Professional Documents
Culture Documents
Fourier Series
Fourier Series
Fourier Series
Jean Baptiste Joseph Fourier (1768-1830) was a French mathematician, physicist and engineer, and the founder of Fourier analysis. In 1822 he made the
claim, seemingly preposterous at the time, that any function of t, continuous
or discontinuous, could be represented as a linear combination of functions
sin nt. This was a dramatic distinction from Taylor series. While not strictly
true in fact, this claim was true in spirit and it led to the modern theory of
Fourier analysis with wide applications to science and engineering.
1.1
We will study this ON set and the completeness and convergence of expansions in the basis, both pointwise and in the norm. Before we get started, it
is convenient to assume that L2 [0, 2] consists of square-integrable functions
on the unit circle, rather than on an interval of the real line. Thus we will
replace every function f (t) on the interval [0, 2] by a function f (t) such
that f (0) = f (2) and f (t) = f (t) for 0 t < 2. Then we will extend
1
Proof Each side of the identity is just the integral of F over one period.
For an analytic proof, note that
Z 2
Z a
Z 2
Z a
Z 2
F (t)dt =
F (t)dt +
F (t)dt =
F (t + 2)dt +
F (t)dt
0
2+a
2+a
F (t)dt
F (t)dt +
F (t)dt =
F (t)dt +
2+a
F (t)dt.
a
Thus we can transfer all our integrals to any interval of length 2 without
altering the results.
R 2+a
Exercise 1 Let G(a) = a
F (t)dt and give a new proof of of Lemma 1
based on a computation of the derivative G0 (a).
For students who dont have a background in complex variable theory we
will define the complex exponential in terms of real sines and cosines, and
derive some of its basic properties directly. Let z = x + iy be a complex
number, where x and y are real. (Here and in all that follows, i = 1.)
Then z = x iy.
Definition 1 ez = exp(x)(cos y + i sin y)
Lemma 2 Properties of the complex exponential:
ez1 ez2 = ez1 +z2
|ez | = exp(x)
2
d int
e
dt
= ineint .
Lemma 4 (en , em ) = nm .
Proof If n 6= m then
Z
1
(en , em ) =
2
ei(nm)t dt =
1 ei(nm)t 2
| = 0.
2 i(n m) 0
R 2
1
If n = m then (en , em ) = 2
1 dt = 1. 2
0
Since {en } is an ON set, we can project any f L2 [0, 2] on the closed
subspace generated by this set to get the Fourier expansion
f (t)
n=
or
f (t)
X
n=
int
cn e ,
1
cn =
2
f (t)eint dt.
(1.1.1)
This is the complex version of Fourier series. (For now the just denotes
that the right-hand side is the Fourier series of the left-hand side. In what
3
2
2
|f (t)| dt
0
|cn |2 .
(1.1.2)
n=
an ibn
,
2
n = 0, 1, 2,
an + ibn
,
n = 1, 2,
2
and rearrange terms, we get the real version of Fourier series:
cn =
a0 X
(an cos nt + bn sin nt),
(1.1.3)
+
f (t)
2
n=1
Z
Z
1 2
1 2
an =
f (t) cos nt dt
bn =
f (t) sin nt dt
0
0
1 2
|a0 |2 X
2
(|an |2 + |bn |2 ),
|f (t)| dt
+
0
2
n=1
and Riemann-Lebesgue Lemma
Z 2
Z
lim
f (t) cos(nt)dt = lim
n
f (t) sin(nt)dt = 0.
|(f, en )|2 .
n=
In terms of the complex and real versions of Fourier series this reads
1
2
or
1
2
2
|f (t)| dt =
0
|cn |2
(1.1.4)
n=
|f (t)|2 dt =
|a0 |2 X
(|an |2 + |bn |2 ).
+
2
n=1
Let f L2 [0, 2] and remember that we are assuming that all such
functions satisfy f (t + 2) = f (t). We say that f is piecewise continuous on [0, 2] if it is continuous except for a finite number of discontinuities. Furthermore, at each t the limits f (t + 0) = limh0,h>0 f (t + h) and
f (t0) = limh0,h>0 f (th) exist. NOTE: At a point t of continuity of f we
have f (t+0) = f (t0), whereas at a point of discontinuity f (t+0) 6= f (t0)
and f (t + 0) f (t 0) is the magnitude of the jump discontinuity.
Theorem 2 Suppose
f (t) is periodic with period 2.
f (t) is piecewise continuous on [0, 2].
f 0 (t) is piecewise continuous on [0, 2].
Then the Fourier series of f (t) converges to
1.2
f (t+0)+f (t0)
2
at each point t.
Examples
We will use the real version of Fourier series for these examples. The transformation to the complex version is elementary.
1. Let
0,
t=0
0 < t < 2
0,
t = 2.
R 2
and f (t + 2) = f (t). We have a0 = 1 0 t
dt = 0. and for n 1,
2
Z
Z 2
t
sin nt 2
1 2 t
1
2
cos nt dt =
sin nt dt = 0,
an =
0 +
0
2
n
2n 0
Z
Z 2
t
cos nt 2
1 2 t
1
1
2
bn =
sin nt dt =
cos nt dt = .
0
0
2
n
2n 0
n
Therefore,
t X sin nt
=
,
0 < t < 2.
2
n
n=1
f (t) =
t
,
2
1 1 1 1
= 1 + + .
4
3 5 7 9
Parsevals identity gives
2 X 1
=
.
6
n2
n=1
2. Let
1
,
2
1,
f (t) =
1
,
2
0
t=0
0<t<
t=
< t < 2.
R
and f (t + 2) = f (t) (a step function). We have a0 = 1 0 dt = 1, and
for n 1,
Z
1
sin nt
an =
cos nt dt =
| = 0,
0
n 0
2
Z
cos nt
1
(1)n+1 + 1
, n odd
n
bn =
sin nt dt =
|0 =
=
0,
n even.
0
n
n
Therefore,
1 2 X sin(2j 1)t
.
f (t) = +
2 j=1
2j 1
= sin t +
+
+ ,
4
3
5
and for < t < 2 it gives
sin 3t sin 5t
= sin t +
+
+ .
4
3
5
2 X
1
=
.
8
(2j 1)2
j=1
1.3
2(x )
in our previous formulas. Every function f (x) L2 [, ] is uniquely associated with a function f(t) L2 [0, 2] by the formula f (x) = f( 2(x)
). The
q
q
2n(x)
2n(x)
2
2
1
,
set {
cos ,
sin } for n = 1, 2, is an ON
basis for L2 [, ], The real Fourier expansion is
a0 X
2n(x )
2n(x )
f (x)
+
(an cos
+ bn sin
),
(1.3.5)
2
n=1
Z
Z
2
2n(x )
2
2n(x )
an =
f (x) cos
dx, bn =
f (x) sin
dx
2
|a0 |2 X
|f (x)|2 dx =
+
(|an |2 + |bn |2 ).
2
n=1
7
a0 X
F (t)
+
an cos nt
2
n=1
on [, ] or
a0 X
f (t)
+
an cos nt, for 0 t
(1.3.6)
2
n=1
Z
Z
2
1
F (t) cos nt dt =
f (t) cos nt dt.
an =
0
Here, (1.3.6) is called the Fourier cosine series of f .
We can also extend the function f (t) from the interval [0, ] to an odd
function on the interval [, ]. We define G(t) L2 [, ] by
on (0, ]
f (t)
0
for t = 0
G(t) =
bn sin nt
n=1
on [, ] or
f (t)
bn sin nt,
for o < t ,
n=1
Z
Z
1
2
G(t) sin nt dt =
f (t) sin nt dt.
bn =
0
Here, (1.3.7) is called the Fourier sine series of f .
8
(1.3.7)
0 t .
X
2(1)n+1
n=1
sin nt,
0 < t < .
1.4
4 X cos(2j 1)t
,
2 j=1 (2j 1)2
0 < t < .
Convergence results
a0 X
f (t)
+
(an cos nt + bn sin nt) = S(t),
(1.4.8)
2
n=1
Z
Z
1 2
1 2
an =
f (t) cos nt dt,
bn =
f (t) sin nt dt.
0
0
For a fixed t we want to understand the conditions under which the Fourier
series converges to a number S(t), and the relationship between this number
and f . To be more precise, let
k
Sk (t) =
a0 X
+
(an cos nt + bn sin nt)
2
n=1
be the k-th partial sum of the Fourier series. This is a finite sum, a trigonometric polynomial, so it is well defined for all t R. Now we have
S(t) = lim Sk (t),
k
if the limit exists. To better understand the properties of Sk (t) in the limit,
we will recast this finite sum as a single integral. Substituting the expressions
for the Fourier coefficients an , bn into the finite sum we find
1
Sk (t) =
2
Z
0
Z 2
k Z 2
1X
f (x) sin nx dx sin nt ,
f (x)dx+
f (x) cos nx dx cos nt +
n=1
0
0
so
#
k
1 X
(cos nx cos nt + sin nx sin nt f (x)dx
+
2 n=1
0
#
Z "
k
1 2 1 X
=
+
cos[n(t x)] f (x)dx
0
2 n=1
Z
1 2
=
Dk (t x)f (x)dx.
(1.4.9)
0
P
We can find a simpler form for the kernel Dk (t) = 21 + kn=1 cos nt =
P
12 + km=0 cos mt. The last cosine sum is the real part of the geometric
series
k
X
(eit )k+1 1
(eit )m =
eit 1
m=0
1
Sk (t) =
so
"
1 X
1
(eit )k+1 1
+
cos mt = + Re
2 m=0
2
eit 1
10
1 iu/2 iu/2
(e
e
ei(k+1/2)u
(eit )k+1 12 eit 21
2
= Re
= Re
eit 1
eiu/2 eiu/2
sin(k + 1/2)u + i(cos u/2 cos(k + 1/2)u)
.
= Re
2 sin u/2
Thus,
Dk (t) =
sin(k + 21 )t
.
2 sin 2t
(1.4.10)
t
t
t+
For fixed t we can write Dk (t x) in the form
Dk (t x) =
Dk (t) =
Dk (x)dx =
0
Proof;
Z
Z
Dk (x)dx =
1 X
cos nx)dx = .
( +
2 n=1
2
Theorem 4 Suppose
f (t) is periodic with period 2.
f (t) is piecewise continuous on [0, 2].
f 0 (t) is piecewise continuous on [0, 2].
12
f (t+0)+f (t0)
2
at each point t.
exists. Thus H1 , H2 are bounded for x 0+. Then, by the RiemannLebesgue Lemma, the last expression goes to 0 as k :
lim [Sk (t) f (t)] = 0.
2
Exercise 3 Suppose f is piecewise continuous and 2-periodic. For any
point t define the right-hand derivative fR0 (t) and the left-hand derivative
fL0 (t) of f by
f (u) f (t 0)
f (u) f (t + 0)
, fL0 (t) = lim
,
ut+
ut
ut
ut
respectively. Show that in the proof of Theorem 4 we can drop the requirement
that f 0 is piecewise continuous and the conclusion of the theorem will still hold
at any point t such that both fR0 (t) and fL0 (t) exist.
fR0 (t) = lim
13
Exercise 4 Show that if f and f 0 are piecewise continuous then for any point
t we have f 0 (t + 0) = fR0 (t) and f 0 (t 0) = fL0 (t). Hint: By the mean value
theorem of calculus, for u > t and u sufficiently close to t there is a point c
such that t < c < u and
f (u) f (t + 0)
= f 0 (c).
ut
Exercise 5 Let
f (t) =
t2 sin( 1t ) for t 6= 0
0
for t = 0.
Show that f is continuous for all t and that fR0 (0) = fL0 (0) = 0. Show that
f 0 (t + 0 and f 0 (t 0) do not exist for t = 0. Hence, argue that f 0 is not a
piecewise continuous function. This shows that the result of Exercise 3 is a
true strengthening of Theorem 4.
Exercise 6 Let
(
f (t) =
2 sin
t
t
2
if 0 < |t| ,
if t = 0.
1.4.1
a0 X
1 2
f (t)
+
f (t) cos nt dt
(an cos nt + bn sin nt) = S(t),
an =
2
0
n=1
Z
1 2
bn =
f (t) sin nt dt,
0
14
converge to
f (t+0)+f (t0)
:
2
k
a0 X
Sk (t) =
+
(an cos nt + bn sin nt),
2
n=1
lim Sk (t) =
f (t + 0) + f (t 0)
.
2
(t0)
(If we require that f satisfies f (t) = f (t+0)+f
at each point then the series
2
will converge to f everywhere. In this section we will make this requirement.)
Now we want to examine the rate of convergence.
We know that for every > 0 we can find an integer N (, t) such that
|Sk (t) f (t)| < for every k > N (, t)|. Then the finite sum trigonometric
polynomial Sk (t) will approximate f (t) with an error < . However, in
general N depends on the point t; we have to recompute it for each t. What
we would prefer is uniform convergence. The Fourier series of f will converge
to f uniformly if for every > 0 we can find an integer N () such that
|Sk (t) f (t)| < for every k > N () and for all t. Then the finite sum
trigonometric polynomial Sk (t) will approximate f (t) everywhere with an
error < .
We cannot achieve uniform convergence for all functions f in the class
above. The partial sums are continuous functions of t, Recall from calculus
that if a sequence of continuous functions converges uniformly, the limit
function is also continuous. Thus for any function f with discontinuities, we
cannot have uniform convergence of the Fourier series.
If f is continuous, however, then we do have uniform convergence.
1 2
a0 X
+
f (t) cos nt dt
(an cos nt + bn sin nt),
an =
f (t)
2
0
n=1
Z
1 2
bn =
f (t) sin nt dt,
0
15
A0 X
f (t)
+
(An cos nt + Bn sin nt),
2
n=1
Z
1 2 0
Bn =
f (t) sin nt dt,
0
0
1
An =
f 0 (t) cos nt dt
Now
1
An =
Z
0
1
n
f (t) cos nt dt = f (t) cos nt|2
0 +
nbn , n 1
=
0,
n=0
0
f (t) sin nt dt
0
0
Using Bessels inequality for f 0 we have
1
(|An | + |Bn | )
n=1
2
hence
|f 0 (t)|2 dt < ,
n2 (|an |2 + |bn |2 ) .
n=1
Now
Pm
m
m
X
X
p
1p
|a
|
n
n=1
2
2
Pm
|an | + |bn | =
|An |2 + |Bn |2
|b
|
n
n
n=1
n=1
m
X
n=1
n=1
m
X
1
)( (|An |2 + |Bn |2 ))
n2 n=1
which converges as
. (We P
have used the Schwarz inequality for the
Pm
X
a0 X
a0
| +
(an cos nt + bn sin nt)| | | +
(|an cos nt| + |bn sin nt|)
2
2
n=1
n=1
X
a0
|+
(|an | + |bn |) < ,
2
n=1
|a0 |2 X
1 2
2
2
+
(|an | + |bn | ) =
|f (t)|2 dt.
2
0
n=1
Proof: The Fourier series of f converges uniformly, so for any > 0 there
is an integer N () such that |Sk (t) f (t)| < for every k > N () and for all
t. Thus
Z 2
k
|a0 |2 X
+
(|an |2 +|bn |2 ) < 22
|Sk (t)f (t)|2 dt = ||Sk f ||2 = ||f ||2 (
2
0
n=1
for k > N (). 2
Remark 1: Parsevals Theorem actually holds for any f L2 [0, 2], as we
shall show later.
Remark 2: As the proof of the preceding theorem illustrates, differentiability of a function improves convergence of its Fourier series. The more
derivatives the faster the convergence. There are famous examples to show
that continuity alone is not sufficient for convergence.
1.5
t=0
0,
t
, 0 < t < 2
h(t) =
2
0,
t = 2.
and h(t + 2) = h(t). In this case, an = 0 for all n and bn = n1 . Therefore,
t X sin nt
=
,
2
n
n=1
0 < t < 2.
Here, Sk (t) =
Pk
n=1
sin nt
n
so
k
X
sin(k + 12 )t 1
sin kt2 cos (k+1)t
1
2
=
cos nt = Dk (t) =
=
.
t
t
2
2
2 sin 2
sin 2
n=1
Sk0 (t)
Sk0 (x)dx =
sin kx
cos (k+1)x
2
2
2 sin x2
!
dx.
Note that Sk0 (0) > 0 so that Sk starts out at 0 for t = 0 and then increases.
Looking at the derivative of Sk we see that the first maximum is at the
dx
x
2 sin 2
2
tk
Z
Sk (tk ) h(tk ) =
0
tk
Z
=
0
sin(k + 12 )x
dx +
x
tk
1
1
1
sin(k + )x dx
x
2 sin 2
x
2
2
0
= I(tk ) + J(tk )
2
Z
where
Z
I(tk ) =
0
tk
sin(k + 12 )x
dx =
x
(k+ 21 )tk
Z
0
sin u
du
u
1
1
x
2 sin 2
x
Z
0
sin u
du 1.851397052
u
1
sin(k + )x dx.
2
18
.280600725
2
1
2
=
.
it
|1 e |
sin 2t
P
int
Note that Sk (t) is the imaginary part of the complex series kn=1 e n .
P
eint
Lemma 8 Let 0 < < < 2. The series
n=1 n converges uniformly
for all t in the interval [, ].
Proof: (tricky)
k
X
eint
n=j
k
X
En (t) En1 (t)
n=j
k
X
En (t)
n=j
k
X
En (t)
n=j
n+1
Ej1 (t)
Ek (t)
+
j
K +1
and
|
k
X
eint
n=j
1
|
n
sin 2t
k
X
1
1
1
1
)+ +
(
n n+1
j k+1
n=j
!
=
2
.
j sin 2t
P
int
This implies by the Cauchy Criterion that kn=j e n converges uniformly on
[, ]. 2
This shows that the Fourier series for h(t) converges uniformly on any
closed interval that doesnt contain the discontinuities at t = 2`, ` =
0, 1, 2, . Next we will show that the partial sums Sk (t) are bounded
for all t and all k. Thus, even though there is an overshoot near the discontinuities, the overshoot is strictly bounded.
From the lemma on uniform convergence above we already know that the
partial sums are bounded on any closed interval not containing a discontinuity. Also, Sk (0) = 0 and Sk (t) = Sk (t), so it suffices to consider the
interval 0 < t < 2 .
19
We will use the facts that 2 sint t 1 for 0 < t 2 . The right-hand
inequality is a basic calculus fact and the left-hand one is obtained by solving
the calculus problem of minimizing sint t over the interval 0 < t < 2 . Note
that
k
X t sin nt
X sin nt
X
sin nt
||
|+|
|.
|
n
nt
n
n=1
1n<1/t
1/tnk
k
X
sin nt
n=1
t+
1n<1/t
1
t
2
1
2
+ 1 t 2 = 1 + 2.
t t
t
sin 2
t2
Thus the partial sums are uniformly bounded for all t and all k.
We conclude that the Fourier series for h(t) converges uniformly to h(t)
in any closed interval not including a discontinuity. Furthermore the partial
sums of the Fourier series are uniformly bounded. At each discontinuity
tN = 2N of h the partial sums Sk overshoot h(tN + 0) by about 17.9%
(approaching from the right) as k and undershoot h(tN 0) by the
same amount.
All of the work that we have put into this single example will pay off,
because the facts that have emerged are of broad validity. Indeed we can
consider any function f satisfying our usual conditions as the sum of a continuous function for which the convergence is uniform everywhere and a finite
number of translated and scaled copies of h(t).
Theorem 6 Let f be a complex valued function such that
f (t) is periodic with period 2.
f (t) is piecewise continuous on [0, 2].
f 0 (t) is piecewise continuous on [0, 2].
f (t) =
Then
f (t+0)+f (t0)
2
at each point t.
a0 X
f (t) =
+
(an cos nt + bn sin nt)
2
n=1
pointwise. The convergence of the series is uniform on every closed interval
in which f is continuous.
20
`
X
s(xj )
j=1
h(t xj )
is everywhere continuous and also satisfies all of the hypotheses of the theorem. Indeed, at the discontinuity xj of f we have
g(xj 0) = f (xj 0)
2s(xj )
f (xj + 0) f (xj 0))
h(0) = f (xj 0)
(
)
f (xj 0) + f (xj + 0)
= f (xj ).
2
Similarly, g(xj + 0) = f (xj ). Therefore g(t) can be expanded in a Fourier
P
s(x )
series that converges absolutely and uniformly. However, `j=1 j h(t xj )
can be expanded in a Fourier series that converges pointwise and uniformly
in every closed interval that doesnt include a discontinuity. But
=
f (t) = g(t) +
`
X
s(xj )
j=1
h(t xj ),
1
|a0 |2 X
+
(|an |2 + |bn |2 ) =
2
n=1
|f (t)|2 dt.
I1 I2 I`
< 2
`
X
+(
|Ij |)(M + P )2 < + = .
4
2 2
j=1
Thus limk ||Sk f || = 0 and the partial sums converge to f in the mean.
Furthermore,
Z 2
k
|a0 |2 X
2
2
2
|Sk (t) f (t)| dt = ||Sk f || = ||f || (
>
+
(|an |2 + |bn |2 )
2
0
n=1
for k > N (). 2
1.6
The convergence theorem and the version of the Parseval identity proved
in the previous section apply to step functions on [0, 2]. However, we already know that the space of step functions on [0, 2] is dense in L2 [0, 2].
Since every step function is the limit in the norm of the partial sums of its
Fourier series, this means that the space of all finite linear
combinations of
int
2
int
the functions {e } is dense in L [0, 2]. Hence {e / 2} is an ON basis
for L2 [0, 2] and we have
Theorem 7 Parsevals Equality (strong form) [Plancherel Theorem]. If f
L2 [0, 2] then
Z
|a0 |2 X
1 2
2
2
+
(|an | + |bn | ) =
|f (t)|2 dt,
2
0
n=1
where an , bn are the Fourier coefficients of f .
22
a0 X
(an cos nt + bn sin nt)
+
f (t)
2
n=1
be the Fourier series of f . Then
Z
0
X
a0
1
f (x)dx = t +
[an sin nt bn (cos nt 1)] .
2
n
n=1
a0
2
A0 X
F (t) =
+
(An cos nt + Bn sin nt).
2
n=1
Now
1
An =
Z
0
2
Z 2
F (t) sin nt
1
a0
F (t) cos nt dt =
(f (t) ) sin nt dt
n
n 0
2
0
=
bn
,
n
n 6= 0,
23
and
1
Bn =
Z
0
2
Z 2
F (t) cos nt
1
a0
F (t) sin nt dt =
+
(f (t) ) cos nt dt
n
n 0
2
0
=
Therefore,
an
.
n
F (t) =
A 0 X bn
an
+
( cos nt +
sin nt),
2
n
n
n=1
A 0 X bn
F (2) = 0 =
.
2
n
n=1
Solving for A0 we find
Z
f (x)dx
F (t) =
0
X1
a0
t=
[an sin nt bn (cos nt 1)] .
2
n
n=1
2
Example 2 Let
f (t) =
Then
t
2
0 < t < 2
t = 0, 2.
t X sin nt
.
2
n
n=1
Integrating term-by term we find
X
1
2t t2
(cos nt 1),
=
2
4
n
n=1
0 t 2.
t X sin nt
,
2
n
n=1
24
1 X
cos nt,
2 n=1
which doesnt converge on [0, 2]. In fact it cant possible be a Fourier series
for an element of L2 [0, 2]. (Why?)
If f is sufficiently smooth and periodic it is OK to differentiate term-byterm to get a new Fourier series.
Theorem 9 Let f be a complex valued function such that
f (t) is periodic with period 2.
f (t) is continuous on [0, 2].
f 0 (t) is piecewise continuous on [0, 2].
f 00 (t) is piecewise continuous on [0, 2].
Let
a0 X
f (t) =
+
(an cos nt + bn sin nt)
2
n=1
be the Fourier series of f . Then at each point t [0, 2] where f 00 (t) exists
we have
X
0
n [an sin nt + bn cos nt] .
f (t) =
n=1
Proof: 0By the 0Fourier convergence theorem the Fourier series of f 0 con(t0 0)
verges to f (t0 +0)+f
at each point t0 . If f 00 (t0 ) exists at the point then
2
the Fourier series converges to f 0 (t0 ), where
A0 X
f (t)
+
(An cos nt + Bn sin nt).
2
n=1
0
Now
1
An =
Z
0
2
Z
f (t) cos nt
n 2
f (t) cos nt dt =
f (t) sin nt dt
+
0
0
0
= nbn ,
25
R 2
A0 = 1 0 f 0 (t)dt = 1 (f (2) f (0)) = 0 (where, if necessary, we adjust the
interval of length 2 so that f 0 is continuous at the endpoints) and
2
Z
Z
1 2 0
f (t) sin nt
n 2
Bn =
f (t) sin nt dt =
f (t) cos nt dt
0
0
0
= nan .
Therefore,
0
f (t)
n=1
Note the importance of the requirement in the theorem that f is continuous everywhere and periodic, so that the boundary terms vanish in the
integration by parts formulas for An and Bn . Thus it is OK to differentiate
the Fourier series
X
2t t2 2
1
f (t) =
=
cos nt, 0 t 2
4
6
n2
n=1
term-by term, where f (0) = f (2), to get
t X sin nt
f (t) =
.
2
n
n=1
0
However, even though f 0 (t) is infinitely differentiable for 0 < t < 2 we have
f 0 (0) 6= f 0 (2), so we cannot differentiate the series again.
1.7
We know that the kth partial sum of the Fourier series of a square integrable
function f :
k
a0 X
(an cos nt + bn sin nt)
Sk (t) =
+
2
n=1
is the trigonometric polynomial of order k that best approximates f in the
Hilbert space sense. However, the limit of the partial sums
S(t) = lim Sk (t),
k
26
doesnt necessarily converge pointwise. We have proved pointwise convergence for piecewise smooth functions, but if, say, all we know is that f is
continuous then pointwise convergence is much harder to establish. Indeed
there are examples of continuous functions whose Fourier series diverges at
uncountably many points. Furthermore we have seen that at points of discontinuity the Gibbs phenomenon occurs and the partial sums overshoot
the function values. In this section we will look at another way to recapture f (t) from its Fourier coefficients, by Ces`aro sums (arithmetic means).
This method is surprisingly simple, gives uniform convergence for continuous
functions f (t) and avoids most of the Gibbs phenomena difficulties.
The basic idea is to use the arithmetic means of the partial sums to
approximate f . Recall that the kth partial sum of f (t) is defined by
Z 2
1
f (x)dx +
Sk (t) =
2 0
Z 2
k Z 2
1X
f (x) cos nx dx cos nt +
f (x) sin nx dx sin nt ,
n=1
0
0
so
"
#
k
1 X
+
(cos nx cos nt + sin nx sin nt f (x)dx
2 n=1
0
#
Z "
k
1 2 1 X
=
cos[n(t x)] f (x)dx
+
0
2 n=1
Z
1 2
=
Dk (t x)f (x)dx.
0
P
P
where the kernel Dk (t) = 21 + kn=1 cos nt = 21 + km=0 cos mt. Further,
1
Sk (t) =
Dk (t) =
sin(k + 12 )t
cos kt cos(k + 1)t
=
.
2 sin 2t
4 sin2 2t
Rather than use the partial sums Sk (t) to approximate f (t) we use the
arithmetic means k (t) of these partial sums:
k (t) =
k = 1, 2, .
(1.7.12)
Then we have
#
Z 2 "
k1 Z
k1
1 X
1 X 2
Dj (t x)f (x)dx =
k (t) =
Dj (t x) f (x)dx
k j=0 0
k j=0
0
1
=
where
Fk (t x)f (x)dx
(1.7.13)
k1
k1
1 X sin(j + 21 )t
1X
Dj (t) =
.
Fk (t) =
k j=0
k j=0 2 sin 2t
Lemma 9
1
Fk (t) =
k
sin kt/2
sin t/2
2
.
ei(j+ 2 )t = ei 2
j=0
sin kt2
eikt 1
i kt
2
=
e
.
eit 1
sin 2t
Lemma 10
2
k (t) =
k
/2
Z
0
f (t + 2x) + f (t 2x)
2
sin kx
sin x
2
dx.
/2
Z
0
sin kx
sin x
2
dx = 1.
Proof: Let f (t) 1 for all t. Then k (t) 1 for all k and t. Substituting
into the expression from lemma 10 we obtain the result. 2
Theorem 10 (Fejer) Suppose f (t) L1 [0, 2], periodic with period 2 and
let
f (t + 0) + f (t 0)
f (t + x) + f (t x)
=
(t) = lim
x0+
2
2
whenever the limit exists. For any t such that (t) is defined we have
lim k (t) = (t) =
f (t + 0) + f (t 0)
.
2
2
k (t) (t) =
k
Gt (x)
sin kx
sin x
2
2
dx +
k
/2
Gt (x)
sin kx
sin x
Now
2
2
Z /2
2 Z
sin kx
sin kx
Gt (x)
dx
dx =
k 0
k 0
sin x
sin x
2
29
2
dx.
and
2
Z /2
2 Z /2
sin
kx
2I
2
Gt (x)
|Gt (x)|dx
,
dx
2
k sin
k
sin x
k sin2
R /2
where I = 0 |Gt (x)|dx. This last integral exists because F is in L1 . Now
choose K so large that 2I/(N sin2 ) < /2. Then if k K we have
2
2
Z /2
2 Z
sin
kx
2
sin
kx
|k (t)(t)|
G (x)
dx+
Gt (x)
dx < .
k 0 t
k
sin x
sin x
2
Corollary 3 Suppose f (t) satisfies the hypotheses of the theorem and also
is continuous on the closed interval [a, b]. Then the sequence of arithmetic
means k (t) converges uniformly to f (t) on [a, b].
Proof: If f is continuous on the closed bounded interval [a, b] then it is
uniformly continuous on that interval and the function Gt is bounded on [a, b]
with upper bound M , independent of t. Furthermore one can determine the
in the preceding theorem so that |Gt (x)| < /2 whenever 0 < x and
uniformly for all t [a, b]. Thus we can conclude that k , uniformly on
[a, b]. Since f is continuous on [a, b] we have (t) = f (t) for all t [a, b]. 2
Corollary 4 (Weierstrass approximation theorem) Suppose f (t) is real and
continuous on the closed interval [a, b]. Then for any > 0 there exists a
polynomial p(t) such that
|f (t) p(t)| <
for every t [a, b].
Sketch of Proof: Using the methods of Section 1.3 we can find a linear transformation to map [a, b] one-to-one on a closed subinterval [a0 , b0 ] of
[0, 2], such that 0 < a0 < b0 < 2. This transformation will take polynomials in t to polynomials. Thus, without loss of generality, we can assume
0 < a < b < 2. Let g(t) = f (t) for a t b and define g(t) outside that
interval so that it is continuous at T = a, b and is periodic with period 2.
Then from the first corollary to Fejers theorem, given an > 0 there is an
integer N and
30
arithmetic sum
N
A0 X
+
(t) =
(Aj cos jt + Bj sin jt)
2
j=1
such that |f (t) (t)| = |g(t) (t)| < 2 for a t b. Now (t) is
a trigonometric polynomial and it determines a poser series expansion in
t about the origin that converges uniformly on every finite interval. The
partial sums of this power series determine a series of polynomials {pn (t)} of
order n such that pn uniformly on [a, b], Thus there is an M such that
|(t) pM (t)| < 2 for all t [a, b]. Thus
|f (t) pM (t)| |f (t) (t)| + |(t) pM (t)| <
for all t [a, b]. 2
This important result implies not only that a continuous function on a
bounded interval can be approximated uniformly by a polynomial function
but also (since the convergence is uniform) that continuous functions on
bounded domains can be approximated with arbitrary accuracy in the L2
norm on that domain. Indeed the space of polynomials is dense in that
Hilbert space.
Another important offshoot of approximation by arithmetic sums is that
the Gibbs phenomenon doesnt occur. This follows easily from the next
result.
Lemma 12 Suppose the 2-periodic function f (t) L2 [, ] is bounded,
with M = supt[,] |f (t)|. Then |n (t)| M for all t.
Proof: From (1.7.13) and (10) we have
1
|k (t)|
2k
|f (t+x)|
sin kx/2
sin x/2
2
M
dx
2k
Z
0
sin kx/2
sin x/2
2
dx = M.
Q.E.D.
Now consider the example which has been our prototype for the Gibbs
phenomenon:
t=0
0,
t
, 0 < t < 2
h(t) =
2
0,
t = 2.
31
t X sin nt
=
,
2
n
n=1
0 < t < 2.
and this series exhibits the Gibbs phenomenon near the simple discontinuities
at integer multiples of 2. Furthermore the supremum of |h(t)| is /2 and
it approaches the values /2 near the discontinuities. However, the lemma
shows that |(t)| < /2 for all n and t. Thus the arithmetic sums never
overshoot or undershoot as t approaches the discontinuities, so there is no
Gibbs phenomenon in the arithmetic series for this example.
In fact, the example is universal; there is no Gibbs phenomenon for arithmetic sums. To see this, we can mimic the proof of Theorem 6. This then
shows that the arithmetic sums for all piecewise smooth functions converge
uniformly except in arbitrarily small neighborhoods of the discontinuities
of these functions. In the neighborhood of each discontinuity the arithmetic
sums behave exactly as does the series for h(t). Thus there is no overshooting
or undershooting.
Remark 1 The pointwise convergence criteria for the arithmetic means are
much more general (and the proofs of the theorems are simpler) than for
the case of ordinary Fourier series. Further, they provide a means of getting
around the most serious problems caused by the Gibbs phenomenon. The technical reason for this is that the kernel function Fk (t) is nonnegative. Why
dont we drop ordinary Fourier series and just use the arithmetic means?
There are a number of reasons, one being that the arithmetic means k (t)are
not the best L2 approximations for order k, whereas the Sk (t) are the best L2
approximations. There is no Parseval theorem for arithmetic means. Further, once the approximation Sk (t) is computed for ordinary Fourier series,
in order to get the next level of approximation one needs only to compute two
more constants:
Sk+1 (t) = Sk (t) + ak+1 cos(k + 1)t + bk+1 sin(k + 1)t.
However, for the arithmetic means, in order to update k (t) to k+1 (t) one
must recompute ALL of the expansion coefficients. This is a serious practical
difficulty.
32
1.8
Additional Exercises
Exercise 8 (1) Let X be the Eucldean space Cn of all n 1 tuples (x1 , .., xn )
of complex numbers. For x, y X, define
(x.y) =
n
X
xj yj
j=1
where z denotes the complex conjugate of z C. Show that (; .; ) satisfies the properties of positivity, homogeneity and linearity and with
symmetry, replaced by compex symmetry (x, y) = (x, y). Hence (; .; )
defines a compex inner product on X.
p
(2) Define ||x|| = (x, x) for x X and verify directly that (1) ||; || is a
norm on X, (2) that the Cauchy-Schwartz inequality
|(x, y)| ||x||||y||
holds on X and (3) that the parallelogram law
||x + y||2 + ||x y||2 = 2||x||2 + 2||y||2
holds on X.
Exercise 9 Let X b the space R[, ] with inner product
Z
1
f (x)g(x)dx.
(f, g) =
Show that the sequence of functions
1
, cos(x), sin(x), cos(2x), sin(2x), .....
2
is an orthonormal sequence in X.
Exercise 10 Let X = R[1, 1] with inner product
Z 1
(f, g) =
f (x)g(x)dx.
1
33
X
2
(1)n1 sin(nx).
n
n=1
(b) Let > 0. Show that f (x) = exp(x), x [, ] has the Fourier
series
!
cos(kx)
k
sin(kx)
1
e e
(1)k
+
.
2 k=1
2 + k 2
(c) Let be any real number other than an integer. Let f (x) = cos(x), x
[, ]. Show that f has a Fourier series
sin() 1 X sin( + n)
sin( n)
+
++
cos(nx).
n=1
+n
n
(d) Find the Fourier series of f (x) = sin(x), x [, ]. Do you
notice any relationship to that in (c)?
(e) Find the Fourier series of f (x) = |x|, x [, ].
34
1, x < 0
0,
x=0
f (x) = sign(x) =
1,
x>0
Do you notice any relationship to that in (e)?
Exercise 12 Show that if f R[, ] is odd, thenPall its Fourier cosine
coefficients an = 0, n 0 so f has a Fourier series
n=1 bn sin(nx). Likewise, if f is even, show that allP
its Fourier sine coeffcients bn = 0, n 0 and
f has a Fourier series a0 /2 +
n=1 an cos(nx).
Exercise 13 By substituting special values of x in convergent Fourier series,
we can often deduce interesting series expansions for various numbers, or just
find the sum of important series.
(a) Use the Fourier series for f (x) = x2 to show that
X
(1)n1
n=1
n2
2
.
12
(b) Prove that the Fourier series for f (x) = x converges in (, ) and
hence that
X
(1)l
= .
2l + 1
4
l=0
(c) Show that the Fourier series of f (x) = exp(x), x [, ) converges
in [, ) to exp(x) and at x = to exp()+exp()
. Hence show
2
that
X
22
=1+
.
2 + 2
tanh()
k
k=1
(d) Show that
1 X (1)n1
= +
.
4
2 n=1 4n2 1
35
X
l=0
1
2
.
=
(2l + 1)2
8
Exercise 14 Let f R[, ] be 2 periodic and let n , n = 0, 1, 3, ... denote its Fejer means. Show that
max {|n (x)| : x [, ]} sup {|f (x)| : x [, ]} .
Exercise 15 The following exercise concerns an extension of Parsevals formula to R[, ]. Consider R[, ] with inner product
Z
1
f (x)g(x)dx, f, g R[, ]
(f, g) =
p
and ||f || = (f, f ). Fix f R[, ] and > 0.
(a) Because f R[, ], choose a partition
= x0 < x1 < x2 < ... < xn =
such that if
Mj := sup {f (x) : x [xj1 , xj ]} , mj := inf {f (x) : x [xj1 , xj ]} , j = 1, ..., n
then
n
X
(Mj mj )(xj xj1 ) < .
j=1
and
|h(x)| M = max {|M1 |, ..., |Mn |} , x [, ].
Hint: Sketch the graph of g and modify g near each xj .
36
and
1
||f h|| =
4
M
(f ) If (an ) and (bn ) are the Fourier series coefficients of f , deduce that
a20 /2
(a2n
b2n )
n=1
1
=
f 2 (x)dx.
X
1
2
=
.
2
n
6
n=1
X
1
4
=
.
n4
90
n=1
(c) Evaluate
X
n=1
1
, > 0.
+ n2
37
1
4
.
=
(2l + 1)4
96
l=0
Exercise 17 Is
X
cos(nx)
n=1
n1/2
the Fourier series of some f R[, ]? The same question for the series
X
cos(nx)
.
log(n
+
2)
n=1
n=2
1
2
n(n + 1)
= 1/3.
n=1
6
1+
(n + 1)(2n + 9)
= 21/8.
n=0 (1 + x ) =
1
, |x| < 1.
1x
=
.
n=1 1 +
(n + a)(n + b)
b+1
Hint: Note that
(b+1)
(a1)
= ab + a b 1, (b + 1) + (a 1) = a + b.
1
Exercise 22 Is
n=0 (1 + n ) convergent?
0 eu 1 3u, u (0, 1)
1 + n1 [e1/n 1] converges.
38
n
X
1
k=1
, n 1.
Z
1/k
1/xdx.
k
Deduce that
Z
Hn
n+1
1/xdx = log(n + 1)
1
and that
1
1
[Hn log(n + 1)] [Hn log n] =
+ log 1
n+1
n+1
k
X
1
=
/k < 0.
n+1
k=2
Deduce that Hn log n decreases as n increases and thus has a non-negative
limit , called the Euler-Mascheroni constant and with value approximately
0.5772... not known to be either rational or irrational. Finally show that
nk=1 (1 + 1/k)e1/k = nk=1 (1 + 1/k)nk=1 e1/k =
n+1
exp((Hn log n)).
n
Deduce that
1/k
= e .
k=1 (1 + 1/k)e
Many functions have nice infinite product expansions which can be derived from first principles. The following examples illustrate some of these
nice expansions.
Exercise 25 The aim of this exercise, is to establish Eulers reflection formula:
, Re(x) > 0.
(x)(1 x) =
sin(x)
39
Z
(x)(1 x) =
0
tx1
dt.
1+t
Step 2: Let now C consist of 2 circles about the origin of radii R and
respectively which are joined alog the negative real axis from R to . Show
that
Z
z x1
dz = 2i
C 1z
where z x1 takes its principle value.
Now let us write,
Z x1
iRx exp(ix)
t
exp(ix)
2i =
d +
dt
1+t
1 R exp(i)
R
Z x
Z R x1
i exp(ix)
t
exp(ix)
+
d +
dt.
1+t
1 exp(i)
Let R and 0+ to deduce the result for 0 < x < 1 and for Rex 0 by
continuation.
Exercise 26 Using the previous exercise, derive the following expansions for
all arguments the RHS is meaningful.
(1)
2
sin(x) = x
.
j=1 1 (x/j)
(2)
cos(x) =
j=1 1
x
(j 1/2)
2 !
.
n=0
1 + azq n
1 zq n
, |q| < 1, |z| < 1, |a| < 1.
n=0
(a + 1)zq n
1+
1 zq n
converges absolutely.
(ii) Show that
(1 z)H(z) = (1 + az)H(zq).
(iii) Write
H(z) =
An z n , A0 = 1.
n=0
(1 + a)(1 + aq)...(1 + aq n1 )
, n 1.
(1 q)(1 q 2 )...(1 q n )
X
(1 + a)(1 + aq)...(1 + aq n1 )
n=1
(1 q)(1 q 2 )...(1 q n )
z =
n=0
1 + azq n
1 zq n
.
(iv) Deduce that if 0 < |b| < |a| and |zb| < 1,
X
(b + a)(b + aq)...(b + aq n1 )
1 + azq n
n
1+
z = n=0
.
(1 q)(1 q 2 )...(1 q n )
1 bzq n
n=1
41
j
D(n)q n =
j=1 (1 + q ).
n=0
42