Convergence of (ln n)^x

  • I
  • Thread starter Happiness
  • Start date
  • #1
665
26

Main Question or Discussion Point

Find the values of ##x## for which the following series is convergent.
Screen Shot 2016-06-13 at 1.25.44 am.png


I compared the series with the harmonic series and deduced it is always divergent. I used ##y^p<e^y## for large ##y##. I used a different method from the answer given, which I don't understand.

Screen Shot 2016-06-13 at 1.25.21 am.png


When ##k=1##, ##M_1=e=2.72## and ##M_0=1##. Since ##M_0+1\leq n\leq M_1##, we have ##n=2##, giving us the term ##\frac{1}{1^X}##.

When ##k=2##, ##M_2=e^2=7.39## and ##M_1=e=2.72##. Since ##M_1+1\leq n\leq M_2##, we have ##n=4, 5, 6, 7##, giving us the terms ##\frac{1}{2^X}+\frac{1}{2^X}+\frac{1}{2^X}+\frac{1}{2^X}##.

Finding the possible values of ##n## is troublesome.

Is there a typo in the given answer?
 
Last edited:

Answers and Replies

  • #2
34,045
9,895
You get of the order of e terms of 1/1X, e2 terms of 1/2X, e3 terms of 1/3X and so on (up to the constant prefactor). Following the same logic as your approximation, ##\displaystyle \frac{e^n}{n^X}## grows to infinity instead of going to zero.
 
  • #3
665
26
You get of the order of e terms of 1/1X, e2 terms of 1/2X, e3 terms of 1/3X and so on (up to the constant prefactor). Following the same logic as your approximation, ##\displaystyle \frac{e^n}{n^X}## grows to infinity instead of going to zero.
Why is the series in the question and the series in the answer equivalent?

Their first terms are clearly different. The former's is ##\frac{1}{(\ln 2)^X}## while the latter's is ##\frac{1}{1^X}##.

If they are not equivalent, how do we show that the former is always bigger than the latter for every term?
 
  • #4
34,045
9,895
The first fraction is always larger than the second by construction of the Mk. So your series is larger than a divergent series.
 
  • #5
665
26
The first fraction is always larger than the second by construction of the Mk. So your series is larger than a divergent series.
Could you explain how? I don't see it.
 
  • #6
34,045
9,895
ln(2) < 1
ln(3) < 2, ln(4) < 2, ln(5) < 2, ln(6) < 2, ln(7) < 2

Therefore 1/ln(2)X > 1/1X and so on for positive X.
 
  • #7
665
26
The given answer is so complicated vs if we just compare the series with the harmonic series.
 
  • #8
34,045
9,895
The given answer is so complicated vs if we just compare the series with the harmonic series.
How do you do that for x=-3?
 
  • #9
665
26
How do you do that for x=-3?
Let ##y=\ln n##. For large ##n, y^3<e^y##. So ##(\ln n)^3<n##. Then ##\frac{1}{(\ln n)^3}>\frac{1}{n}##. The series is bigger than the harmonic series for every term. Hence it is also divergent.
 
  • #10
34,045
9,895
Ah right, that works as well, and it is easier.
 

Related Threads on Convergence of (ln n)^x

Replies
2
Views
3K
Replies
2
Views
882
Replies
7
Views
11K
  • Last Post
Replies
2
Views
1K
  • Last Post
Replies
1
Views
2K
  • Last Post
Replies
6
Views
6K
  • Poll
  • Last Post
Replies
17
Views
5K
  • Last Post
Replies
2
Views
2K
  • Last Post
Replies
2
Views
9K
  • Last Post
Replies
9
Views
2K
Top