PH408: Assignment - 2

1) Find the relevant Fourier series for the function f (x) = π/4 (over x ∈ (0, π]) and = −π/4
(over x ∈ [−π, 0)), and obtain the corresponding Parseval’s relation. Using this and the Parse-
val’s
P identity1 obtained

for the function g(x) = x/2 over x ∈ [−π, π], find the value of the series
n=2,4,6,... n2 .

2) Find the relevant Fourier series representation of | sin x| over x ∈ [−π, π]. Write the cor-
responding Parseval’s identity.

1
3) Start with a relevant Fourier series representation of the function bn (x) = n (for |x| < 2n )
1 1
and = 0 (for |x| > 2n ) over the interval x ∈ [0, l] with l > 2n . Here n is a positive integer. Using
Rl
the Fourier series check that −l bn (x)dx = 1 for any n. As n → ∞, and then l → ∞ show that
that bn (x) approaches the Fourier transform integral representation of the Dirac-delta function.

4) Show that the necessary and sufficient conditions for f (x) to be (i) real is that its Fourier
transform f˜(k) = f˜(−k), and (ii) imaginary is that Fourier transform f˜(k) = −f˜(−k). Here . . .
denotes complex conjugate.
 
1 1
5) Find the Fourier transform of the function f (x) = 2πi E −i Γ −x , for positive and nega-
0 2
tive k. It is given that Γ > 0.

6) Starting with
R π the integral representation of the Bessel function of order zero, namely,
1 iak sin τ
J0 (ak) = 2π −π e dτ , show that its inverse transform is a function
q
g(x) = π2 √a21−x2 (for |x| < a) and = 0 (for |x| > a).
q
1
7) Show that the Fourier transform of the function f (x) = |x|α is f (k) = ˜ 2
sin( πα ) Γ(1−α) .
π 2 |k|1−α
Think of k > 0 first – the integral breaks into two pieces – one of which has to be done with a
square contour in the 1st quadrant (with sides running along x-axis, then up, then horizontal,
then down along y-axis, then close with a small quarter circle) , and another with a similar one
in the 4th quadrant. Then you may use the answer to just extend to k < 0.

8) Solve the following equation for the function φ(x) using Fourier transform:
d2 φ(x)
−D + K 2 Dφ(x) = Qδ(x)
dx2
Usual boundary condition of vanishing values of the function and its derivatives at ±∞ are to
be assumed.

9) The 3-d Fourier transform is defined as g̃(k) = (2π)13/2 g(r)e−ik.r d3 r. If g̃(k) = 1 1

R
(2π)3/2 k2
,
find the inverse transform g(r) (hint: use spherical polar coordinates).

10) Using the Parseval-Plancherel relation, and the Fourier transforms of some relevant func-
R∞
2 R∞
4
tions done in the lectures 1 and 2, calculate (i) −∞ sink k dk and (ii) −∞ sink k dk.

1
11) Let u(t) = (1 − t)[θ(t) − θ(t − 1)] and v(t) = e−t θ(t). Find the convolution function
Y (t) = u(t) ~ v(t), and plot it to check that it has a new peak position which u(t) or v(t) did
not have. Find its Fourier transform Ỹ (ω) directly and then check the convolution theorem that
indeed Ỹ (ω) = ũ(ω).ṽ(ω).
2
12) Solve the equation I d dtθ(t)
2 + kθ(t) = kat, where k, a, I are constants and t is time. It is
given that θ̇(t = 0) = θ(t = 0) = 0. Find θ(t) using the method of Laplace transform.

1 2 √
13) Find the Laplace transforms of: (i) 2a
(sin at + at cos at), (ii) √1πt − a ea t erfc(a t) (where
µ−1
erfc() is complementary error function), (iii) (t−k)
Γ(µ)
θ(t−k) (where θ() is the Heaviside function),
P∞ n−1 √ 1
(iv) n=1 a θ(t − nk). Show that (v) Laplace transform of [ (2n)!n!√πt H2n ( t)] = (1 − s)n /sn+ 2
where you have to use the series representation of the even order Hermite polynomial (check
any book or web).

1
14) Find the Laplace inverse transforms of (i) 4a3 /(s4 + 4a4 ), (ii) 8a3 s2 /(s2 + a2 )3 , (iii) √s(s−a2) .

In the last case, the integral contribution from the branch lips in the Bromwich integral are to
be done by elementary method. At a stage try taking derivative of an integral with respect to
t and then integrate – it is an interesting new method to do definite integrals which sometimes
work.

15) A student is always late for school by a random delay time t. The net delay t is due
to the sum of three delays due to waking up from sleep, taking breakfast, and travelling by
bus. Let those three random times t1 , t2 and t3 be distributed according to three independent
probability distributions: τ11 e−t/τ1 , τ12 e−t/τ2 , and τ13 e−t/τ3 .
(i) Find the probability distribution P (t) of time t.
(ii) If τ1 = τ2 = τ3 = τ , recalculate P (t), and tell what famous distribution is this.
(iii) Find the mean hti and variance σ 2 in case (ii) by using both P (t) and P̃ (s) (and cross-check
the results from the two methods). If τ = 10 mins, find the mean and standard deviation – isn’t
the value of standard deviation a bit surprising to you? Think about it.

16) Suggested exercise (not to be submitted): All of you should carefully understand the proof
of "law of large numbers" and "Central limit theorem" in statistics – they involve all the math-
ematical machinery of Fourier transform and asymptotic analysis which are the most useful
things you have learned in this course.