What is the Upper Bound of this Summation?

Click For Summary

Discussion Overview

The discussion revolves around finding the upper bound of a specific summation involving factorials and powers, expressed as $$\sum\limits_{i=k}^{n} \frac {1}{(n-i)! m^{i-1}}$$. Participants explore various mathematical approaches, including the use of incomplete gamma functions and properties of the exponential series.

Discussion Character

  • Exploratory
  • Mathematical reasoning
  • Technical explanation

Main Points Raised

  • One participant proposes a naive approach to bound the summation by minimizing the denominator, but finds it insufficient.
  • Another participant suggests rewriting the summation in terms of ##\lambda \equiv (n-i)##, relating it to a partial sum of the exponential series, which can be bounded by the exponential or expressed using the incomplete gamma function.
  • A later post discusses the upper incomplete gamma function and its relation to the summation, referencing a theorem that connects it to the gamma function.
  • Concerns are raised about the definition of the gamma function at non-positive integers, specifically ##\Gamma(0)##, and its implications for the calculations.
  • One participant concludes that the expression for ##\gamma *(0,z)## simplifies to 1, but another expresses uncertainty about the derivation, indicating a potential oversight in the reasoning.

Areas of Agreement / Disagreement

Participants express differing views on the handling of the gamma function and its implications for the summation. The discussion remains unresolved regarding the proper treatment of ##\Gamma(0)## and the implications for finding the upper bound.

Contextual Notes

There are limitations related to the assumptions made about the gamma function and its properties, particularly at non-positive integers. The discussion also reflects varying levels of understanding regarding the application of the incomplete gamma function and its relationship to the original summation.

mooncrater
Messages
215
Reaction score
18
There is this summation, that I've been trying to solve, but am not able to do so. It is :
$$\sum\limits_{i=k}^{n} \frac {1}{(n-i)! m^{i-1}}$$
I would be happy to find it's upper bound too. So what I did was intensely naive. I made the denominator the minimum by making ##(n-i)! = 1## and ##m^{i-1} = m^{k-1}## (As that would give an upper bound too, but rather a loose one). That, as expected, didn't suffice my needs. Any ideas about how to solve this one?
Moon.
 
Physics news on Phys.org
If you rewrite the summation in terms of ##\lambda \equiv (n-i)##, you get a partial sum of the exponential series:
\left(\frac{1}{m^{n-1}}\right)\,\sum_{\lambda = 0}^{n-k} \frac{m^{\lambda}}{\lambda!},
which well can be bounded by the exponential or written in terms of the incomplete gamma function.
 
  • Like
Likes   Reactions: mooncrater and mfb
Thanks ! :)
 
Sorry, but I'm facing a problem again. I found about incomplete gamma functions on wikipedia:
https://en.wikipedia.org/wiki/Incomplete_gamma_function

And I found that the upper incomplete gamma function, and the limiting function( defined as ##\gamma *##) suffice my needs. As it's given that :
$$\gamma * (s,z) = e^{-z} \sum_{k=0}^{\infty} \frac {z^{k}}{\Gamma (s+k+1)} $$
which is equivalent to:
$$\gamma *(s,z) = e^{-z} \sum_{k=0}^{\infty} \frac {z^{k}}{(s+k)!} $$
When we put ##s=0## then we get :
$$\gamma *(0,z) = e^{-z} \sum_{k=0}^{\infty} \frac {z^{k}}{ k!} $$
thus in order to find this summation, we need to find :
$$\ {\gamma *(0,z)} {e^{z}} $$
So, basically we've to find ##\gamma *(0,z)## .
There is this Hargot's Theorem given in the wikipedia's page which says :
$$ \gamma (s,z) = z^{s} \Gamma (s) \gamma *(s,z) $$
Thus,
$$\gamma *(s,z) =\frac { \gamma(s,z)} {z^{s} \Gamma(s) } $$
since ##s=0##, therefore :
$$\gamma *(0,z) = \frac{\gamma(0,z)} {1* \Gamma(0)} $$
But, according to this link :
https://en.wikipedia.org/wiki/Particular_values_of_the_Gamma_function
Gamma function is not defined for non-positive integers.
So, how do I tackle this ##\Gamma(0)## ?
Moon.
 
mooncrater said:
Gamma function is not defined for non-positive integers.
So, how do I tackle this Γ(0)\Gamma(0) ?
Quoate from Ahlfors: Γ(z) is a meromorphic function with poles at z = 0, -1, -2, ... but without zeros.
 
  • Like
Likes   Reactions: mooncrater
Huh?
$$\gamma *(0,z) = e^{-z} \sum_{k=0}^{\infty} \frac {z^{k}}{ k!} = e^{-z} e^z = 1$$
Done?
 
  • Like
Likes   Reactions: mooncrater
mfb said:
Huh?
$$\gamma *(0,z) = e^{-z} \sum_{k=0}^{\infty} \frac {z^{k}}{ k!} = e^{-z} e^z = 1$$
Done?
Sorry, forgot about the expansion of the majestic ##e^{x}## itself.
Moon.
 

Similar threads

  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 13 ·
Replies
13
Views
3K
  • · Replies 17 ·
Replies
17
Views
2K
  • · Replies 8 ·
Replies
8
Views
2K
Replies
1
Views
2K
  • · Replies 5 ·
Replies
5
Views
3K
  • · Replies 3 ·
Replies
3
Views
5K