1.1 Introduction To Fourier Series: A+t B+T T
1.1 Introduction To Fourier Series: A+t B+T T
If a function f is periodic with period T > 0 then f (t) = f (t + T ), −∞ < t < ∞. The
smallest of T , for which the equality f (t) = f (t + T ) is true, is called fundamental period
of f (t). However, if T is the period of a function f then nT , n is any natural number, is
also a period of f . Some familiar periodic functions are sin x, cos x, tan x etc.
We consider two important properties of periodic function. These properties will be used
to discuss the Fourier series.
1. It should be noted that the sum, difference, product and quotient of two functions is
also a periodic function. Consider for example:
f(x)
a a+T b b+T
Here an and bn are some constants. Since the sum of the periodic functions again
represents a periodic function. Therefore Sn will be a periodic function. What will
be the period of the function Sn ? The period can be identified simply by looking at
the common period of the functions involved in the sum as
πx πx 2πx nπx nπx
Period of Sn (x) = common period of cos , sin , cos , . . . , sin , cos
l l l l l
= 2π/(π/l) = 2l.
2
Lecture Notes on Fourier Series
Now the question aries whether any function of period T = 2l can be represented as
the sum of a trigonometric series? The answer to this question is affirmative and it
is possible for a very wide class of periodic functions. In the next lesson we will see
how to obtain the constants an and bn in order this trigonometric series to represent
a given periodic function.
Remark 1: Though sine and cosine functions are quite simple in nature but their sum
function may be quite complex. One can see the plot of sin x + sin 2x + cos 3x in Figure 1.3.
However, the function has a period 2π which is a common period of sin x, sin 2x, cos 3x.
1.5
0.5
−0.5
f(x)
−1
−1.5
−2
−2.5
−3
0 2 4 6 8 10 12 14
x
We call two functions φ(x) and ψ(x) to be orthogonal on the interval [a, b] if
Z b
φ(x)ψ(x) dx = 0
a
3
Lecture Notes on Fourier Series
With this definition we can say that the basic trigonometric system viz.
is orthogonal on the interval [−π, π] or [0, 2π]. In particular, we shall prove that any two
distinct functions are orthogonal.
To show the orthogonality we take different possible combination as:
For any integer n 6= 0: We have the following integrals to show the orthogonality of the
function 1 with any member of sine or cosine family
Z π Z π
sin(nx) π cos(nx) π
1 · cos(nx) dx = = 0, 1 · sin(nx) dx = − =0
−π n −π −π n −π
For any integer m and n (m 6= n): Now we show that any two different members of the
same family (sine or cosine) are orthogonal. For the cosine family we have
Z π Z π
1
cos(nx) cos(mx) dx = [cos(n + m)x + cos(n − m)x] dx = 0
−π 2 −π
For any integer m and n: Here we show that any two members of the two different
family (sine and cosine) are orthogonal
Z π
sin(nx) cos(mx) dx = 0
−π
Note that the integrand is an odd function and therefore the integral is zero.
The above result can be summarized in a more general setting in the following theorem.
4
Lecture Notes on Fourier Series
1.4.1 Theorem
πx πx 2πx 2πx
1, cos , sin , cos , sin ,...
l l l l
is 2l. Similar to the evaluation of the integral appeared above to show orthogonality of the
basic trigonometric system, we have the following results:
(
Z l Z a+2l
mπx nπx mπx nπx 0 if m 6= n
a) cos cos dx = cos cos dx =
−l l l a l l l if m = n 6= 0
(
Z l Z a+2l
mπx nπx mπx nπx 0 if m 6= n
b) sin sin dx = sin sin dx =
−l l l a l l l if m = n 6= 0
Z l Z a+2l
mπx nπx mπx nπx
c) sin cos dx = sin cos dx = 0
−l l l a l l
To summarize, the value of the integral over length of period of integrand is equal to zero
if the integrand is a product of two different members of trigonometric system. If the
integrand is product of two same member from sine or cosine family then the value of
the integral will be half of the interval length on which the integral is performed. These
results will be used to establish Fourier series of a function of period 2l defined on the
interval [−l, l] or [a, a + 2l]. It should be noted that for l = π we obtain results for standard
trigonometric system of common period 2π .
5
In this lesson we shall introduce Fourier series of a piecewise continuous periodic func-
tion. First we construct Fourier series of periodic functions of standard period 2π and then
the idea will be extended for a function of arbitrary period.
such that f is continuous on each open sub-interval (a, t1 ), (tj , tj+1 ) and (tn , b) and all the
following one sided limits exist and are finite
This mean that f is continuous on [a, b] except possibly at finitely many points, at each
of which f has finite one sided limits. It should be clear that all continuous functions are
obviously piecewise continuous.
2.1.1 Example 1
At each point of discontinuity the function has finite one sided limits from both sides. At
the end points x = −π and π right and left sided limits exist, respectively. Therefore, the
function is piecewise continuous.
2.1.2 Example 2
Note that f is continuous everywhere except at x = 0. The function f is also not piecewise
continuous on [0, 1] because limx=0+ f (x) = ∞.
An important property of piecewise continuous functions is boundedness and integrability
over closed interval. A piecewise continuous function on a closed interval is bounded and
integrable on the interval. Moreover, if f1 and f2 are two piecewise continuous functions
then their product, f1 f2 , and linear combination, c1 f1 +c2 f2 , are also piecewise continuous.
Let f be a periodic piecewise continuous function on [−π, π] and has the following trigono-
metric series expansion
∞
a0 X
f∼ + [ak cos(kx) + bk sin(kx)] (2.1)
2
k=1
This implies Z π
1
a0 = f (x) dx.
π −π
Multiplying the series by cos(nx), integrating over [−π, π] and assuming its value equal to
the integral of f (x) cos(nx) over [−π, π], we get
Z π ∞
X Z π Z π
f (x) cos(nx) dx = 0 + ak cos(nx) cos(kx) dx + bk cos(nx) sin(kx) dx
−π k=1 −π −π
Rπ
Note that the first term on the right hand side is zero because −π cos(kx) dx = 0. Further,
using the orthogonality of the trigonometric system we obtain
Z π
1
an = f (x) cos(nx) dx
π −π
2
Lecture Notes on Fourier Series
Similarly, by multiplying the series by sin(nx) and repeating the above steps we obtain
Z π
1
bn = f (x) sin(nx) dx
π −π
Remark 1: In the series (2.1) we can not, in general, replace ∼ by = sign as clear
from the determination of the coefficients. In the process we have set two integrals equal
which does not imply that the function f (x) is equal to the trigonometric series. Later we
will discuss conditions under which equality holds true.
Let f (x) be piecewise continuous function defined in [−l, l] and it is 2l periodic. The
Fourier series corresponding to f (x) is given as
∞
a0 X kπx kπx
f∼ + an cos + bn sin (2.2)
2 l l
k=1
where the Fourier coefficients, derived exactly in the similar manner as in the previous
case, are given as
Z l
1 kπx
ak = f (x) cos dx, k = 0, 1, 2, . . .
l −l l
Z l
1 kπx
bk = f (x) sin dx k = 1, 2, . . .
l −l l
In must be noted that just for simplicity we will be discussing Fourier series of 2π periodic
function. However all discussions are valid for a function of an arbitrary period.
3
Lecture Notes on Fourier Series
2.4.1 Problem 1
Solution: The Fourier series of the given function will represent a 2π periodic function
and the series is given by
∞
a0 X
f (x) ∼ + (an cos(nx) + bn sin(nx))
2
n=1
with Z π Z 0 Z π
1 π
a0 = f (x) dx = − π dx + x dx = −
−π π −π 0 2
and the coefficients an , n = 1, 2, . . . as
Z π Z 0 Z π
1 1
an = f (x) cos(nx) dx = − π cos(nx) dx + x cos(nx) dx
π −π π −π 0
0 π Z π
sin(nx) 1 sin(nx) sin(nx)
=− + x − dx
n −π
π n 0 0 n
4
Lecture Notes on Fourier Series
Remark 4: Let a function is defined on the interval [−l, l]. It should be noted that
the periodicity of the function is not required for developing Fourier series. However,
the Fourier series, if it converges, defines a 2l-periodic function on R. Therefore, this is
sometimes convenient to think the given function as 2l-periodic defined on R.
2.4.2 Problem 2
π
Case I: First we treat the function | sin x| as π periodic we have 2l = π ⇒ l = 2. The
coefficient a0 is given as
Z π Z π
1 2 2 4
a0 = π f (x) dx = sin x dx = [− cos x]π0 = .
2 0 π 0 π π
5
Lecture Notes on Fourier Series
6
Lecture Notes on Fourier Series
Remark 5: If we develop the Fourier series of a function considering its period as any
integer multiple of its fundamental period, we shall end up with the same Fourier series.
Remark 6: Note that in the above example the given function is an even function and
therefore the Fourier series is simpler as we have seen that the coefficient bn is zero in this
case. The determination of the Fourier series of a given function becomes simpler if the
function is odd or even. More detail of this we shall see in the next Lesson.
7
We have seen that piecewise continuity of a function is sufficient for the existence of the
Fourier series. We have not yet discussed the convergence of the Fourier series. Conver-
gence of the Fourier series is a very important topic to be explored in this lesson.
In order to motivate the discussion on convergence, let us construct the Fourier series of
the function (
− cos x, −π/2 ≤ x < 0;
f (x) = f (x + π) = f (x).
cos x, 0 ≤ x ≤ π/2.
Note that the Fourier series at x = 0 converges to 0. So the Fourier series of f does not
converge to the value of the function at x = 0.
With this example we pose the following questions in connection to the convergence of
the Fourier series
1. Does the Fourier series of a function f (x) converges at a point x ∈ [−L, L].
2. If the series converges at a point x, is the sum of the series equal to f (x).
The answers of these questions are in the negative because
1. There are Lebesgue integrable functions on [−L, L] whose Fourier series diverge
everywhere on [−L, L].
2. There are continuous functions whose Fourier series diverge at a countable number
of points.
3. We have already seen in the above examples that the Fourier series converges at a
point but the sum is not equal to the the value of the function at that point.
We need some additional conditions to ensure that the Fourier series of a function f (x)
converges and it converges to the function f (x). Though, we have several notions of
convergence like pointwise, uniform, mean square, etc. we first stick to the most com-
mon notion of convergence, that is, pointwise convergence. Let {fm }∞ m=1 be sequence
Lecture Notes on Fourier Series
exist (and are finite), then for each x ∈ (−L, L) the Fourier series converges and we have
∞
f (x+) + f (x−) a0 X kπx kπx
= + an cos + bn sin
2 2 L L
n=1
At both endpoints x = ±L the series converges to [f (L−) + f ((−L)+)] /2, thus we have
∞
f (L−) + f ((−L)+) a0 X
= + (−1)n an
2 2
n=1
Remark 1: If the function is continuous at a point x, that is, f (x+) = f (x−) then we
have
∞
a0 X kπx kπx
f (x) = + an cos + bn sin (3.2)
2 L L
k=1
In other words, if f is continuous with f (−L) = f (L) and one sided derivatives (3.1) exist
then equality (3.2) holds for all x.
Remark 2: In the above theorem condition on f are sufficient conditions. One may
replace these conditions (piecewise continuity and one sided derivatives) by slightly more
restrictive conditions of piecewise smoothness. A function is said to be piecewise smooth
on [−L, L] if it is piecewise continuous and has a piecewise continuous derivative. The
2
Lecture Notes on Fourier Series
difference between the two similar restrictions on f will be clear from the example of the
function (
x2 sin(1/x), x 6= 0;
f (x) =
0, x = 0.
It can easily easily be shown that derivative of the function exist everywhere and thus the
function has one sided derivatives and satisfy the conditions of the convergence Theorem
(3.1). However the function is not piecewise smooth because the limx→0 f 0 (x) does not
exist as (
2x sin(1/x) − cos(1/x), x 6= 0;
f 0 (x) =
0, x = 0.
If a function is piecewise smooth then it can easily be shown that left and right derivatives
exist. Let f be a piecewise smooth function on [−L, L] then limx→a± f 0 (x) exists for all
a ∈ [−L, L]. This implies
f (x + h) − f (x)
lim f 0 (x) = lim lim
x→a+ x→a+ h→0+ h
Similarly one can shown the existence of left derivative. This example confirms that piece-
wise smoothness is stronger condition than piecewise continuity with existence of one
sided derivatives.
Let {fm }∞m=1 be sequence of functions defined on [a, b]. Let f be defined on [a, b]. We say
that the sequence {fm }∞
m=1 converges in the mean square sense to f on [a, b] if
Z b
lim |f (x) − fm (x)|2 dx = 0
m→∞ a
3
Lecture Notes on Fourier Series
Let {fm }∞ m=1 be sequence of functions defined on [a, b] and let f be defined on [a, b].
We say that {fm }∞ m=1 converges pointwise to f on [a, b] if for each x ∈ [a, b] we have
limm→∞ fm (x) = f (x). That is, for each x ∈ [a, b] and ε > 0 there is a natural number
N (ε, x) such that
Let {fm }∞
m=1 be sequence of functions defined on [a, b] and let f be defined on [a, b]. We
say that {fm }∞
m=1 converges uniformly to f on [a, b] if for each ε > 0 there is a natural
number N (ε) such that
|fn (x) − f (x)| < ε for all n ≥ N (ε), and for all x ∈ [a, b]
There is one more interesting fact about the uniform convergence. If {fm }∞ m=1 is a se-
quence of continuous functions which converge uniformly to a function to f on [a, b], then
f is continuous.
3.2.4 Example 1
we seek for a natural number N (ε) such that relation (3.3) holds for n > N . Note that
relation (3.3) holds true if
ln ε
xn < ε ⇐⇒ n >
ln x
It should be evident now that for given x and ε one can define
ln ε
N := , where [ ] gives integer rounded towards infinity
ln x
4
Lecture Notes on Fourier Series
3.2.5 Example 2
xn
Let un = on [0, 1). This sequence converges uniformly and of course pointwise to 0.
n
For given ε > 0 take n > N := 1ε then noting 1ε > 1ε we have |un − 0| < xn /n < 1/n < ε
5
4.1 Different Notions of Convergence
Let {fm }∞m=1 be sequence of functions defined on [a, b]. Let f be defined on [a, b]. We say
that the sequence {fm }∞
m=1 converges in the mean square sense to f on [a, b] if
Z b
lim |f (x) − fm (x)|2 dx = 0
m→∞ a
Let {fm }∞ m=1 be sequence of functions defined on [a, b] and let f be defined on [a, b].
We say that {fm }∞ m=1 converges pointwise to f on [a, b] if for each x ∈ [a, b] we have
limm→∞ fm (x) = f (x). That is, for each x ∈ [a, b] and ε > 0 there is a natural number
N (ε, x) such that
Let {fm }∞
m=1 be sequence of functions defined on [a, b] and let f be defined on [a, b]. We
say that {fm }∞
m=1 converges uniformly to f on [a, b] if for each ε > 0 there is a natural
number N (ε) such that
|fn (x) − f (x)| < ε for all n ≥ N (ε), and for all x ∈ [a, b]
There is one more interesting fact about the uniform convergence. If {fm }∞ m=1 is a se-
quence of continuous functions which converge uniformly to a function to f on [a, b], then
f is continuous.
4.1.4 Example 1
show that for given ε there does not exist a natural number N independent of x such that
|un − 0| < ε. Suppose that the series converges uniformly, then for a given ε with
we seek for a natural number N (ε) such that relation (4.1) holds for n > N . Note that
relation (4.1) holds true if
ln ε
xn < ε ⇐⇒ n >
ln x
It should be evident now that for given x and ε one can define
ln ε
N := , where [ ] gives integer rounded towards infinity
ln x
4.1.5 Example 2
xn
Let un = on [0, 1). This sequence converges uniformly and of course pointwise to 0.
n
For given ε > 0 take n > N := 1ε then noting 1ε > 1ε we have |un − 0| < xn /n < 1/n < ε
• Let f be a piecewise continuous function on [−π, π] and the appropriate one sided
derivatives of f at each point in [−π, π] exists then for each x ∈ [−π, π] the Fourier
series of f converges pointwise to the value (f (x−) + f (x+))/2.
• If f is continuous on [−π, π], f (−π) = f (π), and f 0 is piecewise continuous on
[−π, π], then the Fourier series of f converges uniformly (and also absolutely) to f
on [−π, π].
2
Lecture Notes on Fourier Series
An interesting property of the partial sums of a Fourier series is that among all trigonomet-
ric polynomials of degree N , the partial sum of Fourier Series yield the best approximation
of f in the mean square sense. This result has been summarized in the following lemma.
4.2.1 Lemma
Let f be piecewise continuous function on [−π, π] and let the mean square error is defined
by the following function
Z π hc N
X i 2
0
E(c0 , . . . , cN , d1 , . . . , dN ) = f − + (ck cos kx + dk sin kx) dx
−π 2
k=1
4.3.1 Problem 1
Find the sum of the Fourier series for all point in [−π, π].
Solution: At x = 0, the Fourier series will converge to
f (0+) + f (0−) 0 + (−π) π
= =−
2 2 2
Again, x = ±π are another points of discontinuity and the value of the series at these point
will be
f (π−) + f ((−π)+) π + (−π)
= = 0;
2 2
At all other points the series will converge to functional value f (x).
3
Lecture Notes on Fourier Series
4.3.2 Problem 2
Let the Fourier series of the function f (x) = x + x2 , −π < x < π be given by
∞
π2 X 4 2
h i
2
x+x ∼ + (−1)n 2 cos nx − sin nx
3 n n
n=1
Find the sum of the Fourier series for all point in [−π, π]. Applying the result on conver-
gence of the Fourier series find the value of
1 1 1 1 1 1
1+ 2
+ 2 + 2 + ... and 1− 2
+ 2 − 2 + ...
2 3 4 2 3 4
At the point x = 0 is a point of continuity and therefore the series will converge to 0.
Substituting x = 0 into the series we obtain
∞ ∞
π2 X 4 X 1 π2
+ (−1)(n) 2 = 0 =⇒ (−1)(1+n) 2 = .
3 n n 12
n=1 n=1
4
In this chapter, we start discussion on even and odd function. As mentioned earlier if the
function is odd or even then the Fourier series takes a rather simple form of containing
sine or cosine terms only. Then we discuss a very important topic of developing a desired
Fourier series (sine or cosine) of a function defined on a finite interval by extending the
given function as odd or even function.
A function is said to be an even about the point a if f (a − x) = f (a + x) for all x and odd
about the point a if f (a − x) = −f (a + x) for all x. Further, note the following properties
of even and odd functions:
a) The product of two even or two odd functions is again an even function.
b) The product of and even function and an odd function is an odd function.
Using these properties we have the following results for the Fourier coefficients
Z π
Z π
1 2
f (x) cos(nx) dx, when f is even function about 0
an = f (x) cos(nx) dx = 0
π −π π
0, when f is odd function about 0
1
Z π
2
0,
Z π when f is even function about 0
bn = f (x) sin(nx) dx =
π −π π f (x) sin(nx) dx, when f is odd function about 0
0
5.1.1 Proposition
a) If f is an even function then the Fourier series takes the simple form
∞ Z π
a0 X 2
f (x) ∼ + an cos(nx) with an = f (x) cos(nx) dx, n = 0, 1, 2, . . . .
2 π 0
n=1
5.2.1 Problem 1
2
Lecture Notes on Fourier Series
5.2.2 Problem 2
Determine the Fourier Series of f (x) = x2 on [-π, π ] and hence find the value of the
∞ ∞
n+1 1 1
X X
infinite series (−1) 2
and 2
.
n n
n=1 n=1
Solution: The function f (x) = x2 is even on the interval [-π, π ] and therefore bn =0 for all
n. The coefficient a0 is given as
Z π
1 x3 π 2π 2
2
a0 = x dx = = .
π π 3π −π 3
3
6.1 Half Range Series
Suppose that f (x) is a function defined on (0, l]. Suppose we want to express f (x) in the
cosine or sine series. This can be done by extending f (x) to be an even or an odd function
on [−l, l]. Note that there exists an infinite number of ways to express the function in the
interval [−l, 0]. Among all possible extension of f there are two, even and odd extensions,
that lead to simple and useful series:
a) If we want to express f (x) in cosine series then we extend f (x) as an even function in
the interval [−l, l].
b) On the other hand, if we want to express f (x) in sine series then we extend f (x) as an
odd function in [−l, l].
6.1.1 Proposition
Remark: Note that we can develop a Fourier series of a function f defined in [0, l]
and it will, in general, contain all sine and cosine terms. This series, if converges, will
represent a l-periodic function. The idea of half range Fourier series is entirely different
where we extend the function f as per our desire to have sine or cosine series. The half
range series of the function f will represent a 2l-periodic function.
Lecture Notes on Fourier Series
6.2.1 Problem 1
Taking second term on the right side to the left side and after simplification we get
2nπ [1 − e(−1)n ]
bn =
1 + n2 π 2
Therefore, the sine series of f is given as
∞
x
X n [1 − e(−1)n ]
e = 2π sin nπx for 0 < x < 1
1 + n2 π 2
n=1
6.2.2 Problem 2
Solution: Sine we want to find cosine series of the function f we compute the coefficients
an as
Z l Z l
2 πx nπx 1 (n + 1)πx (1 − n)πx
an = sin cos dx = sin + sin dx
l 0 l l l 0 l l
2
Lecture Notes on Fourier Series
6.2.3 Problem 3
Expand f (x) = x, 0 < x < 2 in a (i) sine series and (ii) cosine series.
Solution: (i) To get sine series we calculate bn as
Z l Z 2
2 nπx 2 nπx
bn = f (x) sin dx = x sin dx
l 0 L 2 0 2
Integrating by parts we obtain
i2 Z 2
nπx 2 2 nπx 4
h
bn = x cos − + cos dx = − cos nπ.
2 nπ 0 nπ 0 2 nπ
Then for 0 < x < 2 we have the Fourier sine series
4 X∞ cos nπ nπx 4
πx 1 2πx 1 3πx
x=− sin = sin − sin + sin + ... .
π n=1 n 2 π 2 2 2 3 2
3
Lecture Notes on Fourier Series
It is interesting to note that the given function f (x) = x, 0 < x < 2 is represented by
two entirely different series. One contains only sine terms while the other contains only
cosine terms.
Note that we have used series equal to the given function because the series converges for
each x ∈ (0, 2) to the function value. It should also be pointed out that one can deduce
sum of several series by putting different values of x ∈ (0, 2) in the above sine and cosine
series.
4
In this lesson we discuss differentiation and integration of the Fourier series of a function.
We can get some idea of the complexity of the new series if looking at the terms of the
series. In the case of differentiation we get terms like n sin(nx) and n cos(nx), where
presence of n as product makes the magnitude of the terms larger then the original and
therefore convergence of the new series becomes more difficult. This is exactly other
way round in the case of integration where n appears in division and new terms become
smaller in magnitude and thus we expect better convergence in this case. We shall deal
these two case separately in next sections.
7.1 Differentiation
We first discuss term by term differentiation of the Fourier series. Let f be a piecewise
continuous with the Fourier series
∞
a0 X
f (x) ∼ + [an cos(nx) + bn sin(nx)] (7.1)
2
n=1
Can we differentiate term by term the Fourier series of a function f in order to obtain the
Fourier series of f 0 ? In other words, is it true that
∞
X
0
f (x) ∼ [−nan sin(nx) + nbn cos(nx)]? (7.2)
n=1
We consider one more simple example to illustrate this fact. Consider the half range sine
series for cos x in (0, π)
∞
8 X n sin(2nx)
cos x ∼
π (4n2 − 1)
n=1
If we differentiate this series term by term then we obtain the series
∞
16 X n2 cos(2nx)
π (4n2 − 1)
n=1
This series can not be the Fourier series of − sin x because it diverges as
16 n2 cos(2nx)
lim 6= 0
n→∞ π (4n2 − 1)
Lecture Notes on Fourier Series
7.1.1 Theorem
If f is continuous on [−π, π], f (−π) = f (π), f 0 is piecewise continuous on [−π, π], and if
∞
a0 X
f (x) ∼ + [an cos(nx) + bn sin(nx)]
2
n=1
(in fact in this case we can replace ∼ by =) is the Fourier series of f , then the Fourier
series of f 0 is given by
∞
X
0
f (x) ∼ [−nan sin(nx) + nbn cos(nx)] .
n=1
Proof: Since f 0 is piecewise continuous and this is sufficient condition for the existence
of Fourier series of f 0 . So we can write Fourier series of as
∞
0 ā0 X
f (x) ∼ + ān cos(nx) + b̄n sin(nx) (7.3)
2
n=1
where Z π Z π
1 0 1
ān = f (x) cos(nx) dx, b̄n = f 0 (x) sin(nx) dx
π −π π −π
Now we simplify coefficients ān and b̄n and write them in terms of an and bn . Using the
condition f (−π) = f (π), we can easily show that
2
Lecture Notes on Fourier Series
f 0 (x+) + f 0 (x−)
Convergence of this series to or f 0 (x) is a direct consequence of conver-
2
gence theorem of Fourier series.
7.2 Integration
In general, for an infinite series uniform convergence is required to integrate the series
term by term. In the case of Fourier series we do not even have to assume the convergence
of the Fourier series to be integrated. However, integration term by term of a Fourier series
does not, in general, lead to a Fourier series. The main results can be summarize as:
7.2.1 Theorem
Let f be piecewise continuous function and have the following Fourier series
∞
a0 X
f (x) ∼ + [an cos(nx) + bn sin(nx)] (7.4)
2
n=1
Then no matter whether this series converges or not we have for each x ∈ [−π, π],
Z x ∞
a0 (x + π) an bn
X
f (t)dt = + sin(nx) − cos(nx) − cos nπ (7.5)
−π 2 n n
n=1
and the series on the right hand side converges uniformly to the function on the left.
Proof: We define Z x
a0
g(x) = f (t)dt − x
−π 2
3
Lecture Notes on Fourier Series
and Z π
a0 a0 a0 π
g(π) = f (t)dt − π = πa0 − π =
−π 2 2 2
Hence, the Fourier series of the function g converges uniformly to g on [−π, π]. Thus we
have
∞
α0 X
g(x) = + [αn cos(nx) + βn sin(nx)]
2
n=1
Using Theorem 7.1.1 we have the following result for the Fourier series of g 0 as
∞
X
0
g (x) ∼ [−nαn sin(nx) + nβn cos(nx)]
n=1
nβn = an − nαn = bn n = 1, 2, . . .
Remark 1: Note that the series on the right hand side of (7.5) is not a Fourier series
due to presence of x.
4
Lecture Notes on Fourier Series
is its Fourier series then no matter whether this series converges or not, it is true that
Z x Z x ∞ Z x
a0 X
f (t)dt = a0 dx + [an cos(nx) + bn sin(nx)] dx
a 2 a a
n=1
where −π ≤ a ≤ x ≤ π and the series on the right hand side of converges uniformly in x
to the function on the left for any fixed value of a.
5
In this lesson some properties of the Fourier coefficients will be given. We will mainly de-
rive two important inequalities related to Fourier series, in particular, Bessel’s inequality
and Parseval’s identity. One of the applications of Parseval’s identity for summing certain
infinite series will be discussed.
Using the orthogonality of the trigonometric system and definition of Fourier coefficients
we get
π n n
a2
Z X X
2
f (x) dx + 0 π + π a2k + b2k − a20 π − 2π a2k + b2k + 0 ≥ 0
−π 2
k=1 k=1
This implies
n π
a20 X 2
Z
1
+ ak + b2k ≤ f 2 (x) dx
2 π −π
k=1
Lecture Notes on Fourier Series
If f is a continuous function in [−π, π] and one sided derivatives exit then we have the
equality
∞ π
a20 X 2
Z
1
+ an + b2n = f 2 (x) dx (8.1)
2 π −π
n=1
Remark: As stated earlier Parseval’s identity can be proved for piecewise continuous
functions. Further, for a piecewise continuous function on [−L, L] we can get Parseval’s
identity just by replacing π by L in (8.1).
2
Lecture Notes on Fourier Series
8.3.1 Problem 1
Solution: a) We first find the Fourier coefficient and the period of the Fourier series just
by comparing the given series with the standard Fourier series
4
a0 = 2, an = [cos(nπ) − 1], n = 1, 2 . . . , bn = 0
π 2 n2
period = 2L = 4 ⇒ L = 2
This implies
Z 2 ∞
1 4 X 16
2
x dx = + 4 4
(cos(nπ) − 1)2
2 −2 2 π n
n=1
3
Lecture Notes on Fourier Series
π4
Then we have the required sum as S = .
90
8.3.2 Problem 2
Find the Fourier series of x2 , −π < x < π and use it along with Parseval’s theorem to
show that ∞
X 1 π4
=
(2n − 1)4 96
n=1
2π 2
Z
2
a0 = πx2 dx =
π 0 3
4
Lecture Notes on Fourier Series
π
2π 4
Z
1
Using x4 dx = we get
π −π 5
∞
4π 4 X 16 2π 4
+ =
18 n4 5
n=1
This implies
∞
X 1 π4
=
n4 90
n=1
Now using the idea of splitting of the series from the Example 8.3.1 (b), we have
∞ ∞ ∞ ∞
X 1 X 1 1 X 1 15 X 1
= − =
(2n − 1)4 n4 16 n4 16 n4
n=1 n=1 n=1 n=1
∞
X 1
Substituting the value of in the above equation we get the required sum.
n4
k=1
8.3.3 Problem 3
we have ∞ Z π
1 16 16 X 1 1
+ = cos2 (x/2) dx = 1
2 π2 π2 (4n2 − 1)2 π −π
n=1
Then,
∞
X 1 π2 − 8
= .
(4n2 − 1)2 16
n=1