Solving Lim x→∞ (x - ln x): Step by Step Guide

  • Thread starter Thread starter r_swayze
  • Start date Start date
  • Tags Tags
    Ln
r_swayze
Messages
65
Reaction score
0
Can somebody show a step by step process of how to get this answer?

I keep getting -∞ when it should be ∞

heres my work:

lim x→∞ (x - ln x) = lim x→∞ ((1 - 1/x ln x) / (1/x)

then apply l'hospitals rule

lim x→∞ (x^(-2) ln x - x^(-2)) / (-x^(-2))

then I cancel all the x^(-2) and I am left with -ln x which would equal to -∞
 
Physics news on Phys.org
You can apply L'Hopital's rule only when you have an indeterminate form of [0/0] or [+/-infinity/infinity].

In this expression -- ((1 - 1/x ln x) / (1/x) -- 1/x approaches zero, but ln x approaches infinity, which makes it another indeterminate form.

I don't think that there's anything you can do with L'Hopital's rule on this one. You might try an approach that looks at the derivative of f(x) = x - ln x. It's easy to show that the derivative approaches 1 as x gets large, which suggests that x - ln x gets larger and larger.
 
Why don't you apply l'Hopital to the ln(x)/x part of your expression and use the result to simplify it?
 
(At the risk of confusing things) or take the first x into the ln(x).
 
Remember that ln(x) = \int_1^x \frac{1}{u} du and use the MVT
 
Dick said:
Why don't you apply l'Hopital to the ln(x)/x part of your expression and use the result to simplify it?

where does the ln(x)/x come from?
 
This is what I would do:

\stackrel{lim}{x\rightarrow\infty}(x-ln(x))\times\frac{(x+ln(x))}{(x+ln(x))}

This would give:\stackrel{lim}{x\rightarrow\infty}\frac{(x^{2}-ln^{2}(x))}{(x+ln(x))}

You could work it out from there, split the limit up by x^{2} and ln^{2}x

Hope this helped, and sorry about the poor formatting, this is probably the first time I am doing latex.

By the way, I'm sure the value is infinity.
 
You could do lim x→∞ (ln e^x - ln x) = lim x→∞ ln (e^x /x) and apply l'hopital's rule inside; the domain change of x to ln e^x is irrelevant because your limit exists within the domain of log.
 
zcd said:
You could do lim x→∞ (ln e^x - ln x) = lim x→∞ ln (e^x /x) and apply l'hopital's rule inside; the domain change of x to ln e^x is irrelevant because your limit exists within the domain of log.

I think I see it, but just to make sure, does ln e equal 1? so that can be plugged into any equation?
 
  • #10
Yes, ln e= 1.
 
  • #11
r_swayze said:
where does the ln(x)/x come from?

Write it (as you basically did) as x*(1-ln(x)/x).
 
Back
Top