Proving Inequalities of Euler-Mascheron Constant with Taylor Expansion

Click For Summary

Homework Help Overview

The discussion revolves around proving inequalities related to the Euler-Mascheron constant using Taylor expansion techniques. The original poster presents inequalities involving logarithmic functions and seeks to establish their validity to demonstrate the finiteness of the Euler-Mascheron constant.

Discussion Character

  • Exploratory, Conceptual clarification, Mathematical reasoning

Approaches and Questions Raised

  • Participants discuss transforming logarithmic expressions into forms suitable for Taylor expansion, specifically considering the use of log(1+x) and its implications. There is uncertainty about the connection between different approaches to the problem.

Discussion Status

Some participants have offered insights into manipulating logarithmic expressions and suggested methods for bounding the Euler-Mascheron constant. There is an ongoing exploration of the relationships between the inequalities and their applications to the problem at hand.

Contextual Notes

Participants note the constraints of the problem, including the requirement to use Taylor series and the potential confusion regarding the interpretation of the problem's parts. There is also mention of the limits of certain logarithmic forms as n approaches infinity.

kde2520
Messages
16
Reaction score
0

Homework Statement


With n>1, show that (a) [tex]\frac{1}{n}[/tex]-ln[tex]\frac{n}{n-1}[/tex]<0
and (b) [tex]\frac{1}{n}[/tex]-ln[tex]\frac{n+1}{n}[/tex]>0

Use these inequalities to show that the Euler-Mascheron constant (eq. 5.28 - page330) is finite.

Homework Equations


This is in the chapter on infinite series, in the section on Taylor Expansion, so I guess Taylor, Maclaurin, and Binomial theorem are fair game.



The Attempt at a Solution


I first wrote the logarithm as a difference of logs and then tried to expand them in the Maclaurin series. But that apparently doesn't work since ln(0) and 1/0 are undefined...

I also don't understand the statement at the end. Is that supposed to be a hint or a third part to the problem?

Any help would be great, thanks.
 
Physics news on Phys.org


For the first two part change the log to log(1+x) so that we want to show
x/(1+x)<log(1+x)<x
for |x|<1
this can be done by using an integral or series representation for log
The third bit is indeed a part of the problem and not a hint, infact one could say the first two parts are a hint for the third
The constant in question is
lim_{n->infinity} [1/1+1/2+1/3+...+1/n-log(n)]
use the given inequalities to bound the constant
 


I'm not exactly sure how to go about changing the log to log(1+x). Do you mean to factor an x out of the demoninator for the first one and then have log(1-1/x), and for the second one log(1+1/x)?
 


I think he means let x= n/(n-1) so that ln(n/(n-1)) becomes just ln x. Of course, then x(n-1)= xn- x= n so n(x-1)= x and n= x/x-1 so that 1/n= (x-1)/x= 1- 1/x. Then your inequality 1/n- ln(n/(n-1))<0 becomes 1- 1/x- ln(x)< 0.
 


Ok, but I'm not seeing how the two approaches are connected since lurflurf said to change the logarithm to ln(1+x).
 


find that log(1+x)~x (x~0)is a simple form to work with
log(x)~-1+x (x~1) amounts to the same thing, matter of taste

we have
1/n-log(n/(n-1))
and
1/n-log((n+1)/n)

n/(n-1)=1+1/(n-1)
let
x=1/(n-1)
n/(n-1)=1+1/(n-1)->1+x
1/n=1/(n-1+1)=[1/(n-1)]/[1+1/(n-1)]->x/(1+x)
1/n-log(n/(n-1))->x/(1+x)-log(1+x)

(n+1)/n=1+1/n
let x=1/n
(n+1)/n=1+1/n->1+x
1/n-log((n+1)/n)->x-log(1+x)

we desire to show
x-log(1+x)>0
and
x/(1+x)-log(1+x)<0
or
x/(1+x)<log(1+x)<x
|x|<1
This is easily done by any number of methods including using series
log(1+x)=x-x/2+x^3/3-x^4/4+...
or integrals
log(1+x)=int(1/(1+t),t,0,x)

once the hint inequalitites are verified use them to bound the constant
 


Ok, that helps a lot. Thanks!
 

Similar threads

Replies
12
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 5 ·
Replies
5
Views
3K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 14 ·
Replies
14
Views
3K
  • · Replies 10 ·
Replies
10
Views
2K
  • · Replies 27 ·
Replies
27
Views
4K
Replies
2
Views
3K
Replies
2
Views
2K
Replies
1
Views
3K