dx
a
Z b
2
or ∥ϕn (x)∥ = ϕ2n (x)dx =⇒ the square of the norm of ϕn
a

Orthonormal Sets
If {ϕn (x)} is an orthogonal set of functions on the interval [a, b] with the
additional property that ∥ϕn (x)∥ = 1 for n = 0, 1, 2, . . ., then {ϕn (x)} is said
to be an orthonormal set on the interval.

2
Z 1 √
∥f1 (x)∥ = 12 dx = x|1−1 = 1 − (−1) = 2 ⇒ ∥f1 ∥ = 2
−1
1
Z 1 r
x3 1 2 2
∥f2 (x)∥2 = x2 dx = = (1 − (−1)) = ⇒ ∥f2 ∥ =
−1 3 1 3 3 3
Z 1 2 2 Z 1
3x 1 1
∥ f3 (x)|2 = 9x4 − 6x2 + 1 dx


− dx =
−1 2 2 −1 4
 5 1  
1 9x 3 1 9 9
= − 2x + x = − 2 + 1 − (− + 2 − 1)
4 5 −1 4 5 5
   
1 18 1 18 10 4
= −4+2 = − = or.
4 5 4 5 5 5
2
∥f3 ∥ = √
5
( r √ )
1 3 5
3x2 − 1


{ϕn (x)} = √ , x ,
2 2 4

is an orthonormal set.
Exercise: Show that the set of functions {1, cos x, sin x, cos 2x, sin 2x, . . .}
is orthogonal on the interval [−π, π]. Find the norms of each function in the
given orthogonal set.

2
Orthogonal set/Weight function
A set of functions {ϕn (x)} , n = 0, 1, 2, . . . is said to be orthogonal w.r.t. a
weight function ω(x) on the interval [a, b] if
Z b
ω(x)ϕm (x)ϕn (x)dx = 0, m ̸= n.
a
Orthogonal Expansion
If {ϕn (x)}∞
n=1 is an orthogonal system with respect to ω(x) on [a, b], we might
ask if we can expand a function ϕ(x) in terms of these functions; that is can
we express ϕ in the form

ϕ(x) = c1 ϕ1 (x) + c2 ϕ2 (x) + · · · (1)

for suitable choice of constants c1 , c2 , . . . ? Such an expansion is an ortho-
gonal expansion, or a generalized Fourier series.

To determine the constants in (1) we can use the orthogonality of the

system. Multiting (1) by ω(x)ϕm (x) and integrate to obtain

Z b Z b Z b
ϕ(x)ϕm (x)ω(x)dx = c1 ϕ1 (x)ϕm (x)w(x)dx + c2 ϕ2 (x)ϕm (x)ω(x)dx · · ·
a a a
∞
X Z b
= cn ϕn (x)ϕm (x)ω(x)dx (2)
n=1 a

Since the system is orthogonal w.r.t. ω(x), every integral on the right
hand side of (2) in zero except when n ≡ m. Solving for cm gives

Rb Rb
ϕ(x)ϕm (x)ω(x)dx ϕ(x)ϕm (x)w(x)dx
cm = a R b = a
where m = 1, 2, 3, . . .
ϕ 2 (x)ω(x)dx
m
∥ϕm ∥2
a

Fourier Series
The set of functions
 
π 2π π 2π 3π
1, cos x, cos x, · · · , sin x, sin x, sin x, · · ·
p p p p p

is orthogonal on the interval [−p, p].

3
 A Fourier series of a funciton f (x) in the interval [−p, p] is given by
∞  
a0 X nπ nπ
f (x) = + an cos x + bn sin x .
2 n=1
p p
Integrating both sides over the interval leads to
Z p ∞ 
a0 p
Z Z p Z p 
X nπ nπ
f (x)dx = dx + an cos xdx + bn sin xdx = a0 p
−p 2 −p n=1 −p p −p p
1 p
Z
⇒ a0 = f (x)dx.
p −p
Multiplying both sides by cos mπ p
and integrating again over the interval leads to
Z p Z p ∞  Z p
mπ a0 mπ X nπ mπ
f (x) cos xdx = cos xdx + an cos x cos xdx
−p p 2 −p p i=1 −p p p
Z p 
nπ mπ
+bn cos x sin xdx
−p p p
p
an p
Z   
2nπ an p 2nπ
= 1 + cos x dx = x+ sin x
2 −p p 2 2nπ p −p

Z p
an 1 nπ
= (2p) ⇒ an = f (x) cos xdx.
2 p −p p
Lastly multiplying both sides by sin mπ p
and integrating again over the interval leads to
Z p Z p ∞  Z p
mπ a0 mπ X nπ mπ
f (x) sin xdx = sin xdx + an cos x sin xdx
−p p 2 −p p n=1 −p p p
Z p 
nπ mπ
+bn sin x sin xdx
−p p p
p
bn p
Z   
2nπ bn p 2nπ
= 1 − cos x dx = x− sin x
2 −p p 2 2nπ p −p
1 p
Z
bn nπ
= (2p) ⇒ bn = f (x) sin xdx.
2 p −p p
Example.
Compute the Fourier series for the given function
(
1, −2 < x < 0
f (x) =
x, 0 < x < 2

4
 " #
2 0 2 2
x2
Z Z Z 
1 1 1 0
a0 = f (x)dx = (1)dx + xdx = x|−2 +
2 −2 2 −2 0 2 2 0
1
= (2 + 2) = 2
2Z
1 2
Z 0 Z 2 
nπx 1 nπx nπx
an = f (x) cos dx = cos dx + x cos dx
2 −2 2 2 −2 2 0 2
" #
0 2 Z 2
1 2 nπx 2x nπx 2 nπx
= sin + sin − sin dx
2 nπ 2 2 nπ 2 0 0 nπ 2
" #
2
1 4 nπx 2
= 2 2
cos = 2 2 [(−1)n − 1]
2 nπ 2 0 nπ
−4
⇒ a2n−1 =
(2n − 1)2 π 2
Z 2 Z 0 Z 2 
1 nπx 1 nπx nπx
bn = f (x) sin dx = sin dx + x sin dx
2 −2 2 2 −2 2 0 2
" #
0 2 Z 2
1 2 nπx −2x nπx 2 nπx
= − cos + cos + cos dx
2 nπ 2 −2 nπ 2 0 nπ 0 2
" #
2
1 2 2 4 nπx
= − [1 − (−1)n ] − [2(−1)n − 0] + 2 2 sin
2 nπ nπ nπ 2 0
 
1 2 4 2
= − [1 − (−1)n ] − (−1)n = [−1 + (−1)n − 2(−1)n ]
2 nπ nπ 2nπ
1
=− [1 + (−1)n ]
nπ
1
⇒ b2n = −
nπ
So the Fourier series for f (x) is
∞  
X 4 nπx 1 nπx
f (x) = 1 − 2 2
cos + sin
n=1
(2n − 1) π 2 nπ 2
Fourier Series
Definition 1: Let f be a piecewise continuous function on the interval [−T, T ].
The Fourier series of f is the trigonometric series

5
 ∞
a0 X n nπx nπx o
f (x) = + an cos + bn sin
2 n=1
T T
Where the an ’s and bn ’s are given by the formulae

1 T
Z
nπx
an = f (x) cos dx, n = 0, 1, 2, · · · (3)
T −T T
1 T
Z
nπx
bn = f (x) sin dx n = 1, 2, 3, · · · (4)
T −T T

Formulae 3 & 4 are called the Euler formulae or Fourier coefficients of f .

Note that f (x) need not be defined for every x in [−T, T ]; we only need that
integrals in (3) & (4) to exist.
The even and odd functions

There are two symmetry properties of functions that will be in the study
of Fourier series. A function f (x) that satisfies f (−x) = f (x) for all x in the
domain of f has a graph that is symmetric w.r.t the y-axis. We say such a
function is an even function. A function f (x) that satisfies f (−x) = −f (x)
for all x in the domain of f has a graph that is symmetric w.r.t to the origin.
It is said to be an odd function.

Knowing that a function is even or odd can be helpful in evaluating

definite integrals.
Properties of symmetric functions:
If f (x) is an even piecewise continuous function on [−a, a], then
Z a Z a
f (x)dx = 2 f (x)dx (4)
−a 0

If f (x) is an odd piecewise continuous. function on [−a, a], then

Z a
f (x)dx = 0 (5)
−a