Introduction And Applications of Gamma

Functions

A project report
submitted by

Akshay Kumar Ranwa

(IVR No: 201700027139)

under the supervision of

Dr.
Contents

1 Introduction 3

2 Mathematical preliminaries and definitions 4

2.1 The factorial function. . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2 Convex function. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.3 The gamma function. . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4 Definition form Gauss. . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.5 Weierstrass formula. . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3 Properties of gamma function. 10

3.1 Complement Formula. . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.2 Stirling formula. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.3 Relation between beta function gamma function. . . . . . . . . . . . 13
3.4 Multiplication formula. . . . . . . . . . . . . . . . . . . . . . . . . . 14

4 Incomplete gamma function. 16

4.1 Upper incomplete gamma function. . . . . . . . . . . . . . . . . . . . 16
4.2 Lower incomplete gamma function. . . . . . . . . . . . . . . . . . . . 17
4.3 Properties of incomplete gamma function . . . . . . . . . . . . . . . . 17
4.3.1 Integration by part. . . . . . . . . . . . . . . . . . . . . . . . . 17
4.3.2 Derivative of upper incomplete gamma function. . . . . . . . . 18
4.4 Inequality hold by incomplete gamma function . . . . . . . . . . . . . 18
4.4.1 Some inequality of gamma function . . . . . . . . . . . . . . . 20

5 Generalization of gamma function 22

5.1 Generalize gamma function in Gaussian form . . . . . . . . . . . . . . 22

1
 5.2 Eulerian Form and Convergence of Generalize gamma function . . . . 24
5.3 Generalize gamma function of Weierstrassian Form, and Derivatives . 25
5.4 Integral for the Asymptotic Constant. . . . . . . . . . . . . . . . . . 26
5.5 The Associated Periodic Functions of . . . . . . . . . . . . . . . . . . 27

6 Generalization of gamma function 29

6.1 Generalize gamma function in Gaussian form . . . . . . . . . . . . . . 29
6.2 Eulerian Form and Convergence of Generalize gamma function . . . . 31
6.3 Generalize gamma function of Weierstrassian Form, and Derivatives . 32
6.4 Integral for the Asymptotic Constant. . . . . . . . . . . . . . . . . . 33
6.5 The Associated Periodic Functions of . . . . . . . . . . . . . . . . . . 34

7 Extended incomplete gamma function 36

7.1 Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
7.1.1 Corollary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

8 Application of gamma function 41

8.1 THE GAMMA FUNCTION AND BALLS . . . . . . . . . . . . . . . 41
8.1.1 Volume Of The N-Dimensional Ball. . . . . . . . . . . . . . . 41
8.1.2 Derivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
8.2 Psi and Polygamma function . . . . . . . . . . . . . . . . . . . . . . . 43
8.2.1 The Computation Of Infinite Sums. . . . . . . . . . . . . . . . 44
8.3 Application in Probability . . . . . . . . . . . . . . . . . . . . . . . . 46

2
Chapter 1

Introduction

18 century Eular‘s introduced the Gamma function in his aim to generalzing the fac-
torial function of the nature number, that is the problem of finding an expression that
has the value n! for any positive integer n. Because of it‘s great importance , it is
studies by many of great mathematician.
Christian goldbach‘s and Daniel Bernoulli also worked in on the factorial function
and found a integral representation of n! for n is positive integer. James Stirling also
attempted for a continuous expression for the factorial and which is known as stirling
formula .
Gauss, Weierstreass and Legendra is also work on the gamma function and some
properties of gamma function.
The gamma function is represent by greek alphabet Γ and its value for ϑ > 0

Γ(ϑ) = (ϑ − 1)!

Gamma function is analytic everywhere except 0 and negative integer‘s so it is a

meromorphic function .
so its residue at =-k
(−1)k
Res(ϑ=−k) Γ(ϑ) =
k!

the gamma is belong to a to a class of transcendental function. Application of gamma

function in various area like analytic physics, engineering and probability.

3
Chapter 2

Mathematical preliminaries and

definitions

2.1 The factorial function.

A function which is represent as

ωn = ω(ω + 1)(ω + 2) · (ω + n); x ̸= 0 and ω0 = 1 (2.1)

is called factorial function . y = wn function as pochhamner symbol.

2.2 Convex function.

A function f (ϑ) define in interval (a,b) and let ϑ1 , ϑ2 in interval (a,b) ,then difference
quotient.
f (ϑ1 ) − f (ϑ2
ϕ(ϑ1 , ϑ2 ) = = ϕ(ϑ2 , ϑ1 )
ϑ1 − ϑ2

for ϑ1 , ϑ2 and ϑ3 in interval in (a,b), then difference quotient

ϕ(ϑ1 , ϑ3 ) − ϕ(ϑ2 − ϑ3 )
ϕ(ϑ1 , ϑ2 , ϑ3 ) =
ϑ1 − ϑ3
and if ϕ(ϑ1 ,ϑ2 ,ϑ3 )≥0 for ϑ1 < ϑ2 < ϑ3 then f (ϑ) is called convex function.
and A weakly convex function F (ϑ).

4
 ϑ1 + ϑ2 + ... + ϑn 1
F( ) = (F (ϑ1 ) + F (ϑ2 ) + ...F (ϑn ))
n n

2.3 The gamma function.

Definition 1
Definition introduced by Eular to define factorial function in term of infinite product.

∞
Y (1 + n1 )ϑ
Γ(ϑ + 1) = ϑ! = (2.2)
n=1
1 + nϑ

Definition 2
Gamma function in term of defined integral for ϑ > 0

Z1
Γ[ϑ] = (−log(u))ϑ−1 du (2.3)
0

Now using change variable and take t = log(u) so it can convert most useful form of
gamma function.
Z∞
Γ[ϑ] = uϑ−1 e−u du (2.4)
0

then

Z ∞
Γ(1) = e−u du = 1
0

and for ϑ > 0

Z ∞
Γ[ϑ + 1] = un e−u du
0
Z ∞
= [−uϑ e−u ]∞
0 +u uϑ−1 e−u du = ϑΓ[ϑ]
0

then it is obvious that by using equation

5
 Γ(ϑ + n) = (ϑ + n − 1)(ϑ + n − 2)...(ϑ + 1)ϑΓ(ϑ) (2.5)

it is only for ϑ > 0 then for extended all real number. If ϑ lies in interval (−n, −n+1).

1
Γ(ϑ) = Γ(ϑ + n) (2.6)
ϑ(ϑ + 1)...(ϑϑ + n − 1)
some value of gamma function
1. Γ(1 − ϑ) = −ϑΓ(−ϑ)
√
2. Γ( 21 ) = π
1·3·5...(2n−1) √
3. Γ(n + 21 ) = 2n
π
1·3·5...(2n−1)
4. Γ(n + 31 ) = 2n
Γ( 31 )

Theorem 1.1 A function f (t) : (0, ∞) → (0, ∞) is gamma function if it satisfied

following condition .
1.f (1) = 1.

2.f (t + 1) = tf (t).

3. log f (t) is a convex f unction.

Proof. let f (t) be a function that satisfies all conditions. Then Eq.[2.4] which s
is valid because of condition (1), and f (ϑ) = (ϑ − 1)! for all integers for all positive
integers because of condition (1).
we need to show that f (t) agrees with Γ(t) on the interval 0 < t ⩽ 1. If this is the
case, then f (t) must agree with Γ(t) everywhere because of condition (2). Let t be a
real number, 0 < t ⩽ 1, and ϑ and integer ⩾ 2.
so inequality

log f (−1 + ϑ) − log f (ϑ) log f (t + ϑ) − log f (ϑ) log f (1 + ϑ) − log f (ϑ)
⩽ ⩽
(−1 + ϑ) − ϑ (t + ϑ) − ϑ (1 + ϑ) − ϑ

expresses the monotonic growth of the difference quotient for particular values, and

6
is therefore valid because of condition (2). Since f (ϑ) = (ϑ − 1)!, we have

log f (t + ϑ) − log(ϑ − 1)!

log(ϑ − 1) ⩽ ⩽ log(ϑ)
t

or
log(ϑ − 1)t (ϑ − 1)! ⩽ log f (t + ϑ) ⩽ log ϑ2 (ϑ − 1)!

But the logarithm is a monotonic function; so

(ϑ − 1)t (ϑ − 1)! ⩽ f (t + ϑ) ⩽ ϑ2 (ϑ − 1)!

the help of Eq. (2.4), we get the following inequality for f (t) itself:

(ϑ − 1)t (ϑ − 1)! ϑ2 (ϑ − 1)!

⩽ f (t) ⩽
ϑ(ϑ + 1) · · · (t + ϑ − 1) t(t + 1) · · · (t + ϑ − 1)
ϑt ϑ! t+ϑ
=
t(t + 1) · · · (t + ϑ) ϑ

Since it is true for all ϑ ⩾ 2, we can change ϑ by ϑ + 1 on the left side. Thus

ϑt ϑ! ϑ2 ϑ! t+ϑ
⩽ f (t) ⩽
t(t + 1) · · · (t + ϑ) t(t + 1) · · · (t + ϑ) ϑ
.
An easy calculation gives the inequality

ϑ ϑ2 ϑ!
f (t) ⩽ ⩽ f (t)
t+ϑ t(t + 1) · · · (t + ϑ)

As ϑ approaches infinity, we get

ϑ2 ϑ!
f (t) = lim
ϑ→∞ t(t + 1) · · · (t + ϑ)

But Γ(t) is also a function that satisfies our three conditions. Hence the relation we
have just derived is still valid if we put Γ(t) instead of f (t) on the left side.

7
2.4 Definition form Gauss.
Gauss studied Eular infinite product and introduced ϑ > 0 then definition.

p! pϑ
Γp (ϑ) = (2.7)
ϑ(ϑ + 1)...(ϑ + p)
then

Γ(ϑ) = lim Γp (ϑ) (2.8)

p→0

then clearly

p! p
Γp (1) = p=
1(1 + 1)(1 + 2)...(1 + p) p+1
Γ(1) = lim Γp (1) = 1
p→∞

and
p! pϑ+1 p
Γp (ϑ + 1) = = ϑΓp (ϑ)
(ϑ + 1)(ϑ + 2)...(ϑ + p + 1) ϑ+p+1

then

Γ(ϑ + 1) = lim Γp (ϑ + 1) = ϑΓ(ϑ)

p→∞

And by the definition gamma function does not exist on negative integer so id follow
all properties of gamma function.

2.5 Weierstrass formula.

for ϑ be any real number except negative integers then infinite product.

∞
1 γϑ
Y ϑ −ϑ
= ϑe (1 + )e p (2.9)
Γ(ϑ) p=1
p

γ is eular‘s constant and his value is define γ=0.5772156...

8
 proof. let a relation that pϑ =eϑ ln(p) .

1 ϑ ϑ ϑ
eϑ ln(p) = eϑ(ln(p)−1− 2 −...− p ) eϑ+ 2 +... p

put value eq[2.6] when

1 1 2 p ϑ(ln(p)−1− 12 −...− ϑp ) ϑ+ ϑ2 +... ϑp

Γp (ϑ) = ... e e
ϑ +1 ϑ + 2 p + 1

n
1 ϑ(ln(p)−1− 12 −...− ϑp ) Y ϑ ϑ
Γp (ϑ) = e (1 + )−1 e p
ϑ p=1
p

then

∞
1 γϑ
Y ϑ −ϑ
= ϑe (1 + )e p
Γ(ϑ) p=1
p

and the value of

1 ϑ
γ = − lim (ln(p) − 1 − − ... − ) = 0.5772156649...
n→∞ 2 p

9
Chapter 3

Properties of gamma function.

3.1 Complement Formula.

It is the relation of Gamma function between complementary values ϑ and 1 − ϑ.it is
valid when ϑ and ϑ − 1 are non negative integer.
Using equation [2.8] for ϑ
∞
1 Y ϑ −ϑ
=γϑ (1 + )e p
Γ(ϑ) p=1
p

put for value for -ϑ

∞
1 Y ϑ −ϑ
= −ϑe−γϑ (1 − )e p
Γ(−ϑ) p=1
p

multiply value of gamma ϑ and -ϑ then

∞ ∞
1 1 γϑ
Y ϑ −ϑ −γϑ
Y ϑ −ϑ
= −ϑ (1 + )e e
p (1 − )e p
Γ(ϑ) Γ(−ϑ) p=1
p p=1
p

using the functional equation Γ(−ϑ) = − Γ(1−ϑ)

ϑ

then
∞
1 1 Y ϑ2
= ϑ (1 − 2 )
Γ(ϑ) Γ(1 − ϑ) p=1
p

we know that

10
 ∞
Y ϑ2
sin(πϑ) = πϑ (1 − )
p=1
p2

then
π
Γ(ϑ)Γ(1 − ϑ) = (3.1)
sinπϑ
it is complement formula.
let example and using complement formula and take ϑ= 21 then.

1 1 π
Γ( )Γ(1 − ) =
2 2 sin π2

then

1 √
Γ( ) = π.
2

3.2 Stirling formula.

If we studies gamma function for ϑ tends to infinity then we using gamma the stirling
formula .so if ϑ is integer and it goes to infinity then asymptotic formula.

√
Γ(ϑ + 1) = ϑ! ∼ 2πϑϑϑ e−ϑ (3.2)

we observe weaker approximation under the curve log(t) with t belong to [1,ϑ] here ϑ
is very large. we know that

ϑ
X
log(ϑ!) = log(m)
m=1

using area approximation

ϑ
X Z ϑ
log(m) = log(u)du
m=1 1

Z ϑ
log(t)dt = ϑlog(ϑ) − ϑ + 1
1

11
solving integration

n−1
X log(m) + log(m + 1)
= + Rϑ
m=1
2

1
= log((ϑ − 1)!) + log(ϑ) + Rϑ
2
becauseRϑ = O(1) then

1
log(ϑ!) ≈ ϑlog(ϑ) + log(ϑ) − ϑ + c
2
then weaker result
√
ϑ! ≈ ec ϑϑ ϑe−ϑ .

Stirling‘s formula is very important

√
( 2, π, c) Suppose we will write

√  a1 a2 
ϑ! = 2πϑϑϑ e−ϑ 1 + + 2 + ···
ϑ ϑ

it is also valid for ϑ + 1

 
p (ϑ+1) −(ϑ+1) a1 a2
(ϑ + 1)! = 2π(ϑ + 1)(ϑ + 1) e 1+ + + ...
ϑ + 1 (ϑ + 1)2

but we also have (ϑ + 1)! = (ϑ + 1)ϑ!

√ ϑ −ϑ
 a1 a2 
(ϑ + 1)! = (ϑ + 1) 2πϑϑ e 1+ + 2 + ···
ϑ ϑ

We now compare relations both equation when ϑ become large This gives .

 ϑ+1/2  
 a1 a2  1 −1 a1 a2
1+ + 2 + ... = 1+ e 1+ + + ...
ϑ ϑ ϑ ϑ + 1 (ϑ + 1)2
1 13 1
a1 a2 − a1 + 12 a − 2a2 + a3 − 12
22 1
=1+ + + + ...
ϑ ϑ2 ϑ3

then
1
−a1 + =0
12

12
 1 1
a12 − 2a1 − =0
13 12
So using equation‘s we have found, by elementary mean,
then
1 1
: a1 = 12 , a2 = 288 , and more
so famous Stirling‘s asymptotic formula for ϑ >0.

√
 
ϑ −ϑ 1 1 139 571
Γ(ϑ + 1) = 2πϑϑ e 1+ + − − ··· (3.3)
12ϑ 288ϑ2 51840ϑ3 2488320ϑ4

3.3 Relation between beta function gamma func-

tion.
for c > 0and d > 0 then beta is define.

Z 1
B(c, d) = uc−1 (1 − u)d−1 du = B(d, c). (3.4)
0

let u=sin(t) ant putting in equation [8] then

Z π
2
B(c, d) = 2 sin(2c−1) t cos(2d−1) (t)dt
0

for finding relation between gamma function then for c>0,d>0.

Z ∞ Z ∞
2c−1 −u2 2
Γ(c)Γ(d) = 2 u e du 2 v 2c−1 e−v dv
0 0

let u=r cos(t),v=r sin(t),then

π
Z ∞ Z
2 2
Γ(c)Γ(d) = 4 e−r r2(c+d)−1 sin2c−1 tcos2d−1 tdrdt
0 0

π
Z ∞ Z
2
−r2 2(c+d)−1
Γ(c)Γ(d) = 2 e r dr sin2c−1 tcos2d−1 dt
0 0

13
 Γ(c)Γ(d) = Γ(c + d)B(c, d)

so
Γ(c)Γ(d)
B(c, d) = (3.5)
Γ(c + d)

so it is relation of gamma function beta function .

Some value of beta function using gamma function.
1
1. let c = 2
and d= 12 then value of function beta function.

1 1 Γ( 21 )Γ( 12 )
B( , ) = =π
2 2 Γ(1)

2. let c= 41 and d= 34 then b

1 3 Γ( 1 )Γ( 34 ) √
B( , ) = 4 =π 2
4 4 Γ(1)

3. for any c ≥ 1 and d ≥ 1.

(c − 1)!(d − 1)!
B(c, d) = .
(c + d − 1)!

3.4 Multiplication formula.

Gauss series is define for gamma function.

1 n−1 (n−1) 1
Γ(ϑ)Γ(ϑ + )...Γ(ϑ + ) = (2π) 2 n 2 −nϑ Γ(nϑ) (3.6)
n n
for proof using equation [2.6] and [2.7] for m is 0 to n − 1

m k m
Γp (ϑ + ) = (ϑ + − 1)Γ(ϑ + − 1)
n n n

14
 1
p!pϑ+ n −1
=
(ϑ + m
n
)(ϑ + 1 + n1 )...(ϑ + p + n1 − 1)
√ 1 1
2π( pn
e
)p pϑ+ n − 2
=
(nϑ + m)(nϑ + n + m)...(nϑ + m − p + np)

taking product m = 0 to n − 1 gamma function

n−1
m Y m
Γ(ϑ + ) = lim Γp (ϑ + )
n p→∞
m=0
n

n−1 √ 1 1
Y 2π( pn
e
)p pϑ+ n − 2
= limp→∞
m=0
(nϑ + m)(nϑ + n + m)...(nϑ + m − p + np)
p n pn pn nϑ− n Pn−1 m
2π) ( e ) p ϑ p m=0 n
= limp→∞
(nϑ + m)(nϑ + n + m)...(nϑ − 1 + np)

sum arithmetic sequence and substituting pn → p

p n−1 nϑ−1 1
2π) p!p p 2−nϑ
= limp→∞
(nϑ + m)(nϑ + n + m)...(nϑ − 1 + p)
using stirlings formula

(n−1) 1
= (2π) 2 n 2 −nϑ (nϑ − 1)Γ(nϑ − 1)

so
(n−1) 1
= (2π) 2 n 2 −nϑ Γ(nϑ)

putn = 2 in eq[3.5]then we known that legendre multiplication formula

√
1 π
Γ(ϑ)Γ(ϑ + ) = (2ϑ−1) Γ(2ϑ) (3.7)
2 2

and if put ϑ= n1 in eq.[3.5] is known as Eular form of multiplication.

n−1
1 2 n−1 (2π) 2
Γ( )Γ( )...Γ( )= √ (3.8)
n n n n

15
Chapter 4

Incomplete gamma function.

Incomplete gamma is generalisation form of gamma function .We can work on two
types of incomplete gamma functions.
1.Upper incomplete.
2.lower incomplete.

4.1 Upper incomplete gamma function.

for x>0 and k is any real number then Upper incomplete gamma function define .
Z ∞
Γ(k, x) = uk−1 e−u du (4.1)
x

for k is any integer ϑ so

ϑ−1 p
−x
X x
Γ(ϑ, x) = (ϑ − 1)!e (4.2)
p=0
p!

And gamma function is represent by

Γ(ϑ) = Γ(ϑ, 0) (4.3)

16
4.2 Lower incomplete gamma function.
for x>0 lower incomplete gamma function define
Z x
γ(k, x) = uk−1 e−u du (4.4)
o

for k is any integerϑ

ϑ−1 p
x
X x
γ(ϑ, x) = (ϑ − 1)!(1 − e ) ) (4.5)
p=0
p!

and the gamma function is define.

Γ(ϑ) = limx→∞ γ(ϑ, x) (4.6)

4.3 Properties of incomplete gamma function

4.3.1 Integration by part.

for x>0 and ϑ is any real number then Upper incomplete gamma function relation .

Γ(ϑ + 1, x) = ϑΓ(ϑ, x) + xe−x (4.7)

for x>0 lower incomplete gamma function relation.

γ(ϑ + 1, x) = sγ(ϑ, x) − xϑ e−x (4.8)

Since the gamma function is defined bu equation. [2.3]

Z ∞
Γ(ϑ) = uϑ−1 e−u du
0

we have
Γ(ϑ) = Γ(ϑ, 0) = lim γ(ϑ, x)
x→∞

so relation between Upper incomplete gamma function , lower incomplete gamma

function and gamma function

γ(s, x) + Γ(s, x) = Γ(s) (4.9)

17
4.3.2 Derivative of upper incomplete gamma function.

Use integral representation above, derivative of the upper incomplete gamma function
Γ(ϑ, x) with respect to x is

∂Γ(ϑ, x)
= −xϑ−1 e−x
∂x

The derivative with respect to its s is given by

∂Γ(ϑ, x)
= ln xΓ(ϑ, x) + xT (3, ϑ, x)
∂ϑ

second derivative by

∂ 2 Γ(ϑ, x)
2
= ln2 xΓ(ϑ, x) + 2x[ln xT (3, ϑ, x).
∂ϑ

where the function T (m, s, x) is a special case of the Meager G-function

 
0, 0, . . . , 0
T (m, ϑ, x) = Gm,0
m−1,m
 | x .
s − 1, −1, . . . , −1

4.4 Inequality hold by incomplete gamma function

We know psi function in term of gamma function ψ(a)

Γ′ (a)
ψ(a) = (4.10)
Γ(a)

and also about these identities.

Z ∞  
−tk 1 1 p
e dt = Γ ,x , (4.11)
x k k

Z ϑ    
−tk 1 1 1 1 k
e dt = Γ − Γ ,x (4.12)
0 k k k k
hold for all ϑ, k > 0. Let Z ∞
ϑk k
Ik (ϑ) = e e−t dt.
ϑ

18
for k=2 Z ∞
ϑ2 2
I2 (ϑ) = e e−t dt.
ϑ

Komatu and Pollak have proved that the double inequality[5]

1 1
√ < I2 (ϑ) < q
ϑ3 + 2 + ϑ ϑ2 + π4 + ϑ

for every ϑ > 0. , Gautschi has[5] proved this double inequality

1h 2
1/k i 1 h k
1/k i
ϑ +2 − ϑ < Ik (ϑ) < ϑ + a0 −ϑ
2 a0

for every ϑ > 0 and ϑ > 1, where

 
k/k−1 1
ao = Γ 1+
p

An application of inequality (13) was given in (25). Alzer [26] proved that the double
inequality
  1/2    1/k 
1 
−αϑk 1 
−βϑk
r 1+ 1− 1−e < Γk (ϑ) < r 1 + 1− 1−e
k k

for all ϑ ≥ 0 and k > 0 with k ̸= 1 and

α ≥ (1, Γ−k (1 + 1/k))

and
β ≤ min(1, Γ−k (1 + 1/k))

Motivated by inequality [ ], in the article we deal with the monotonicity of the function

1/k
1/k
ϑk + a −ϑ ϑk + a −ϑ
R(ϑ) = ϑk R ∞ −tk =
e ϑ e dt Γk (ϑ)

so we can prove that

1h k
1/k i 1h k
1/k i
ϑ +a − ϑ < I(x) < ϑ +b −ϑ
a b

19
for all ϑ > 0 and k > 1 if and only if a ≥ 2 and
b ≤ a0 = Γp/(1−p) (1 + 1/p).

4.4.1 Some inequality of gamma function

P roof
1
Γ1/(1−ϑ) (1 + ϑ) >
2
for each ϑ ∈ (0, 1).

P roof We can see easily inequality is equivalent to

ln Γ(1 + ϑ) + (1 − ϑ) ln 2 > 0

for each ϑ ∈ (0, 1).

Let
E(ϑ) = ln Γ(1 + ϑ) + (1 − ϑ) ln 2

put ϑ = 0 in equation ()

E(1) = 0

and differentiate equation

E ′ (ϑ) = ψ(ϑ + 1) − ln 2 < ψ(2) − ln 2 = 1 − γ − ln 2 < 0

for each ϑ ∈ (0, 1),

where γ is the Euler-Mascheroni constant which value is 0.5772 . . . .
Therefore,
1
Γ1/(1−ϑ) (1 + ϑ) > (4.13)
2

20
2.P rovethat.
Γ1/ϑ (1 + ϑ) is strictly increasing on (0, ∞), and t

ϑ < Γ1/ϑ (1 + ϑ) < 1

for each ϑ ∈ (0, 1).

Proof
Let
β1 (ϑ) = log Γ(1 + ϑ)

β2 (ϑ) = ϑ
β1 (ϑ) log Γ(1 + ϑ)
β(ϑ) = =
β2 (ϑ) ϑ
γ(ϑ) = log Γ(1 + ϑ) − ϑ log ϑ

we can easily solve

β1 (0) = β2 (0) = 0

γ (0+ ) = γ(1) = 0,
h ′ i′
β1 (ϑ)
β ′ (x)
= γ ′ (ϑ + 1) > 0
2

for ϑ ∈ (0, ∞), and

1
γ ′′ (a) = γ ′ (1 + a) − <0
a
for (a ∈ (0, 1) so for using all equation than β(ϑ) and eβ(ϑ) = Γ1/ϑ (1 + ϑ) is strictly
increasing on (0, ∞).
and also we can say γ(ϑ) is strictly concave on the interval (0, 1) and the inequality

γ(ϑ) > γ(0)(1 − ϑ) + γ(1)ϑ

for all ϑ ∈ (0, 1).

so, the monotonicity Γ1/ϑ (1 + ϑ) on the interval (0, 1).

21
Chapter 5

Generalization of gamma function

let the difference equation

β(t + 1) = s(t)β(t) (5.1)

has been studied directly and indirectly through

β(t + 1) − β(t) = ϕ(ϑ)

g(t) is a meromorphic function. As par solution (5.1) is obtained under an entirely

different assumption with regard to g(t).
One of other functions satisfying this condition is of the form

f (t)h(t)eϕ (ϑ)

there f (ϑ) is regular at infinity,

and h(ϑ) and ψ(ϑ) are any algebraic functions , Abelian integral, finite combination
of algebraic functions and Abelian integral It is also bring out the similarity of the
solution and its properties to those of the Gamma function.

5.1 Generalize gamma function in Gaussian form

Take g(t) satisfying two condition:
1.log g(t) be analytic in [0, ∞)
2.for some value of p, ∈ is any positive , may be found

22
such that
dp+1
Limk→+∞ k ∈ log s(t + k) = 0 (5.2)
dtp+1

uniform over a finite region of the t plane. Under conditions of solution

β(t + 1) = s(t)β(t)

denote by Γs(w) (t), it will be ordinary Gamma function

using equation []
1 · 2 · · · (k − 1)
Γ(t) =k→0 kt
t(t + 1) · · · (t + k − 1)
s(1)s(2) · · · s(k − 1)
Γs(w) (t) = lim [s(k)]t S(p, t) (5.3)
k→∞ s(t)s(t + 1) · · · s(t + k − 1)

G(k, t) is determined and (5.2) converges, and satisfies the fundamental difference
equation.
for any condition.
S(k, t + 1)S(k, t)
k→∞ =1 (5.4)
s(t + k)s(k)
we may have condition (5.3) fulfilled by setting

d β2 (t) dp βp+1 (t)

ln S(k, t) = ln s(k) + · · · + p ln s(k) (5.5)
dt 2! dk (p + 1)!

there β 2 (ϑ), · · · βt+1 (ϑ) are the Bernouillian polynomials. we need to show that the
same value of F (p, ϑ) insures the convergence of (5.3). βn (1) = 0,

Γs(w) (1) = 1

where q =0,1,2,· · · · ·n

Γs(w) (q) = s(1)s(2) · · · · s(q − 1) = f (q − 1)!

The general solution of (5.1) can be written

β(t) = Γs(w) (t)P (ϑ) (5.6)

23
where P (ϑ) is a periodic function with period unity.
Clearly Γf (w) (t) is the only solution of (5.1)
then
ϕ(ϑ + p)
Limϑ→∞ =1 (5.7)
f (p − 1)|f (p)|−F (p, ϑ)
Equation (5.7) in connection with (5.1) may therefore can be as defining. Γ(g(w))(t).

5.2 Eulerian Form and Convergence of Generalize

gamma function .
Equation [5.3], can be rewritten and k replace to b

∞ t
[s(1)]t S(1, t) Y s(k)

s(b + 1) S(b + 1, t)
Γs(w) (t) = (5.8)
s(t) b=1
s(t + b) s(b) F (b, t)

∞
[s(1)]t S(1, t) Y
= Q(b)
s(t) b=1

Since

β2 (ϑ) β3 (t) βz (k) tz − t

+ + ··· + =
(t − 1)!2! (t − 2)!3! z! z!
we have
p
X pz − ϑ dz log s(b)
log S(b + 1, t) = log S(b, t) +
s=1
t! dbz
p−1
Y dp+1 log s(k + ϕz ) βz+1 (t)
+ λz p+1
z=1
dk (p − z + 1)!(z + 1)!

|λz | ≦ 1; 0 < θz < 1,

by using Darboux’s form of the remainder in Taylor’s series for complex variables
in connection with . substitute in (5.8) and reduce obtain

p−1
X dp+1 log f (k + θz ) βz+1 (t)
log Q(k) = λz p+1
z=1
dk (p − z + 1)!(z + 1)!

24
 tλp dp+1 tp+1 λp+1 dp+1
+ log g (b + θp ) − log g (b + θp+1 t) .
(p + 1)! dbp + 1 p + 1)! dto+1
By means of condition (b) we see that from some value of (k) on, the terms of
Σ log Q(k) are less in absolute value than those of the convergent series Σ (1/ (k 1+ϵ )).
Hence (1) is absolutely convergent for all finite values of ϑ which are not zeros of some
f (t + k). Condition (b) likewise proves (1) to be uniformly convergent over any finite
region of the ϑ plane which excludes these points. Γf (w) (t) is therefore an analytic
function, except for isolated points, in the entire sector at least over which f (ϑ) is
analytic.

5.3 Generalize gamma function of Weierstrassian

Form, and Derivatives
When p ≦ 1, equations (5.8)and (5.3) become and

s(1)s(2) · · · s(b − 1)
Γs(w) (t) = Limb→∞ [s(b)]t , (5.9)
s(ϑ)s(ϑ + 1) · · · s(t + b − 1)

∞
( t )
[s(1)]t Y

s(b) s(b + 1)
Γs(w) (t) = . (5.10)
s(t) b=1 s(t + b) s(b)

(5.9) is uniformly convergent,

differentiate logarithmically
then

Γ′s(w) (ϑ) f ′ (ϑ) f ′ (1) f ′ (ϑ + 1) f ′ (2)

    
= −γf (u) − − + − + ··· , (5.11)
Γf (u) (ϑ) f (ϑ) f (1) f (ϑ + 1) f (2)

where  
s‘(1) s‘(2) s‘(b)
γs(w) = Limb→∞ + + ··· + − log g(b) (5.12)
s(1) s(2) s(b)
convergence γs(w) being easily founded.
Clearly
Γ′s(w) (1) = −γs(w) . (5.13)

25
The analogy with Γ(t) is further brought out by transforming (1) into

∞  
1 γs(w)
Y s(b + t) s‘(b)
− s(b) ϑ
= s(t)e e , (5.14)
Γs(w) (t) b=1
s(b)

the generalization of the Weierstrassian form of Γ(t). From (5.12) we find

d2 d2 d2
− log Γ s(w) (t) = log s(t) + log s(t + 1) + · · · .
dt2 dt2 dt2

More generally for p unrestricted, we have

dp+1 dp+1 dp+1

− log Γ s(w) (t) = log s(t) + log g(t + 1) + · · ·
dtp+1 dtp+1 dtp+1

5.4 Integral for the Asymptotic Constant.

In the present section and in the one following we will obtain results are very useful
in establishing particular relations between the generalised Gamma functions.
if Z t+1
u= log Γs(w) (x)dx
t

so
dw
= log Γs(w) (t + 1) − log Γs(w) (t) = log s(t)
dt
and Z t
w= log s(x)dx + C
a
Z t
1
log Γs(w) (t) ∼ log Fs(w) + log s(x)dx − log s(t)
a 2
r
X (−1)p−1 Ap d2p−1 log s(t)
+
p=1
(2p)! dt2p−1

it easily follows that

Z t+1 Z t+1 Z x 
log Γs(w) (x)dx ∼ log Fs(w) + log s(w)dw dx
t t a
t+1 n
(−)p−1 Bp d2p−2
Z
1 X s(t + 1)
− log s(x)dx + 2p−2
log
2 t p=1
(2p)! dt s(t)

26
But, using the Euler-Maclaurin Sum formula, † we get
Z t+1 Z x  Z t
log s(w)dw dx ∼ log s(x)dx
t a a
t+1 n
(−)p Ap d2p−2
Z
1 X s(t + 1)
+ log s(x)dx + 2p−2
log
2 t p=1
(2p)! dt s(t)

Z t+1 Z t
log Γs(w) (x)dx ∼ log s(x)dx + log Fs(w)
t a

Comparing with the above, we see that and

C = log Fs(w)

Z t+1 Z t
log Γs(w) (x)dx = log f (x)dx + log Fs(u)
t a

Letting t = a, we obtain
Z a+1
log Fu(w) = log Γs(w) (x)dx,
a

an analytical expression for the constant Fs(w) . Since Fs(w) . depends on the value of
a chosen, we shall write it a Fs(w) . Then
Z a1
loga1 Fs(w) = loga2 Fs(w) + log f (x)dx.
a2

5.5 The Associated Periodic Functions of .

Let s(t) be a function and both Γs(w) (t) and Γs(−w) (t) exist,

Γs(w) (t)Γs(−w) (1 − t) = S(t)

so
S(t + 1) = S(t)

27
Then
it can be denote this periodic function by Ps(w) (t), i. c.,

Γs(w) (t)Γs(−w) (1 − t) = Ps(w) (t) (5.15)

if a Fs(w)−a Fs(−w) = a πs(w) . Then using the integral assimtotic equation, we easily
obtain Z a+1
log a πs(w) = log Ps(w) (t)dt. (5.16)
a

so
1
Ps(w) (0) = ,
s(0)
or
Limt→0 Ps(w) (t)s(t) = 1,
       2
1 1 1 aπs(w)
Ps(w) = Γs(w) Γs(−w) = (5.17)
2 2 2 2πs(w/2)
then
Ps(w) (t) = Ps(−w) (−t), (5.18)
1 Ps(w) (t + a)
Ps(w+a) (t) = , (5.19)
s(a) Ps(w) (1 + a)
using multiplication theorem
     n
1 n−1 a Fs(w)
Γs(w) (t)Γs(w t+ · · · Γs(w) t + = Γs(w)/n) (nt)
n n na Fs(w/n)

,
     n
1 2 a Fs(−w)
Γs(−w) − t Γs(−w) − t · · · Γs(−w) (1 − t) = Γs(−w/n) (1 − n
n n a Fs(/n)

Inverting and multiplying, we finally obtain

     n
1 n−1 a πs(w)
Ps(w) (t)Ps(w) t + · · · Ps(w) t + = Pf (w/n) (nt).
n n na πf (u/n)

28
Chapter 6

Generalization of gamma function

let the difference equation

β(t + 1) = s(t)β(t) (6.1)

has been studied directly and indirectly through

β(t + 1) − β(t) = ϕ(ϑ)

g(t) is a meromorphic function. As par solution (5.1) is obtained under an entirely

different assumption with regard to g(t).
One of other functions satisfying this condition is of the form

f (t)h(t)eϕ (ϑ)

there f (ϑ) is regular at infinity,

and h(ϑ) and ψ(ϑ) are any algebraic functions , Abelian integral, finite combination
of algebraic functions and Abelian integral It is also bring out the similarity of the
solution and its properties to those of the Gamma function.

6.1 Generalize gamma function in Gaussian form

Take g(t) satisfying two condition:
1.log g(t) be analytic in [0, ∞)
2.for some value of p, ∈ is any positive , may be found

29
such that
dp+1
Limk→+∞ k ∈ log s(t + k) = 0 (6.2)
dtp+1

uniform over a finite region of the t plane. Under conditions of solution

β(t + 1) = s(t)β(t)

denote by Γs(w) (t), it will be ordinary Gamma function

using equation []
1 · 2 · · · (k − 1)
Γ(t) =k→0 kt
t(t + 1) · · · (t + k − 1)
s(1)s(2) · · · s(k − 1)
Γs(w) (t) = lim [s(k)]t S(p, t) (6.3)
k→∞ s(t)s(t + 1) · · · s(t + k − 1)

G(k, t) is determined and (5.2) converges, and satisfies the fundamental difference
equation.
for any condition.
S(k, t + 1)S(k, t)
k→∞ =1 (6.4)
s(t + k)s(k)
we may have condition (5.3) fulfilled by setting

d β2 (t) dp βp+1 (t)

ln S(k, t) = ln s(k) + · · · + p ln s(k) (6.5)
dt 2! dk (p + 1)!

there β 2 (ϑ), · · · βt+1 (ϑ) are the Bernouillian polynomials. we need to show that the
same value of F (p, ϑ) insures the convergence of (5.3). βn (1) = 0,

Γs(w) (1) = 1

where q =0,1,2,· · · · ·n

Γs(w) (q) = s(1)s(2) · · · · s(q − 1) = f (q − 1)!

The general solution of (5.1) can be written

β(t) = Γs(w) (t)P (ϑ) (6.6)

30
where P (ϑ) is a periodic function with period unity.
Clearly Γf (w) (t) is the only solution of (5.1)
then
ϕ(ϑ + p)
Limϑ→∞ =1 (6.7)
f (p − 1)|f (p)|−F (p, ϑ)
Equation (5.7) in connection with (5.1) may therefore can be as defining. Γ(g(w))(t).

6.2 Eulerian Form and Convergence of Generalize

gamma function .
Equation [5.3], can be rewritten and k replace to b

∞ t
[s(1)]t S(1, t) Y s(k)

s(b + 1) S(b + 1, t)
Γs(w) (t) = (6.8)
s(t) b=1
s(t + b) s(b) F (b, t)

∞
[s(1)]t S(1, t) Y
= Q(b)
s(t) b=1

Since

β2 (ϑ) β3 (t) βz (k) tz − t

+ + ··· + =
(t − 1)!2! (t − 2)!3! z! z!
we have
p
X pz − ϑ dz log s(b)
log S(b + 1, t) = log S(b, t) +
s=1
t! dbz
p−1
Y dp+1 log s(k + ϕz ) βz+1 (t)
+ λz p+1
z=1
dk (p − z + 1)!(z + 1)!

|λz | ≦ 1; 0 < θz < 1,

by using Darboux’s form of the remainder in Taylor’s series for complex variables
in connection with . substitute in (5.8) and reduce obtain

p−1
X dp+1 log f (k + θz ) βz+1 (t)
log Q(k) = λz p+1
z=1
dk (p − z + 1)!(z + 1)!

31
 tλp dp+1 tp+1 λp+1 dp+1
+ log g (b + θp ) − log g (b + θp+1 t) .
(p + 1)! dbp + 1 p + 1)! dto+1
By means of condition (b) we see that from some value of (k) on, the terms of
Σ log Q(k) are less in absolute value than those of the convergent series Σ (1/ (k 1+ϵ )).
Hence (1) is absolutely convergent for all finite values of ϑ which are not of some
f (t + k). Condition (b) likewise proves (1) to be uniformly convergent over any finite
region of the ϑ plane which excludes these points. Γf (w) (t) is therefore an analytic
function, except for isolated points, in the entire sector at least over which f (ϑ) is
analytic.

6.3 Generalize gamma function of Weierstrassian

Form, and Derivatives
When p ≦ 1, equations (5.8)and (5.3) become and

s(1)s(2) · · · s(b − 1)
Γs(w) (t) = Limb→∞ [s(b)]t , (6.9)
s(ϑ)s(ϑ + 1) · · · s(t + b − 1)

∞
( t )
[s(1)]t Y

s(b) s(b + 1)
Γs(w) (t) = . (6.10)
s(t) b=1 s(t + b) s(b)

(5.9) is uniformly convergent,

differentiate logarithmically
then

Γ′s(w) (ϑ) f ′ (ϑ) f ′ (1) f ′ (ϑ + 1) f ′ (2)

    
= −γf (u) − − + − + ··· , (6.11)
Γf (u) (ϑ) f (ϑ) f (1) f (ϑ + 1) f (2)

where  
s‘(1) s‘(2) s‘(b)
γs(w) = Limb→∞ + + ··· + − log g(b) (6.12)
s(1) s(2) s(b)
convergence γs(w) being easily founded.
Clearly
Γ′s(w) (1) = −γs(w) . (6.13)

32
The analogy with Γ(t) is further brought out by transforming (1) into

∞  
1 γs(w)
Y s(b + t) s‘(b)
− s(b) ϑ
= s(t)e e , (6.14)
Γs(w) (t) b=1
s(b)

the generalization of the Weierstrassian form of Γ(t). From (5.12) we find

d2 d2 d2
− log Γ s(w) (t) = log s(t) + log s(t + 1) + · · · .
dt2 dt2 dt2

More generally for p unrestricted, we have

dp+1 dp+1 dp+1

− log Γ s(w) (t) = log s(t) + log g(t + 1) + · · ·
dtp+1 dtp+1 dtp+1

6.4 Integral for the Asymptotic Constant.

In the present section and in the one following we will obtain results are very useful
in establishing particular relations between the generalised Gamma functions.
if Z t+1
u= log Γs(w) (x)dx
t

so
dw
= log Γs(w) (t + 1) − log Γs(w) (t) = log s(t)
dt
and Z t
w= log s(x)dx + C
a
Z t
1
log Γs(w) (t) ∼ log Fs(w) + log s(x)dx − log s(t)
a 2
r
X (−1)p−1 Ap d2p−1 log s(t)
+
p=1
(2p)! dt2p−1

it easily follows that

Z t+1 Z t+1 Z x 
log Γs(w) (x)dx ∼ log Fs(w) + log s(w)dw dx
t t a
t+1 n
(−)p−1 Bp d2p−2
Z
1 X s(t + 1)
− log s(x)dx + 2p−2
log
2 t p=1
(2p)! dt s(t)

33
But, using the Euler-Maclaurin Sum formula, † we get
Z t+1 Z x  Z t
log s(w)dw dx ∼ log s(x)dx
t a a
t+1 n
(−)p Ap d2p−2
Z
1 X s(t + 1)
+ log s(x)dx + 2p−2
log
2 t p=1
(2p)! dt s(t)

Z t+1 Z t
log Γs(w) (x)dx ∼ log s(x)dx + log Fs(w)
t a

Comparing with the above, we see that and

C = log Fs(w)

Z t+1 Z t
log Γs(w) (x)dx = log f (x)dx + log Fs(u)
t a

Letting t = a, we obtain
Z a+1
log Fu(w) = log Γs(w) (x)dx,
a

an analytical expression for the constant Fs(w) . Since Fs(w) . depends on the value of
a chosen, we shall write it a Fs(w) . Then
Z a1
loga1 Fs(w) = loga2 Fs(w) + log f (x)dx.
a2

6.5 The Associated Periodic Functions of .

Let s(t) be a function and both Γs(w) (t) and Γs(−w) (t) exist,

Γs(w) (t)Γs(−w) (1 − t) = S(t)

so
S(t + 1) = S(t)

34
Then
it can be denote this periodic function by Ps(w) (t), i. c.,

Γs(w) (t)Γs(−w) (1 − t) = Ps(w) (t) (6.15)

if a Fs(w)−a Fs(−w) = a πs(w) . Then using the integral assimtotic equation, we easily
obtain Z a+1
log a πs(w) = log Ps(w) (t)dt. (6.16)
a

so
1
Ps(w) (0) = ,
s(0)
or
Limt→0 Ps(w) (t)s(t) = 1,
       2
1 1 1 aπs(w)
Ps(w) = Γs(w) Γs(−w) = (6.17)
2 2 2 2πs(w/2)
then
Ps(w) (t) = Ps(−w) (−t), (6.18)
1 Ps(w) (t + a)
Ps(w+a) (t) = , (6.19)
s(a) Ps(w) (1 + a)
using multiplication theorem
     n
1 n−1 a Fs(w)
Γs(w) (t)Γs(w t+ · · · Γs(w) t + = Γs(w)/n) (nt)
n n na Fs(w/n)

,
     n
1 2 a Fs(−w)
Γs(−w) − t Γs(−w) − t · · · Γs(−w) (1 − t) = Γs(−w/n) (1 − n
n n a Fs(/n)

Inverting and multiplying, we finally obtain

     n
1 n−1 a πs(w)
Ps(w) (t)Ps(w) t + · · · Ps(w) t + = Pf (w/n) (nt).
n n na πf (u/n)

35
Chapter 7

Extended incomplete gamma

function

The Generalized inverse Gaussian distribution proposed by [8]

1 −1
g(k) := k α−1 e−ck−dk (k > 0, c > 0, d > 0, −∞ < α < ∞) (7.1)
I(α; c, d)

where I(α; c, d) is normalising constant. Density (6.1) can be generalised in an unlim-

ited number of ways. However, If the right cumulative density function and reliability
functions can be found to be helpful and less taxing, a generalisation of (6.1) will be
useful. The density is a natural extension of (6.1).

−η


 C(ζ, c, η, d)k ζ−1 e−ck−dk

y(k) := (t > 0, c > 0, d > 0, η > 0, −∞ < ζ < ∞) (7.2)



 0 elsewhere

where −1
Z ∞
ζ−1 −ck−dk−η
C := C(ζ, c, η, d) = k e , (7.3)
0

normalizing constant. A simple modification of the variable k yields special cases

of (6.2).These distributions can be used to solve many problems in the technical,
physical, social, and biological sciences.. The density’s cumulative density function
(6.2) is Z ϑ
−η
Y (ϑ) = C k ζ−1 e−ck−dk dk (ϑ > 0), (7.4)
0

36
reliability function
R(ϑ) = 1 − Y (ϑ).

In statistics and reliability theory, study into the functions F (ϑ) and R(ϑ) is critical.
The generalised incomplete gamma functions were first developed by Chaudhry and
Zubair [8]. Z ϑ
−1
γ(ζ, ϑ; d) = k ζ−1 e−k−dk dk (7.5)
0

and Z ∞
−1
Γ(ζ, ϑ; d) = k ζ−1 e−k−dk dk, (7.6)
ϑ

then pair of functions.

Z ϑ
−η
γ(ζ, ϑ; d, η) = k ζ−1 e−k−dk dk (ϑ ⩾ 0, d > 0; d = 0, Re(ζ) > 0) (7.7)
0

and Z ∞
−η
Γ(ζ, ϑ; d, η) = k ζ−1 e−k−dk dk (ϑ ⩾ 0, d ⩾ 0), (7.8)
ϑ

When we consider the generalised incomplete gamma functions (6.45)-(6.6), we can

see that these are special instances of (6.7)-(6.8). η = 1,

γ(ζ, ϑ; d; 1) = γ(ζ, x; d),

Γ(ζ, ϑ; d; 1) = Γ(ζ, x; d).

The extended incomplete gamma functions can be used to simplify the cumulative
density function (6.4) and the reliability function (6.5).

Y (ϑ) = Cc−ζ γ (ζ, cϑ; dcη ; β) ,

R(ϑ) = Cc−ζ Γ (ζ, cϑ; dcη ; η) .

Anderson et al. [9] given integral functions

R∞
xv exp −x − x−1/2 dx


I1 (ϑ, m) := 0
Rb
I2 (ϑ, b, m) := 0 xm exp −x − x−1/2 dx


R∞
I3 (ϑ, t, m) := 0 xm exp −x − (x + t)−1/2 dx

37
and Z ∞
xm exp −x − dxδ − x−1/2 dx,


I4 (ϑ, δtc, d; m) :=
0

The extended incomplete gamma functions can be used to simplify astrophysical ther-
monuclear functions.[9]

I1 (ϑ, m) = Γ m + 1, 0; ϑ; 12

I2 (ϑ, b, m) = γ m + 1, b; ϑ; 12


 
m
I3 (ϑ, t, m) = et m  (−t)m−r Γ r + 1, t; z, 1
P


r=0 2
r
(−b)r
I4 (ϑ, δ, d; m) = ∞ 1
P

r=0 r! Γ m + rδ + 1, 0; ϑ; 2

The extended gamma functions are clearly visible.

The Abramowitz’s function

Z ∞
2 −ϑt−1
yk (ϑ) := tm e−t dt, (7.9)
0

The function (6.9) be a special case of the extended incomplete gamma function.
which yields  
1 k+1 1
yk (ϑ) = Γ , 0; ϑ; (7.10)
2 2 2
Another integral function that Goodwin and Stalon explore is [9].

∞ 2
e−k
Z
f (n) := dx (7.11)
0 k+n

The exponential integral function ([9] is connected to the function in (6.11).

√
Z n
1 2 2 2
y(n) = e−n Ei n2 + πe−n ek dk
2 0

where Z n
−n2 2
Y (n) := e ek dk
0

38
The Dowson integral function is a type of integral function. The extended incomplete
gamma function can be used to simplify the function (6.11).
 
1 −n2 2 1
y(n) = e Γ 0, n ; 2n; −
2 2

7.1 Theorem
Γ(ζ + 1, ϕ; d; η) =ζΓ(ζ, ϕ; d; η) + bηΓ(ζ − η, ϕ; d; η)
−η
(7.12)
+ ϕζ e−ϕ−bϕ

Proof. Let
−η
y(k) := e−k−dk H(k − ϕ), (7.13)

where 
 1 if k >, ϕ
H(k − ϕ) := (7.14)
 0 if k < ϕ,

Heaviside unit step function.

The Mellin transform of the function y(k) is the extended gamma function (6.8). .

Γ(ζ, ϕ; d; η) = M {f (k); k → ζ}. (7.15)

The differentiation of (6.13) in the sense of distribution yields

d −η
{y(k)} = −1 + dηk −η−1 f (k) + e−k−dk δ(k − ϕ),


(7.16)
dk

where
d
δ(t − x) := (H(t − x)) (7.17)
dt
Dirac delta function.
A function’s Mellin transform and derivative are related.
 
d
−(ζ − 1)M {f (k); k → ζ − 1} = M (f (k)); k → ζ (7.18)
dk

39
From (6.18) and (6.17)

−(ζ − 1)Γ(ζ − 1, ϕ; d; η) = − Γ(ζ, ϕ; d; η) + dηΓ(ζ − η − 1; ϕ; d; η)

−η
+ ϕζ−1 e−ϕ−bϕ

which simplifies to give

Γ(ζ, ϕ; d; η) =(ζ − 1)Γ(ζ − 1, ϕ; d; η)

−η
(7.19)
dηΓ(ζ − η − 1, ϕ; d; η) + ϕζ−1 e−ϕ−bϕ

Replacing ζ by ζ + 1 in (6.18).

Γ(ζ + 1, ϕ; d; η) =ζΓ(ζ, ϕ; d; η) + bηΓ(ζ − η, x; d; η)

−η
+ ϕζ e−ϕ−bϕ

7.1.1 Corollary
−1
Γ(ζ + 1, ϕ; d) = ζΓ(ζ, ϕ; d) + dΓ(ζ − 1, ϕ; b) + ϕζ e−ϕ−bϕ . (7.20)

Proof.
In eq. (6.12) take η = 1. it is prove, and if b = 0

Γ(ζ + 1, x) = ζΓζ, x) + xζ e−x ,

40
Chapter 8

Application of gamma function

8.1 THE GAMMA FUNCTION AND BALLS

8.1.1 Volume Of The N-Dimensional Ball.

Any Euclidean Space’s Largest Unit Ball, Jeffrey Nunemacher lays the groundwork for
an intriguing gamma function application.the open ball defines by Jeffrey Nunemacher
lays of radius r and dimension n, Bn(r),[10]

X
yk2 < r2 (8.1)

V n(r) is the volume. Nunemacher derives his formula

ZZ Z
Vn (r) = ··· 1dy1 dy2 . . . dyn (8.2)
Bn (r)

By applying (7.1) to the limits of the iterated integral in (7.2) and conducting trigono-
metric substitutions, he derives the following - more important - identity, which is
specific to the unit ball: Z π/2
Vn = 2Vn−1 cosn θdθ (8.3)
0

41
8.1.2 Derivation

For the volume of an i-dimensional ball, David Singmaster employs the formula shown
below.: [10]
π i/2 rn
Vi (r) = (8.4)
Γ(i/2 + 1)
We’ll pick up with equations where Nunemacher left off. (7.4). using beta function
and notice its similarity to (7.3).
 
1 i 1
Vi (1) = Vi−1 (1)β , + (8.5)
2 2 2

recursion,  
1 i
Vi−1 (1) = Vi−2 (1)β ,
2 2
Consequently,
1 3 1 i 1 i 1
Vi (1) = Vi (1)β( , ) · · · β( , )β( , + )
2 2 2 2 2 2 2
where V1 (1) = 2 = B 12 , 1 . Gamma as a replacement for Beta:

"
#
1 1 3 1 i 1 i+1







Γ 2
Γ(1) Γ 2
Γ 2
Γ 2
Γ 2
Γ 2

Γ 2
Vi (1) = 3

... i+1
Γ 2
Γ(2) Γ 2

then "

n #

Γ 21 Γ(1)
Vi (1) =
Γ(i/2 + 1)
1

√
Since Γ 2
= π and Γ(1) = 1,

π i/2
Vi (1) =
Γ(i/2 + 1)

We don’t see the r in this equation when r = 1. Instead, we get the modified recursion
relation when we take the more general form:
Z π/2
Vi = 2rVi−1 cosi θdθ
0

42
With the exception of the ball’s size being dilated by a factor of r and its volume
being dilated by a factor of ri, the derivation will be virtually identical. Finally,

π i/2 ri
Vi (r) = ,
Γ(i/2 + 1)

This is in line with our previous assertion of (2.4). The volume of the n-ball can
now be calculated using the gamma function, and we can go on to an interesting
application.

8.2 Psi and Polygamma function

Gamma function Weierstrassformula

∞
1 Y
= zeγϑ (1 + ϑ/i)e−ϑ/i
Γ(ϑ) i=1

where γ is the Euler-Mascheroni constant.

ψ(ϑ) be defined
Γ′ (ϑ)
Γ(ϑ)
. for {ϑ ∈ C, ϑ ̸= 0, −1, −2, . . .}.
The most well-known representation of ,[1]

∞
1 X ϑ
ψ(ϑ) = −γ − + (8.6)
ϑ i=1 i(ϑ + i)

though the one we’ll eventually use to compute sums in the following subsection is
defined as follows:
1 ϑ−1
−1
Z
t
ψ(ϑ) = −γ − dt (8.7)
0 1−t
This integral holds true for Re(ϑ) > −1,
The polygamma functions, ψ(k), shall now be defined. The gamma and digamma
functions give rise to this family of functions. They’re useful since they lead to series
that are better and better convergent. The polygamma functions are the higher-order
derivatives of ψ(z), as the nomenclature suggests. Take a look at these instances of

43
recurrent differentiation. (7.6):

∞
X ∞
X
′ −2
ψ (ϑ) = (ϑ + i) , ψ (k)
(ϑ) = (−1) k+1
k! (ϑ + i)−k−1
i=0 i=0

8.2.1 The Computation Of Infinite Sums.

”A finite series of psi and polygamma functions can always be converted to an infinite
series with a general term in the index.” [10].

Example 8.1. Evaluate

∞
X 1
j=1
(j + 1)(3j + 1)

let uj ,  
1 1
uj =
3 (j + 1)(j + 1/3)
1 3/2 3/2
= −
(j + 1)(j + 1/3) j + 1 j + 1/3
so  
1 1 1
uj = −
2 j + 1 j + 1/3
for all A > 0, Z ∞
1
= e−Aϑ dϑ
A 0

Because both fractions’ denominators must be larger than 0, this identity may be
applied. As a result, the sum in equation may be rewritten as:

∞ ∞ Z ∞ Z ∞ 
X 1 1X −(j+1)x −(j+1/3)x
= e dx − e dx
j=1
(j + 1)(3j + 1) 2 j=1 0 0
∞ Z ∞ Z ∞ 
1X −jϑ −x −jϑ −(1/3)ϑ
= e e dϑ − e e dϑ
2 j=1 0 0
∞ Z ∞ 
1X −jϑ −ϑ −(1/3)ϑ


= e e −e dϑ
2 j=1 0

N Z ∞ !
1 X
e−iϑ e−x − e−(1/3)ϑ dϑ


= lim
2 N →∞ i=1 0

1−ϑJ+1
PN
Since , i=0 ϑj = 1−ϑ
.When we take out the first term of the series, we get,

44
ϑ0 = 1,then
J

X
jϑ 1 − ϑJ
ϑ =
j=1
1−ϑ

put e−ϑ for ϑ,

J
e−ϑ 1 − e−Jϑ


X
e−Jϑ =
j=1
1 − e−ϑ

Consider the following summary. We may swap summation and integration and keep
manipulating the sum because of suitable convergences resulting from the monotone
convergence theorem.

∞
"Z #
∞ −ϑ
1 − e−J


X 1 1 e
e−ϑ − e−(1/3)ϑ dϑ


= lim
j=1
(j + 1)(3j + 1) 2 J→∞ 0 1 − e−ϑ
1 ∞ e−ϑ e−ϑ − e−(1/3)ϑ
Z

= dϑ
2 0 1 − e−ϑ

We are now using a variable change. Let n = e−ϑ be the case. As a result, −e−ϑdϑ =
di. This substitution will be made.
∞ 1
i − i1/3
Z
X 1 1
= di
j=1
(j + 1)(3j + 1) 2 0 1−i


1 1 (i − 1) − i1/3 − 1
Z
= di
2 0 1−i
1 1 i−1 1 1 i1/3 − 1
Z Z
= di − di
2 0 1−i 2 0 1−i

Compare ψ(ϑ) .
∞
X 1 1 1
= ψ(4/3) − ψ(2).
i=1
(i + 1)(3i + 1) 2 2

by Professor Efthimiou
∞
X 1
S(c, d) = ,
k=1
(k + c)(k + d)

where c ̸= d, and {c, d ∈ C; Re(c), Re(d) > 0} .

and
ψ(d + 1) − ψ(c + 1)
S(c, d) =
d−c

45
8.3 Application in Probability
The distribution function D(κ) of waiting times until the Poisson event is given a
Poisson distribution with a rate of change of lambda.

D(κ) = P (K ≤ κ)

= 1 − P (K > κ)
(λκ)j e−λκ
ϑ j!
X
=1−
j=0
ϑ−1
−λκ
X (λκ)j
=1−e
j=0
j!

Γ(ϑ, κλ)
=1−
Γ(ϑ)
for κ ∈ [0, ∞), The corresponding probability function P (κ) of waiting times until the
ϑ the Poisson event is then obtained by differentiating D(κ),

P (κ) = D′ (κ)
ϑ−1 ϑ−1
−λκ
X (λt)j −λt
X j(λκ)j−1 λ
= λe −e
j=0
j! j=0
j!
ϑ−1 ϑ−1
X (λκ)j X j(λκ)j−1 λ
= λe−λκ + λe−λκ − e−λκ
j=1
j! j=1
j!
ϑ−1 
j(λϑ)j−1 (λt)j
X 
−λκ −λκ
= λe − λe −
j=1
j! j!

( ϑ−1  )
X (λx)j−1 (λκ)j
= λe−λκ 1− −
j=1
(j − 1)! j!

(λκ)ϑ−1
  
−λκ
= λe 1− 1−
(ϑ − 1)!
λ(λv)ϑ−1 −λκ
= e .
(ϑ − 1)!
let α ≡ ϑand The time between modifications is defined as θ ≡ 1/λ. The above

46
equation can then be represented as[11].

κα−1 e−κ/θ
P (κ) =
Γ(α)θα

for κ ∈ [0, ∞). This is the probability function for the gamma distribution.

47
Bibliography

[1] Ranvillie E.D, Special Functions, Macmillan, 1960

[2] E.Artin, The Gamma Function, New York, Holt.,Rinehart and Winston, (1964).

[3] G.E.Andrews,R.Askey and Roy,Special function, cambridge university,(1999)

[4] Arfken.G, The Incomplete Gamma Function and Related Functions. chapter
10.5 in Mathematical Methods for Physicists, 3rd ed. Orlando, FL: Academic
Press,(1985)

[5] Gautschi, Some elementary inequalities relating to the gamma and incomplete
gamma function. Math. Phys. 38, 77-81 (1959/60)

[6] Emil L.,The Generalized Gamma Function Annals of Mathematics ( 1919) pp.
202-217

[7] Shantanu das,Functional Fractions Calclus. Springer Science and Business Me-
dia, LLC, (2011)

[8] B. Jorgensen,,Statistical Properties of Generalized Inverse Gaussian Distribu-

tions, Springer-Verlag, New York, 1982.

[9] Haubold, A.M. Mathai, Astrophysical thermonuclear functions, Astrophys. Space

Sci. 214 (1994) 49â70.

[10] Jeffrey Nunemacher, The largest unit ball in any euclidean space. Mathematics
Magazine, 59(3):170â171, 1986.

[11] Jambunathan M. V.”Some Properties of Beta and Gamma Distribu-

tions.” Ann. Math. Stat. 25, 401-405, 1954.

48