National Institute of Technology Durgapur

Department of Mathematics

Assignment for MA531

Lakshmi Kanta Dey

November 5, 2014

1 Fourier Transform
Definition 1.1. Let f : R R be a function. f is called an absolutely integrable (a.i.)
on R if
Z
|f (x)| dx < .

Definition 1.2. Let f (t) be piecewise

continuous on (, ). Assume that f (t) is
R
absolutely integrable, that is, |f (t)|dt converges. Then, the Fourier transform of
f (t) denoted by F[f (t)] is defined as
Z
1

F[f (t)] =
f (t)eikt dt = F (k).
2
The inverse Fourier transform of F (k) is defined by
Z
1
f (t) =
F (k)eikt dk.
2
Definition 1.3. The Fourier cosine transform of f (x) is defined by
r Z
2
Fc [f (t)] =
f (x) cos kx dx = Fc (k).
0
The inverse Fourier cosine transform is defined by
r Z
2
Fc (k) cos kx dk.
f (x) =
0

Definition 1.4. The Fourier sine transform of f (x) is defined by

r Z
2
f (x) sin kxdx = Fs (k)
Fs [f (t)] =
0
The inverse Fourier sine transform is defined by
r Z
2
f (x) =
Fs (k) sin kx dk.
0
Definition 1.5. The Dirac delta function is defined as follows:

, if x = 0;
(x) =
0, if x 6= 0.
This is also restricted by
Z

(x) dx = 1.

Definition 1.6. The Heavisides unit step function is defined as follows:


0, t < a;
H(t a) =
1, t a.
We also call this function as unit step function only and it is denoted by ua (t).
Property 1.7. If F[f (t)] = F (k) and t0 is any real number, then
1. F[f (t t0 )] = F (k)eikt0 [Shifting on time domain].
2. F[eik0 t f (t)] = F (k k0 ) [Frequency shifting].
3. F[f (t)] = F (k).[Symmetry]
Property 1.8. Let f (x) be piecewise continuous on every interval [l, l]. Let
converge. Then
F[tn f (t)] = in F (n) (k).

Property 1.9. Let f (x) be piecewise continuous on every interval [l, l]. Let
converge. Then
Z x
F (k)
F[
f (t) dt] =
.
ik

|x

n f (x)|dx

|f (x)|dx

Remark 1.10. Fourier Transform of an absolutely integrable function is also a.i.

Definition 1.11. Suppose f and g are two piecewise continuous functions on every
interval [l, l] and a.i. on R. Then we define the convolution of f and g, denoted by
(f g)(x) as
Z
1
(f g)(x) =
f (u)g(x u) du.
2

Theorem 1.12. Let F[f (x)] = F (k), F[g(x)] = G(k). Then F[(f g)(x)] = F (k)G(k)
and F[f (x)g(x)] = [F G](k).
Remark 1.13. Using convolution theorem, we have
F 1 [F (k)G(k)] = (f g)(x)
and
F 1 [(F G)(k)] = f (x)g(x).
Note 1.14. Proofs of Scaling and shifting properties of F.T. are to be known.

1.1 Exercise
1. Define absolutely integrable function and state Fourier Integral Theorem.
2. Give example to show that absolutely integrability is sufficient condition but not
necessary to have F.T.
3. Show that if f is even function, then F (k) = Fc (k) and if f is odd, then F (k) =
Fs (k).
4. Find the F.T. of the following functions:
a) u3 (t)e4t , where u0 (t) is the unit step function.
b) e|t| .
c) ea|t| , < t < , a > 0. Write the inverse transform also.
d) eat H(t), a > 0, where H(t) is Heavisides unit step function.
e) Dirac delta function.
f) 1.
g) H(t).
h)

1
5+it .

5. Find the Fourier transform of the following function:


f (x) =

1, for |x| a,
0, otherise, a > 0.

6. Find the Fourier transform of eat2 . Also find its inverse transform and hence show
x
that the Fourier transform of e 2 is self-reciprocal.
7. Find the Fourier transform of

f (x) =
Hence deduce that

R
0

a |x|, for |x| < a,

0,
otherise, a > 0.

( sinx x )2 dx = 2 .

8. Find Fourier cosine and sine transform of Rthe function defined

= eat , a is
R byxfsin(t)
cos kx
a constant and hence deduce the value of 0 (a2 +x2 ) dx and 0 (a2 +xkx
2 ) dx.
9. Find the Fourier sine transform of f (t) = teat , a > 0 and g(t) = 1t .
10. Find the Fourier cosine transform of f (t) = eat , a > 0. Deduce the inversion
formula.
11. Find the Fourier cosine transform of f (t) =

1
.
a2 +t2

12. State the convolution Theorem for F.T. Also find the inverse F.T. of the following:
a)
b)

1
12+7ikk2
1
,a
(ik+a)2

e)

1
6+5ikk2
1
1+k2
1
4+k2

f)

e4ik
3+ik

g)

e2ik
.
2(4+k2 )

c)
d)

>0

13. Find the inverse Fourier sine transform of the following:

a)
b)

1
1+k2
sin ak
k .

14. Find the solution of the differential equations using Fourier transformation
a) y1 4y = u0 (t)e4t , < t < .
b) y1 + 3y = u0 (t)et , < t < .
c) y2 + 3y1 + 2y = e|t| .
d) y2 + 5y1 + 4y = (t 2).

2 Z-TRANSFORM
Definition 2.1. Let {fn } is a function of the discrete integer valued arguments. The Z
transform of the sequence {fn } is defined as
n
Z{fn } =
= F (z),
n= fn z

whenever the series is convergent. Note that z may be real or complex. The inverse Z
transform is defined as
fn = Z 1 [F (z)].

n , known as causal part. Another

Note 2.2. Above series has two parts. One is
n=0 fn z
n = f
n
part is 1
n= fn z
n=1 n z , known as anti-causal. For a given sequence {fn },
n converges for all z such that
there exists a number R1 such that the series
n=0 fn z
| z1 | < R1 , that is |z| > R11 and diverges for all |z| < R1 . The Z transform F (z), valid in
its region of convergence, is unique.

Definition 2.3. Let {fn } is a sequence of exponential type, that is, there exist numbers
N > 0, t0 0 and n0 0, such that
|fn | < N ent0 f orall n n0 .
Theorem 2.4. (Existence theorem)Every sequence {fn } of exponential type has the
Z transform.
Property 2.5. (Linearity property) Let {fn } and {gn } be two sequences such that
Z{fn } = F (z) and Z{gn } = G(z) with radius of convergence |z| > R11 and |z| > R12
respectively. Then, for any constants a, b
Z{afn + bgn } = aZ{fn } + bZ{gn } = aF (z) + bG(z),
for all |z| >

1
R,

where

1
R

= max[ R11 , R12 ].

Theorem 2.6. (Shifting theorem) Let F (z) be the Z transform of the sequence {fn }
valid in the region |z| > R1 , and k be a positive integer. Then
1. Z{fnk } = z k F (z), n k (shifting to the right)
n ] (shifting to the left)
2. Z{fn+k } = z k [F (z) k1
n=0 fn z

valid in the region |z| >

1
R.

Property 2.7. (Scaling property) Let F (z) be the Z transform of the sequence {fn }
valid in the region |z| > R1 , and a be a any real or complex constant. Then
1. Z{an fn } = F (az)
2. Z{an fn } = F ( az )
Result 3.7.1 is also known as the damping rule.
Property 2.8. (Differentiation of Z transform) Let F (z) be the Z transform of the
sequence {fn } valid in the region |z| > R1 . Then,
Z{nfn } = z
which is also convergent in the region |z| >

d
[F (z)],
dz

1
R.

Theorem 2.9. (Convolution theorem) Let {fn } and {gn } be two sequences such that
Z{fn } = F (z) and Z{gn } = G(z) with radius of convergence |z| > R11 and |z| > R12
respectively. Then,
Z{fn gn } = F (z)G(z),
valid in the region |z| >

1
R,

where

1
R

= max[ R11 , R12 ].

Theorem 2.10. Let F (z) be the Z transform of the sequence {fn } exist in the region
|z| > R1 . Then,
f

1. fp = limz [z p {F (z)f0 fz1 zf22 zp1

p1 }], p = 0, 1, 2, . . . (Initial value theorem)
2. limn = limz1 (z 1)F (z). (F inal value theorem)
Remark 2.11. Proofs of the properties of Z-transforms have to be learnt.

2.1 PROBLEMS
1. Define Z-transform. What do you mean by ROC of Z-transform?
2. Find the Z-transform (with ROC) of the following functions:

1 for n = 0,
a) [n] =
0 for n 6= 0.
1
n for n < 0,
0 for n = 0,
b) s[n] =
n
2 for n > 0.
c) s[n] = an , where a is any real or complex number.
d) s[n] = an , where a is any real or complex number.
e) s[n] = ean , where a is any real or complex number.
f) s[n] =

an
n! ,

where a is any real or complex number.

g) s[n] = (n k).

1 for n 0,
h) u[n] =
0 for n < 0.
i) r[n] = nu[n].
j) s[n] =

1
(n+1) , n

= 0, 1, 2, . . .

3. Using linearity property, obtain the Z-transform of {fn }, where fn = cosh(n).

4. Find the Z-transform of n2 u[n].
5. Find the inverse Z-transform of the following functions by Contour Integral Method
or any other method
a) Z(z) =

z
(z1) , |z|

> 1.

b) Z(z) =

z3
.
(z 2 1)

c) Z(z) =

2z
(z1)(z2) .

d) Z(z) =

(2z 2 +3z)
(z4)(z+2) .

e) Z(z) =

(7z11z 2 )
(z1)(z2)(z+3) .

f) Z(z) =

(z 3 +2z 2 +29z)
.
(z1)(z+3)2

g) Z(z) =

(z 3 +5z 2 +6z)
.
(z2)(z3)3

h) Z(z) =

(z 2 3z+5)
(z1)(z+2) .

Also find limn fn