Help with proving that Improper Integral is Divergent

Click For Summary

Homework Help Overview

The problem involves proving that an improper integral diverges, specifically the integral of 1/(x ln x) from 1 to infinity. The subject area is calculus, focusing on improper integrals and convergence tests.

Discussion Character

  • Exploratory, Assumption checking, Conceptual clarification

Approaches and Questions Raised

  • Participants discuss the use of limits and substitution methods, particularly u-substitution, to evaluate the integral. There are questions about the validity of expressions involving infinity, such as ∞ - ∞ and ∞ + ∞. Some participants suggest exploring the Direct Comparison Theorem as an alternative method to prove divergence.

Discussion Status

The discussion is ongoing, with participants exploring various approaches and questioning assumptions about the mathematical expressions involved. Some guidance has been offered regarding the nature of indeterminate forms and the application of different convergence tests, but no consensus has been reached on a definitive method for proving divergence.

Contextual Notes

Participants are grappling with the implications of using limits that involve infinity and the conditions under which certain mathematical operations are valid. There is a focus on ensuring that the reasoning aligns with established mathematical principles.

student93
Messages
83
Reaction score
0

Homework Statement



The problem is attached in this post.

Homework Equations



The problem is attached in this post.

The Attempt at a Solution



Lim t -> ∞ ∫ dx/xlnx from 1 to t

u-substitution:

u=lnx
du=1/x dx

Lim t -> ∞ ∫ 1/u du

Lim t -> ∞ ln u

Lim t -> ∞ ln(lnx) from 1 to t

Lim t -> ∞ ln(lnt) - ln(0)

= ∞ - ∞ = 0 (This is incorrect since the answer is that the integral is divergent).
 

Attachments

  • Screen shot 2014-03-06 at 1.59.24 AM.png
    Screen shot 2014-03-06 at 1.59.24 AM.png
    8.3 KB · Views: 558
Physics news on Phys.org
student93 said:

Homework Statement



The problem is attached in this post.


Homework Equations



The problem is attached in this post.

The Attempt at a Solution



Lim t -> ∞ ∫ dx/xlnx from 1 to t

u-substitution:

u=lnx
du=1/x dx

Lim t -> ∞ ∫ 1/u du

Lim t -> ∞ ln u

Lim t -> ∞ ln(lnx) from 1 to t

Lim t -> ∞ ln(lnt) - ln(0)

= ∞ - ∞ = 0 (This is incorrect since the answer is that the integral is divergent).

##\ln(1) = 0##, not ##\infty##.

Anyway, never, ever use expressions of the form ##\infty - \infty##; these are meaningless and are illegal in mathematics.
 
Last edited:
That's true but it's not ln(1), it's ln(ln(1)) which is ln(0), which is equal to ∞, right?
 
student93 said:
That's true but it's not ln(1), it's ln(ln(1)) which is ln(0), which is equal to ∞, right?

OK, you're right. However, ##\int_a^b dx/(x \ln x) = \ln \ln b - \ln \ln a \to \infty + \infty## as ##a \to 1+## and ##b \to \infty##. As ##a \to 1+##, ##\ln a \to 0+## and so ##\ln \ln a \to -\infty##, or ##-\ln \ln a \to +\infty##.
 
Ray Vickson said:
OK, you're right. However, ##\int_a^b dx/(x \ln x) = \ln \ln b - \ln \ln a \to \infty + \infty## as ##a \to 1+## and ##b \to \infty##. As ##a \to 1+##, ##\ln a \to 0+## and so ##\ln \ln a \to -\infty##, or ##-\ln \ln a \to +\infty##.

Is there another way to prove that this integral is divergent? (For example, would the Direct Comparison Theorem work in this case?)

Also, why is ∞ + ∞ mathematically allowed, if ∞ - ∞ isn't?
 
student93 said:
Is there another way to prove that this integral is divergent? (For example, would the Direct Comparison Theorem work in this case?)

Also, why is ∞ + ∞ mathematically allowed, if ∞ - ∞ isn't?

∞ - ∞ is meaningless (along with 0/0, for example). ∞-∞ can be made equal to anything you want---anything from -∞ to +∞ and everything in between. However, ∞+∞ = +∞ is unambiguous. Think about it and you will understand why.
 
Ray Vickson said:
∞ - ∞ is meaningless (along with 0/0, for example). ∞-∞ can be made equal to anything you want---anything from -∞ to +∞ and everything in between. However, ∞+∞ = +∞ is unambiguous. Think about it and you will understand why.

So ∞-∞ is an Indeterminate Form then right? So can I just L'Hopital's rule to solve for the limit instead of the other method that you showed in your previous post?
 
student93 said:
So ∞-∞ is an Indeterminate Form then right? So can I just L'Hopital's rule to solve for the limit instead of the other method that you showed in your previous post?

So by using L'Hopital's rule, could I take the derivative of ln(lnx) - ln(ln1) and then try solving for the limit etc.?
 
Last edited:
student93 said:
So by using L'Hopital's rule, could I take the derivative of ln(lnx) - ln(ln1) and then try solving for the limit etc.?


You keep missing the point: you do not have an indeterminate form, and do not need anything like l'Hospital's rule. I suggest you look much more carefully at what you are doing, and don't rush. Anyway, I already showed you a correct argument, but for some reason you are just not "getting" it.
 
  • #10
Ray Vickson said:
You keep missing the point: you do not have an indeterminate form, and do not need anything like l'Hospital's rule. I suggest you look much more carefully at what you are doing, and don't rush. Anyway, I already showed you a correct argument, but for some reason you are just not "getting" it.

Ok, I think I now understand your explanation, however would it also be possible to prove this integral is divergent via the Direct Comparison Test? And if it is possible, then how would I prove that the integral is divergent via the Direct Comparison Test?
 

Similar threads

Replies
5
Views
2K
  • · Replies 5 ·
Replies
5
Views
3K
  • · Replies 7 ·
Replies
7
Views
10K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 4 ·
Replies
4
Views
4K
  • · Replies 2 ·
Replies
2
Views
2K
Replies
3
Views
2K
  • · Replies 14 ·
Replies
14
Views
2K
  • · Replies 105 ·
4
Replies
105
Views
10K
  • · Replies 12 ·
Replies
12
Views
3K