0% found this document useful (0 votes)
194 views2 pages

Derivative of Gamma Function

This document defines the gamma function and its derivative, the digamma or psi function. It shows that the psi function can be written as the sum of the harmonic series plus a constant term involving psi(1). Taking the limit as p approaches infinity allows writing psi(1) solely in terms of the Euler-Mascheroni constant.

Uploaded by

Andhee Jacobe
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
194 views2 pages

Derivative of Gamma Function

This document defines the gamma function and its derivative, the digamma or psi function. It shows that the psi function can be written as the sum of the harmonic series plus a constant term involving psi(1). Taking the limit as p approaches infinity allows writing psi(1) solely in terms of the Euler-Mascheroni constant.

Uploaded by

Andhee Jacobe
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Derivative of the Gamma Function

Jacobe, Andhee M.
January 1, 2020

Definition: Gamma Function Now, what we need is to find the value of Ψ(1) and to do so, first
For any real number z > 0, the gamma function is given as define another function.
Z ∞ Definition: p-Gamma Function
Γ(z) := xz−1 e−x dx (1) Let p be a positive integer and z > 0. Define the p-gamma func-
0 tion as
The gamma function can also be defined in a recursive manner as p!pz pz
Γp (z) := = . (14)
Γ(z) = (z − 1)Γ(z − 1) (2) z(z + 1)...(z + p) z(1 + z/1)...(1 + z/p)

which then gives us the Factorial Function Equation (14) also has an integral form given as
Z p p
Γ(z) = (z − 1)! (3) x
Γp (z) := 1− xz−1 dx (15)
0 p
whenever z is a positive integer.
We also note that
Derivative of Gamma Function pz = ez ln p = ez(ln p−Hp ) ezHp . (16)
Using the Leibniz rule in differentiating an integral, we have
Z ∞ Manipulating (14) and using (16) gives us
0
Γ (z) = xz−1 ln x e−x dx. (4) p
0 zΓp (z) Y  z
1= 1 + (17)
However, applying differentiation (product rule) to equation (2) pz k
k=1
also gives us p 
z(Hp −ln p) −zHp
Y z
0 0 = zΓ p (z)e e 1+ (18)
Γ (z) = (z − 1)Γ (z − 1) + Γ(z − 1) (5) k
k=1
p 
which if we divide by Γ(z) and using (2) becomes Y z  −z/k
= zΓp (z)ez(Hp −ln p) 1+ e (19)
Γ0 (z) (z − 1)Γ0 (z − 1) + Γ(z − 1) Γ0 (z − 1) 1 k
k=1
= = + . (6)
Γ(z) Γ(z) Γ(z − 1) z−1 by noting that
p
Now, we define the Psi or Digamma function as Y
e−zHp = e−z/k . (20)
0
Γ (z) k=1
Ψ(z) := . (7)
Γ(z) We also define the Euler-Mascheroni Constant as
Definition: Euler-Mascheroni Constant
Looking at (6), it is clear that
1 γ := lim (Hp − ln p) = 0.57721566490... (21)
Ψ(z) = Ψ(z − 1) + (8) p→∞
z−1
If for (19) we let p → ∞, we are interested what happens to the
and for any positive integer z right-hand side.
1
Ψ(z) = Ψ(z − 1) + (9) Claim: Γ (z) → Γ(z) as p → ∞.
z−1 p
1 1 This is equivalent to showing Γp (z) − Γ(z) → 0 as p → ∞.
= Ψ(z − 2) + + (10)
z−2 z−1 Z p
x
p Z ∞
.. Γp (z) − Γ(z) = 1− xz−1 dx − xz−1 e−x dx
. (11) 0 p 0
z−1
X 1 = I 1 + I2 + I 3
= Ψ(1) + . (12)
k where
k=1
Z ∞
It can also be noted that the second term in the right-hand side I1 = − xz−1 e−x dx (22)
th
of (12) is the (z − 1) harmonic sum which we denote as Hz−1 . p
Z p  p 
x −x
And hence, we manipulate (7) and (12) and arrives at I2 = 1− −e xz−1 dx (23)
p/2 p
z−1
! Z p/2  p 
1 x −x
xz−1 dx
X
0 I3 = 1− −e (24)
Γ (z) = Ψ(1) + Γ(z) (13)
k 0 p
k=1
Now, (22) approaches zero as p → ∞. Using the Taylor Series Equation (35) can then be simplified as
p
Expansion of ln (1 − z/p) and the interval 0 ≤ x ≤ p give us
∞  
1 X 1 1

x
p 
x
 Ψ(z) = − − γ + −
ln 1 − = p ln 1 − = −x − L (25) z k k+z
k=1
p p z−1   ∞   ∞  
X 1 X 1 X 1
where = −γ + + −
k k k+z
x2 x2 k=1 k=z+1 k=1
 
1 x
L= + + 2 + ... . (26) z−1   ∞   X ∞  
p 2 3p 4p X 1 X 1 1
= −γ + + −
Which in turn gives us k k+z k+z
k=1 k=1 k=1
z−1  
 p X 1
−x x = −γ +
0<e − 1− = e−x − e−L e−x ≤ e−x . (27) k
p k=1

Then (27) implies that (23) converges to zero as p → ∞. More- which if compared to (12) will enable us to conclude that
over, using (26) and the fact that x/p ≤ 1/2, we get
Ψ(1) = −γ (36)
 p
x
0 < e−x − 1 − = e−x − e−L e−x ≤ e−x L (28) which ultimately states that for any positive integer z,
p
z−1
!
but it is clear also that 0
X 1
Γ (z) = −γ + Γ(z) (37)
k
∞ k=1
2 X
x 1
L≤ . (29) Remark: The derived results also apply to complex values with
p (k + 1)2k−1
k=0
some exceptions (zero and negative integers).
Combining (28), (29), and (24) will result to
====== End of Investigation =======

1X 1
|I3 | ≤ Γ(z + 2) (30)
p (k + 1)2k−1
k=0

which ultimately converges to zero as p → ∞. Therefore,


ultimately, we proved our claim.

We go back to the limit of (19) as p → ∞. Combining the


claim and the definition for the Euler-Mascheroni constant, we
have
∞ 
Y z  −z/k
1 = zΓ(z)ezγ 1+ e . (31)
k
k=1

Rewriting (31) , we have


∞ 
1 Y z  −z/k
= zezγ 1+ e . (32)
Γ(z) k
k=1

Equation (32) is called the Weierstrass Form of the gamma


function.

We now extract the logarithm of (32), that is


∞ 
X  z z
− ln Γ(z) = ln z + zγ + ln 1 + − (33)
k k
k=1

Taking the derivative of (33) with respect to z,



Γ0 (z)
 
1 X 1 1
− = +γ+ − (34)
Γ(z) z k+z k
k=1

One may notice that left-hand side is the negative of the Psi or
Digamma function, hence:
∞  
1 X 1 1
Ψ(z) = − − γ + − (35)
z k k+z
k=1

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy