Final Exam in Math2 (Spring 2022-2023)

1 Department of Applied Physics and Astronomy, University of Sharjah, P.O. Box 27272 Sharjah, UAE.
(Dated: 15 May 2023)

REMARK : Total score: 45 pts + 16 pts bonus.

Exercise 1: (10 pts)

a2 x 2
q
Find the Fourier transform of the functions f ( x ) = 2
a
√
π
e− 4 and g( x ) = π 1 − a| x |
2 ae .

================================================================================
The Fourier transform of the function f ( x ) is
ˆ +∞ ˆ +∞ ˆ +∞
T 1 a 2
− a4 x2 iωx a a2 x2 +iωx
F (ω ) = [ f ( x )] = √ dx f ( x )e iωx
= √ dx e e = √ dx e− 4 . (1)
2π −∞ 2 2π −∞ 2 2π −∞

We have
2
a2 2 a2 a2 ω2 a2 ω2
    
4 2iω 2iω 2iω
− x + iωx = − x2 − iωx =− x2 − 2 x + ( 2 )2 − =− x− − ,
4 4 a2 4 a2 a a 2 4 a2 a2
2
 2 2 2
− a4 x − 2iω − ω2 − ω2 a2 2
then by making the variable change t = x − 2iω
a2
, we get e 2 a a =e a e− 4 t , or
2 ˆ +∞ √
a −ω 2
− a4 t2 π − ω2
F ( ω ) = √ e a2 dx e = √ e a2 . (2)
2 2π −∞ 2 2
For the function g( x ), we have
ˆ +∞ ˆ
π 1 +∞
r
1 1
G (ω ) = [ g( x )] T = √ dx g( x )eiωx = √ dx e− a| x| eiωx
2π −∞ 2π 2 a −∞
ˆ 0 ˆ +∞ 0 +∞
1 1 1 eax+iωx 1 e− ax+iωx
= dx eax+iωx + dx e− ax+iωx = +
2a −∞ 2a 0 2a a + iω −∞ 2a − a + iω 0
1 1−0 1 0−1
 
1 1 1 1
= + = − = 2 . (3)
2a a + iω 2a − a + iω 2a a + iω − a + iω a + ω2

================================================================================

Exercise 2: (12 pts)

A 2D space with coordinates (y, z) is defined by the metric

!
1 0
gµν = . (4)
0 ρ
!
y
The position vector contravariant components are xµ = , transform as
z

y0 = αy + βz
z0 = βy + αz. (5)
 2

1- If (5) can be written as x 0µ = Λ ν x ν , find the transformation matrix Λ ν .

µ µ
−
→
2- If Aµ = (q, r ) is a vector, i.e., A0µ = Λ ν Aν , simplify the scalar products − →
µ
x . A = gµν x µ Aν and
−
→ −
→
x 0 . A 0 = gµν x 0µ A0ν .
3- The scalar product of any two vectors should be invariant under the transformation (5), i.e., the equality
−
→ −
→ −
→
x .A = − →x 0 . A 0 must hold whatever the vectors components y, z, q and r. Then, find the value of ρ and a
relation between α and β.

================================================================================
1- From (5), the transformation matrix is
!
α β
Λ ν=
µ
. (6)
β α

2- The two scalar products can be simplified as

−
→ −
→
x . A = gµν x µ Aν = g11 x1 A1 + g12 x1 A2 + g21 x2 A1 + g22 x2 A2
= qy + ρzr
−
→ −
→
x 0 . A 0 = gµν x 0µ A0ν = y0 q0 + ρz0 r 0 . (7)

3- The scalar product is invariant under the transformation (5); means

qy + ρzr = y0 q0 + ρz0 r 0 = (αy + βz)(αq + βr ) + ρ( βy + αz)( βq + αr )

= (α2 + ρβ2 )yq + ( β2 + ρα2 )zr + βα(1 + ρ)yr + αβ(1 + ρ)zq.

This means ρ = −1 and α2 − β2 = 1.

================================================================================
Exercise 3: (8 pts)

Let’s consider the ordinary differential equation

x2 f 00 ( x ) + x f 0 ( x ) + ( x2 − ν2 ) f ( x ) = 0, (8)

that gives the Bessel solutions Jν ( x ) and J−ν ( x ).

1- Show that the solution can be written as a series

∞
(−1)k  x ν+2k
f (x) = ∑ k!(k + ν)! 2
. (9)
k =0

√
2- Let’s consider the function g( x ) = x f ( x ), re-werite (8) in function of g( x ); and find the solution for
thye special case ν = 1/2.

================================================================================
1- In order to prove that (9) is a solution for (8), we need to estimate
∞
(ν + 2k)(−1)k  x ν+2k−1
f 0 (x) = ∑ 2k!(k + ν)! 2
,
k =0
∞
(ν + 2k − 1)(ν + 2k)(−1)k  x ν+2k−2
f 00 ( x ) = ∑ 4k!(k + ν)! 2
,
k =0
 3

then
∞ ∞ ∞
(ν + 2k − 1)(ν + 2k)(−1)k  x ν+2k−2 (ν + 2k)(−1)k  x ν+2k−1 (−1)k  x ν+2k
x2 ∑ 4k!(k + ν)! 2
+ x ∑ 2k!(k + ν)! 2 + ( x 2
− ν 2
) ∑ k!(k + ν)! 2 =0
k =0 k =0 k =0
∞ ∞
(−1)k h i  x ν+2k 4(−1)k  x ν+2k+2
∑ k!(k + ν)! ( ν + 2k − 1 )( ν + 2k ) + ( ν + 2k ) − ν 2
2
+ ∑ k!(k + ν)! 2 = 0.
k =0 k =0

We separte the first term (k = 0) of the first integral from the summation, we get

1 h i  x ν
( ν − 1) ν + ν − ν2 +
ν! 2
∞ ∞
(−1)k h i  x ν+2k 4(−1)k  x ν+2k+2
∑ k!(k + ν)! ( ν + 2k − 1 )( ν + 2k ) + ( ν + 2k ) − ν 2
2
+ ∑ k!(k + ν)! 2 = 0.
k =1 k =0

As it is clear, the firest line vanishes and the second one becomes after (k = q + 1), as
∞ ∞
(−1)q+1 h i  x ν+2q+2 4(−1)k  x ν+2k+2
∑ ( q + 1) ! ( q + 1 + ν ) ! ( ν + 2q + 1 )( ν + 2q + 2 ) + ( ν + 2q + 2 ) − ν 2
2
+ ∑ k!(k + ν)! 2 = 0.
q =0 k =0

Then by changing the summation index is both summations q → n, k → n, we get

∞   
4(n + 1)(ν + n + 1) 4 x ν+2n+2
∑ (− 1 ) n
− +
(n + 1)!(n + 1 + ν)! n!(n + ν)! 2
= 0,
n =0

which clearly vanishes.

2- We have
g( x ) g0 ( x ) g( x ) g00 ( x ) g0 ( x ) 3g( x )
f ( x ) = √ , f 0 ( x ) = √ − 3/2 , f 00 ( x ) = √ − 3/2 + 5/2 ,
x x 2x x x 4x

then (8) becomes

g00 ( x ) g0 ( x ) 3g( x ) g0 ( x )
   
2 g( x ) g( x )
x √ − 3/2 + 5/2 + x √ − 3/2 + ( x2 − ν2 ) √ = 0
x x 4x x 2x x
 
1 1
√ x2 g00 ( x ) + ( x2 − ν2 + ) g( x ) = 0,
x 4

and thus
1
x2 g00 ( x ) + ( x2 − ν2 + ) g( x ) = 0.
4

In case of ν = 1/2, we got

g00 ( x ) + g( x ) = 0,

which has the solutions g( x ) = a cos(t) + b sin(t) or g( x ) = aeit + be−it .

================================================================================
Exercise 4: (15 pts)
¸
Here, we will be interested by the integral f (z)dz over the closed contour as shown in the figure, where
C
1
the complex function is given by f (z) = 2.
(z2 + 14 )
 4

1- Find the poles of f (z). What is the order of each pole.

2- Caluculate the residus at each pole according to the values of R > 1/2 and R < 1/2.
¸ ´ +R
3- In part C1 , the intergral can be written as f (z)dz = − R f ( x )dx, where x is real. Simplify the integral
C1
¸ ´ 2π
over the part C2 as f (z)dz = 0 Q(θ )dθ , i.e., find Q(θ ) (Hint: try with z = Rei` ).
C2
¸ ¸
4- Guess the value of the limit lim f (z)dz (Hint: you do not need to evaluate try f (z)dz); and deduce
R→∞C C2
´ +∞ dx
2
the integral −∞ 2 .
( x2 + 14 )
================================================================================
1- The function f (z) = 2 1 1 2 has the poles z1 = +i/2 and z2 = −i/2. In order to determine which
(z + 4 )
order has each pole, one has to evaluate the limits

i (z − 2i )m (z − 2i )m (z − 2i )m−2
lim (z − )m f (z) = lim  2 = lim 2  2 = lim 2 . (10)
2
 
z→ 2i z→ 2i z2 + 1 z→ 2i z − i z + i z→ 2i z + i
4 2 2 2

This limit is zero for m > 2, infinite for m < 2 and finite for m = 2. Therefore the pole ordr is 2. Similarly
for the pole z2 = −i/2,

i ( z + i ) m −2
lim (z + )m f (z) = lim  2 2 . (11)
z→− 2i 2 z→ 2i z − 2i

2- The residus calculation does nor depend on the considered contour. Then,
 
(2) i d i 1
a−1 ( ) = lim  (z − )2 

2 
2 z→ 2i dz 2 1
z2 + 4
" # " #
d 1 2
= lim = lim − = −2i.
z→ 2i dz (z + 2i )2 z→ 2i (z + 2i )3
 
(2) i d i 1
a−1 (− ) = lim  (z + )2 

2 
2 z→− 2 dz
i 2
z2 + 1
4
" # " #
d 1 2
= lim = lim − = 2i.
z→− 2i dz (z − 2i )2 z→− 2i (z − 2i )3
¸ ¸ ¸
3- We have f (z)dz = f (z)dz + f (z)dz. In the part C2 , we have z = Reiθ and dz = iReiθ dθ, then
C C1 C2
˛ ˆ π
eiθ
f (z)dz = iR  2 dθ,
0 1
C2 R2 e2iθ + 4
 5

iRe iθ
which means Q(θ ) = 2.
( R2 e2iθ + 14 )
iReiθ
¸
4- The function Q(θ ) vanishes if R → +∞, i.e., is means lim 2 = 0 ⇒ lim f (z)dz = 0. Then,
R→∞ ( R e2iθ + 14
2 ) R→∞C
2
for R → +∞, we have
ˆ +∞ ˆ +R
dx dx (2) i
2 = lim 2 = 2πi × a−1 ( ) = 2πi × (−2i ) = 4π,
R→∞ − R 2
 
−∞ x2 + 1
x2 + 1
4 4

since only the pole z = 2i belongs to the contour.

================================================================================

Bonus Qusetions:
´∞ ´π ´∞ x p log x
1- Calculate the integrals 0 cos( x2 )dx (5 pts), 0 cosn θdθ (3 pts); and 0 x 2 +1
dx (3 pts).
d2
2- In case of the functional F (y( x ), y0 ( x ), y00 ( x ), x ), where y0 ( x ) = d
dx y ( x ) and y00 ( x ) = dx2
y ( x ). Find the
Euler-Lagrange equation. (5 pts)

================================================================================
´∞ ´∞
1- We have the integral Q = 0 cos( x2 )dx and let’s assume the integral Z = 0 sin( x2 )dx, would be very
´∞ 2
useful in our case. Then, let’s estimate the integral 0 e−z dz over the contour

2
Then, since the function f (z) = e−z is analytic, then
˛ ˛ ˛ ˛
0 = f (z)dz = f (z)dz + f (z)dz + f (z)dz. (12)
C C1 C2 C3
¸
For the contour part C2 , the integral vanishes once R → ∞, while it is given over the part C1 by f (z)dz =
C1
´∞ 2
√
0 e− x dx = 2π . Over the part C3 , one uses z = √t (1 + i ) the integral can be written in fuction of the
2
integrals Q and Z as
˛ ˆ 0 ˆ 0 2
ˆ ∞
− t2 (1+ i )2 −t2 (1+2i) (1√
+ i) 2 (1 + i ) (1 + i )
f (z)dz = e (1 + i )dt = e dt = − e−it √ dt = − √ ( Q − iZ ).
∞ ∞ 2 0 2 2
C3

Then, opne writes

√
π (1 + i )
− √ ( Q − iZ ) = 0,
2 2
 6

which implies
√
π − ( Q + Z ) − i [− Z + Q] = 0,
√
i.e., Z = Q = 2π .
´π
For the integral Rn = 0 cosn θdθ it is vanishing for odd values for n R2n+1 = 0, while for even values
´π ´ 2π
R2n = 0 cos2n θdθ = 12 0 cos2n θdθ, one put z = ei` and − zi dz = dθ, then
˛  2n
i 1 dz
R2n = − z+ ,
22n+1 z z
C

where the contour C represents the unit circle. Then, it gets simplified as
˛
2n ˛
i z2 + 1 i 2n
(2n)! dz π (2n)!
R=−
22n+1 z 2n + 1
dz = − 2n+1
2
∑ k!(2n − k)! z2n−2k+1
=
22n n!n!
.
k =0
C C

This function has the z = 0, with the roder 2n + 1. Then,

 2n  2n   n  n 
i 1 d 2 π d 2 π π
R = − n+1 2πi lim z +1 = 2n lim z +1 = n n! = n .
2 (2n)! z→0 dz2n 2 (2n)! z→0 dzn 2 n! 2

´∞ x p log x ¸ z p log z
For the integral 0 x 2 +1
dx, we consider the integral z2 +1
dz over the contour shown in the figure
C
below. Here, we divide the contour into four parts: C1 for z = x for x from 0 to ∞, C2 for z = xe2iπ with x
from ∞ to 0, C3 a full round with z = Rei` with θ from 0 to 2π and R → ∞; and C4 a full round with z = eei`
with θ from 2π to 0 and e → 0.

´ ∞ x p log x
The integral over C1 gives the required value I = 0 x2 +1 dx; and over C2 the integral gives
´ 0 x p e2iπ p log( xe2iπ ) ´ ∞ x p e2iπ p (log( x)+2iπ ) ´∞ p
∞ ( xe2iπ )2 +1
dx = − 0 x 2 +1
dx = −e2iπ p I − 2iπe2iπ p 0 x2x+1 . The integral over C3 and
C4 tend to zero for R → ∞ and e → 0, repectively.
´∞ p
One can find that 0 x2x+1 = π
pπ , then
2 cos( 2 )
˛
z p log z 
2iπ p
 π 2 e2iπ p
2
dz = I 1 − e −i pπ = 2πi ( Res (i ) + Res (−i ))
z +1 cos( 2 )
C
!
eiπ p/2 (iπ/2) e3iπ p/2 (3iπ/2) iπ 2  iπ p/2 
= 2πi + = e − 3e3iπ p/2
2i −2i 2
 7

then
  iπ 2 −iπ p  iπ p/2 
e−iπ p I 1 − e2iπ p = e e − 3e3iπ p/2
2
  iπ 2  
− I eiπ p − e−iπ p = e−iπ p/2 − 3eiπ p/2
2
thus
π 2 sin( pπ/2)
I= .
4 cos2 ( pπ/2)

================================================================================