Special functions

i
(Exercises with Solutions)
un
Master 2, PDEs & Applications
no
Pr. Dr. Abdelaziz Mennouni, Department of Mathematics,
University of Batna 2, Algeria, E-mail: [email protected]
en

February 13, 2021

M
 2

M
en
no
un
i
Contents

1 Special functions 5
1.1 Solutions . . . . . . . . . . . . . . . . . . . . . . . 6

i
un
no
en
M

3
4 CONTENTS

i
un
no
en
M
Chapter 1

Special functions

i
un
no
en
M

5
6 CHAPTER 1. SPECIAL FUNCTIONS

1.1 Solutions
Solution 1.1. 1. We have
1.2.3 . . . n
Γ(x + 1) = lim nx+1
n→∞ (x + 1)(x + 2)(x + 3) . . . (x + n + 1)
 
xn 1.2.3 . . . n
= lim × nx
n→∞ x + n + 1 x(x + 1)(x + 2) . . . (x + n)
= xΓ(x).

2. On the other hand,

Γ(1) = 1!.

i
On the one hand,

Γ(n + 1) =
=
nΓ(n)
un
n(n − 1)Γ(n − 1)
no
.
.
en

.
= n(n − 1)(n − 2)(n − 3) . . . + .2.1
= n!.
M

3. It is evident that
Γ(5) = 4! = 24.
Γ(7) = 6! = 720.

4. Note that    
22 17 17
Γ = Γ .
5 5 5
This shows that

    
22 17 17 17
Γ /Γ = Γ
5 5 5 5
1.1. SOLUTIONS 7

22


Γ 5 17
17

= .
Γ 5
5
5. Also, we have
   
23 19 19
Γ = Γ
4 4 4
 
19 15 15
= Γ
4 4 4
 
19 15 11 11
= Γ .
4 4 4 4

i
Thus,
Γ
Γ
11
4

23


=
un 1
19 15 11
no
4 4 4 4
64
= .
3135
en

Solution 1.2. 1. Since

nx = ex log n
M

= ex(log n−1−1/2−...−1/n) ex+x/2+...+x/n ,

we get
1 1 2 n x ex(log n−1−1/2−...−1/n ex+x/2+...+x/n)
Γn (x) = ... n = .
x x+1x+2 x+n x(1 + x)(1 + x/2) . . . (1 + x/n)

2. We have
1 ex ex/2 ex/n
Γn (x) = ex(log n)−1−1/2−...−1/n +...+ .
x 1 + x 1 + x/2 1 + x/n
8 CHAPTER 1. SPECIAL FUNCTIONS

3. Note that
1
γn − γn+1 = log(n + 1) − log(n) −
  n+1
1 1
= log 1 + −
n n+1
 
1 1 1
= 1− − 2 +0 3
n + 1 2n n
 
1 1 1
= − 2 +0 3
n(n + 1) 2n n
 
n−1 1
= + 0 .

i
2n2 (n + 1) n3

un
This shows that, γn − γn+1 ≥ 0 for all n ≥ 1.
no
4. Since
" #
1 ex ex/2 ex/n
Γ(x) = lim ex(log n)−1−1/2−...−1/n +...+ ,
en

n→∞ x 1 + x 1 + x/2 1 + x/n

we obtain
x
e−γx +∞ e n
M

Γ(x) = ∏ x .
x n=1 (1 + )
n
Solution 1.3. We have
 
−1 1 1
ψ(x) = −γ + ∑ − .
x n≥1 n x + n

By deriving the formula, we get

1
ψ 0 (x) = ∑ (x + n)2 .
n≥0
1.1. SOLUTIONS 9

Solution 1.4. It follows from Stirling’s Lemma that, for Re (z) > 1
and |z| = R large enough, we have
1
≤ ez z−z+1/2
Γ(z)
≤ eR RR+1/2
1
≤ RR eR R 2 .
1
≤ c1 RR c1 := eR R 2 .

For Re (z) < 1 and |z| = R, large enough:

i
1
Γ(z)
≤

≤ e|z|+1 (1 + R)R+1/2
≤ c2 RR .
un
ez−1 (1 − z)−z+1/2 sin(πz)
no
Solution 1.5. We have
γz
 z −z
ξ (z) = ze ∏ 1 + e n,
en

n≥1 n

so that
M

z −x x x
|ξ (z)| = |z|eγx ∏ 1 + e n ≥ |x|eγx ∏ 1 + e− n ,
n≥1 n n≥1 n

with x = Re (z) = 1. Note that since

x −x  x x   x x
log ∏ 1 + e n = ∑ log 1 + − = ∑ log 1 + −
n≥1 n n≥1 n n n≥1 n n
1  x 2 x2
≥− ∑ ≥ − ζ (2) > −1,
2 n≥1 n 2

we have
1
|ξ (z)| ≥ |x|eγx e−3 > .
2
10 CHAPTER 1. SPECIAL FUNCTIONS

Solution 1.6. We use integration by parts

Z ∞
Γ(z + 1) = e−t t z dt
0
 z −t +∞ Z ∞ −t z−1
= −t e 0 + z e t dt
0
= zΓ(z).

Solution 1.7. 1. Letting u = e−t , so that t = − log u, that is dt =

− du
u , and therefore
t = 0 =⇒ u = 1,

i
t = +∞ =⇒ u = 0.
√ un
2. Letting t = u2 we get u = t that is dt = 2udu, moreover
t = 0 =⇒ u = 0,
no
t = +∞ =⇒ u = +∞.
Solution 1.8. 1. We have
Z +∞ Z +∞
en

2 +v2 )
2
I =4 e−(u dudv.
0 0
Evaluate a double integral using a change of variables when
we substitute u = r cos θ , v = r sin θ , we get
M

Z π/2 Z +∞
2
I 2
= −2 − 2re−r drdθ
0 0
Z π/2 Z +∞  0
−r2
= −2 e drdθ
0 0
Z π/2
= −2 − dθ
0
= π.

Hence √
I= π.
1.1. SOLUTIONS 11

2.
  Z +∞
1 2
Γ = e−u du
2 0
√
= π.

3.
   
3 1 1
Γ = Γ
2 2 2
√
π

i
= .
2

Γ

−1

=
Γ
un
−1
1
2


no
2 2
 
1
= −2Γ
2
√
en

= −2 π.
M

−1
 

−3 Γ 2
Γ = −3
2 2
4√
= π.
3

Solution 1.9. We will prove this by induction. For n=0, on the one
hand, we have √
πΓ(1) √
= π,
20 Γ(1)

√
on the other hand, Γ 21 = π.
12 CHAPTER 1. SPECIAL FUNCTIONS

Let n ∈ N, we assume that

√
1 πΓ(2n + 1)
Γ(n + ) = 2n .
2 2 Γ(n + 1)
We will prove that
√
3 πΓ(2n + 3)
Γ(n + ) = 2n+2 .
2 2 Γ(n + 2)
We have
3 1 1
Γ(n + ) = (n + )Γ(n + )

i
2 √ 2 1 2
=

=
√
un
π(n + 2 )Γ(2n + 1)
22n Γ(n + 1)
π(2n + 1)Γ(2n + 1)
no
22n+1 Γ(n + 1)
√
πΓ(2n + 2)
= 2n+1
2 Γ(n + 1)
en

√
π(2n + 2)Γ(2n + 2)
= 2n+1
2 2(n + 1)Γ(n + 1)
√
M

πΓ(2n + 3)
= 2n+2 .
2 Γ(n + 2)

Solution 1.10. 1. We have

Z+∞ Z+∞
Γ(s)Γ(1 − s) = us−1 v−s e−u e−v dv du
0 0
Z+∞ Z+∞
= dv v−s e−v ( us−1 e−u du),
0 0
1.1. SOLUTIONS 13

u
under the change of variables w = , we obtain
v
Z+∞ Z+∞
Γ(s)Γ(1 − s) = dv e−v ( us−1 e−uv du)
0 0
Z+∞ Z+∞
= du us−1 dv e−v e−uv )
0 0
Z+∞ Z+∞
= du us−1 ( dv e−v(1+u) )

i
0 0

=
Z+∞

0
Z1
us−1
1+u
un
du

Z+∞ s−1
no
us−1 u
= du + du.
1+u 1+u
0 1
en

1
Again, under the change of variables w = in the second inte-
u
gral, we get
M

Z1 s−1 Z1
u u−s
Γ(s)Γ(1 − s) = du + du.
1+u 1+u
0 0

Thus,
Γ(s)Γ(1 − s) = ϕ(s) + ϕ(1 − s).

2. For all n ∈ N, we have

n−1
1 k k (−1)n un
= ∑ (−1) u + .
1 + u k=0 1+u
14 CHAPTER 1. SPECIAL FUNCTIONS

Hence
n−1 Z1 Z1 n+s−1
u
ϕ(s) = ∑ (−1)k uk+s−1 du + (−1)n
1+u
du
k=0 0 0
n−1 Z1
(−1)k un+s−1
= ∑ k + s + (−1)n 1+u
du.
k=0 0

But
Z1 n+s−1 Z1
u 1 1
0≤ du ≤ un+Re(s)−1 du = ≤ → 0.

i
1+u n + Re(s) n n→+∞
0 0

Thus,

ϕ(s) = ∑
+∞
(−1)n
,
un
pour touts ∈ C, avec0 < Re(s) < 1.
no
n=0 n + s

3. We use the development of ϕ(s), we obtain

en

+∞
(−1)n +∞ (−1)n
Γ(s)Γ(1 − s) = ∑ +∑
n=0 n + s n=0 n + 1 − s
+∞
(−1)n +∞ (−1)n+1
M

= ∑ n+s + ∑ n−s
n=0 n=1
+∞
(−1)n +∞ (−1)n
= ∑ n+s + ∑ s−n
n=0 n=1
1 +∞ 2(−1)n s
= +∑ 2 .
s n=1 s − n2

4. For t = 0, we have
" #
+∞
sin πs 1 2(−1)n s
1= +∑ 2 ,
π s n=1 s − n2
1.1. SOLUTIONS 15

so that
1 +∞ 2(−1)n s
Γ(s)Γ(1 − s) = + ∑ 2
s n=1 s − n2
π
= .
sin πs
Solution 1.11. It is easy to show that
√
 
1
Γ = π,
2
    √
1 2 2π 3
= ,

i
Γ Γ
3 3 3

Solution 1.12.
Γ
   
1
4
Γ
3
4
√
= π 2. un
1. We will prove this by induction.
no
For n = 0, we have
Z 1  z 1
z−1 s 1
J0 (z) = s ds = = .
en

0 z 0 z

Let n ≥ 0, assume that

M

n!
Jn (z) = .
z(z + 1) . . . (z + n)
We have
Z 1
Jn+1 (z) = sz−1 (1 − s)n+1 ds
0 z 1 Z 1 z
s n+1 s
= (1 − s) + (n + 1)(1 − s)n ds
z 0 0 z
n+1
= 0+ Jn (z + 1)
z
(n + 1)n!
= .
z(z + 1) . . . (z + n)(z + n + 1)
16 CHAPTER 1. SPECIAL FUNCTIONS

2. Consider the sequence of functions

 t n z−1
gn (t) = 1 − t χ]0;n[ (t).
n
It follows
(a) gn (t) −→ e−t t z−1 χ]0;+∞[ .
n→+∞
(b) Since
1 − u ≤ e−u for all u ∈ [0; 1],
we get
|gn (t)| ≤ e−t t Re (z)−1 .

i
Following the Lebesgue’s theorem,

Γ(z) = lim

For t = ns, we obtain,

n→+∞ 0
un
Z n
t z−1

1−
t n
n
dt.
no
Z 1
Γ(z) = lim n z
sz−1 (1 − s)n dt = lim nz Jn (z).
n→+∞ 0 n→+∞

Solution 1.13. 1. We have

en

Γ(y)Γ(x)
B(y, x) =
Γ(y + x)
Γ(x)Γ(y)
M

=
Γ(x + y)
= B(x, y)

2. We consider the following change of variables u = sin2 θ , so

that dt = 2 cos θ sin θ , and that,
Z π
2
B(x, y) = 2 (sin2 θ )(x−1) (cos2 θ )(x−1) sin θ cos θ dθ
0
Z π
2
= 2 sin(2x−1) θ cos(2y−1) θ dθ .
0
1.1. SOLUTIONS 17

Solution 1.14. 1. We have

Γ(x + 1)Γ(y)
B(x + 1, y) =
Γ(x + y + 1)
xΓ(x)Γ(y)
=
(x + y)Γ(x + y)
x
= B(x, y).
x+y

i
2. Also, we have
un
B(x, y + 1) = B(y + 1, x)
y
no
= B(y, x)
y+x
y
= B(x, y).
x+y
en
M

3. It follows that

B(x, y) = B((x − 1) + 1, y)
x−1
= B(x − 1, y).
x+y−1

Solution 1.15. We use the change of variables s = t 5 , so that,

1 4
dt = s− 5 ds.
5
18 CHAPTER 1. SPECIAL FUNCTIONS

Thus, we obtain
Z 1
tdt 1 1 1
Z
1 4
√ = s 5 (1 − s)− 2 s− 5 ds
0 1 − t5 5 0
1 1 −3
Z
1
= s 5 (1 − s)− 2 ds
5 0
1 1 2 −1
Z
1
= s 5 (1 − s) 2 −1 ds
5 0
1 2 1
= B( , ).
5 5 2

Solution 1.16. We use the change of variables s = t 4 , hence

i
and hance
un
1 3
dt = s− 4 ds,
4
no
Z 1
tdt 1 1 1
Z
1 3
√ = s 2 (1 − s)− 2 s− 4 ds
0 1 − t5 4 0
1 1 −1
Z
1
= s 4 (1 − s)− 2 ds
en

4 0
1 1 3 −1
Z
1
= s 4 (1 − s) 2 −1 ds
5 0
M

1 3 1
= B( , ).
4 4 2

Solution 1.17. 1. We have

Z 1
1 1 1 1
B( , ) = s− 2 (1 − s)− 2 ds
2 2 0
Z 1
du
= 2 √ under the change of variables s = u2
0 1 − u2
= 2 arcsin 1
= π.
1.1. SOLUTIONS 19

But
Γ 12 Γ 12



1 1
B( , ) =
2 2 Γ(1)
 
1
= Γ2 .
2

Thus
√
 
1
Γ = π.
2

i
2. We have

Γ
 
1
2
=
Z +∞

0
Z +∞
1
e−s s− 2 ds
2
e−u du
un
no
= 2 under the change of variables s = u2
0
Z +∞
2
= e−u du.
−∞
en

This implies that

M

Z +∞
2 √
e−u du = π.
−∞

Solution 1.18. It is evident that

+∞
zk
E1,2 (z) = ∑
k=0 Γ(k + 2)
1 +∞ zk+1
= ∑ (k + 1)!
z k=0
ez − 1
= ,
z
20 CHAPTER 1. SPECIAL FUNCTIONS

+∞
zk
E1,3 (z) = ∑
k=0 Γ(k + 3)
1 +∞ zk+2
= 2∑
z k=0 (k + 2)!
ez − z − 1
= ,
z2

and

i
+∞
E2,1 (z) = ∑ un zk
k=0 Γ(2k + 1)
= cosh z.
no
Solution 1.19. We have
en

+∞
xk
M

Eα,β (x) = ∑
k=0 Γ(αk + β )
+∞
xk+1
= ∑
k=−1 Γ(α(k + 1) + β )
+∞
xxk
= ∑
k=−1 Γ(αk + (α + β ))
+∞
1 xk
= +x∑
Γ(β ) k=0 Γ(αk + α + β )
1
= + xEα,α+β (x).
Γ(β )
1.1. SOLUTIONS 21

Solution 1.20. Proceeding as before, we obtain

+∞ +∞
d β xk d xk
β Eα,β +1 (x) + αx [Eα,β +1 (x)] = ∑ Γ(αk + β + 1) + ∑ αx dx Γ(αk + β + 1)
dx k=0 k=0
+∞
(αk + β )xk
= ∑
k=0 Γ(αk + β + 1)
+∞
xk
= ∑
k=0 Γ(αk + β )
= Eα,β (x).

i
Solution 1.21. We remark that un
no
+∞
(−1)k+1 xαk
ϑα (x) := ∑ Γ(αk + 1) .
k=1
en

Thus, we obtain

d +∞ (−1)k+1 αkxαk−1
M

d
ϑα (x) = ∑ Γ(αk + 1)
dx dx k=1
+∞
(−1)k+1 xαk−1
= ∑
k=1 Γ(αk)
+∞
(−1)k xαk+α−1
= ∑
k=0 Γ(αk)
+∞
α−1 (−1)k xαk
= x ∑
k=0 Γ(αk)
= xα−1 Eα,α (−xα ).
22 CHAPTER 1. SPECIAL FUNCTIONS

Solution 1.22. We have

+∞ 
zn
Z +∞ 
λ ,p 1 p
λ +n−1 −t− t 1
Eβ ,γ (z) = ∑ t e dt
n=0 Γ(λ ) 0 Γ(λ )Γ(β n + γ) n!
+∞
(tz)n
Z +∞
1 λ −1 −t− pt
= t e ∑ Γ(β n + γ)n! dt.
[Γ(λ )]2 0 n=0
Z +∞
1 p
= t λ −1 e−t− t Eβ ,γ (tz)dt.
[Γ(λ )]2 0

i
un
no
en
M