Inequality from Stirling's formula

  • Context: Graduate 
  • Thread starter Thread starter neginf
  • Start date Start date
  • Tags Tags
    Formula Inequality
Click For Summary

Discussion Overview

The discussion revolves around the derivation and validity of an inequality related to Stirling's formula, specifically focusing on the expression involving binomial coefficients and its comparison to a function of n. Participants explore the implications of the inequality and its correctness, referencing a book on function approximations.

Discussion Character

  • Technical explanation
  • Debate/contested
  • Mathematical reasoning

Main Points Raised

  • One participant requests clarification on an inequality from a book, specifically C(n,n/2)/2^(n+1) > 1/(2*sqrt(n)).
  • Another participant expresses difficulty accessing the relevant page of the book and suggests typing out the question instead.
  • A participant presents the "sharp form of Stirling's inequality" and questions its correctness when tested with specific values of n.
  • One participant indicates that the inequality does not hold for n = 4 and n = 2, expressing uncertainty about its validity.
  • Another participant challenges the correctness of the original inequality and proposes a different inequality involving binomial coefficients and a modified expression.

Areas of Agreement / Disagreement

Participants do not reach a consensus on the correctness of the original inequality. There are competing views regarding its validity, with some participants asserting it is incorrect while others are unsure.

Contextual Notes

Some participants mention limitations in accessing the book, which may affect their ability to verify the inequality. There is also uncertainty regarding the specific values of n that were tested.

Mathematics news on Phys.org
I don't see how to navigate to page 17 in that link. (I'm not enthusiastic about how the interface to Google books behaves on my browser and internet connection. It seems darn slow and tedious.) You'd best type out your question.
 
Thank you for writing back.

The inequality is

n is even
C(n,n/2)/2^(n+1) > 1/(2*sqrt(n)).

"Sharp form fo Stirling's inequality" is

sqrt(2*pi*k) * k^k * e^-k < k! < sqrt(2*pi*k) * k^k * e^-k * (1+1/(4*k))

Is it right? Tried with 4.

With Google books, by clicking on the book in the upper left hand corner, it will appear big on the screen and you can click on the big image and page up and down.
Think I'm missing something here.
 
With Google books, by clicking on the book in the upper left hand corner, it will appear big on the screen and you can click on the big image and page up and down.
Think I'm missing something here.

I'm missing all pages after page 15. It does say "some pages are omitted from the preview".

neginf said:
The inequality is
n is even


\frac{ \binom{n}{n/2}}{2^{n+1}} &gt; \frac{1}{2 \sqrt{n}}

"Sharp form fo Stirling's inequality" is

(\sqrt{2 \pi k}) k^k e^{-k} &lt; k! &lt; (\sqrt{2 \pi k}) k^k e^{-k } (1+\frac{1}{ 4k} )

Is it right? Tried with 4.

Is what right? Do you mean that you used k = 4 or n = 4 ?
 
Sorry, n.
I tried it with 4 and it seemed not to hold, the inequality. Tried with 2 and same problem.
 
This inequality is not correct:
\frac{ \binom{n}{n/2}}{2^{n+1}} &gt; \frac{1}{2 \sqrt{n}}


Assuming the inequality
(\sqrt{2 \pi k}) k^k e^{-k} &lt; k! &lt; (\sqrt{2 \pi k}) k^k e^{-k } (1+\frac{1}{ 4k} )

the only similar inequality that I see is:

\frac{ \binom{n}{n/2}}{(1 + \frac{1}{2n})^2} &gt; \frac{\sqrt{2}}{\sqrt{\pi n}} &gt; \frac{1}{\sqrt{2n}}
 

Similar threads

  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 14 ·
Replies
14
Views
2K
Replies
7
Views
3K
Replies
44
Views
6K
  • · Replies 11 ·
Replies
11
Views
3K
Replies
2
Views
1K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 333 ·
12
Replies
333
Views
20K