Check for the convergence or divergence of the following series

Click For Summary
SUMMARY

The discussion focuses on the convergence of two series: 1) \( \sqrt{n}(1 - \cos(1/n)) \) and 2) a series defined by \( a_n = \frac{1}{n + \sqrt{n}} \) for odd \( n \) and \( a_n = -\frac{1}{n} \) for even \( n \). The first series converges, supported by the Taylor expansion of cosine and the fact that \( |\cos(1/n)| \leq 1 \). The second series is conditionally convergent, resembling the alternating series \( (-1)^n/n \), but requires careful analysis of the absolute values to confirm convergence.

PREREQUISITES
  • Understanding of series convergence tests, including the Integral Test and Comparison Test.
  • Familiarity with Taylor series expansions, particularly for trigonometric functions.
  • Knowledge of alternating series and their convergence criteria.
  • Basic algebraic manipulation of series and limits.
NEXT STEPS
  • Study the Taylor expansion of cosine to understand its application in series convergence.
  • Learn about the Alternating Series Test and its conditions for convergence.
  • Explore the concept of absolute convergence and its implications for series.
  • Investigate the Integral Test for series convergence in more depth.
USEFUL FOR

Mathematics students, educators, and anyone studying series convergence in calculus or advanced mathematics courses.

freshman2013
Messages
43
Reaction score
0

Homework Statement


Here are some series I'm completely stuck on.
1.sqrt(n)*(1-cos(1/n))

2. a series in which if n is odd, then an is 1/(n+\sqrt[]{n}) while if n is even, then an is -1/n


Homework Equations





The Attempt at a Solution


For 1., I tried integral test which seemed impossible to integrate, and then I tried comparison test but i can't find anything to compare to

For 2. I thought that as n approaches infinity for 1/(n+\sqrt[]{n}), which equals (1/\sqrt{n})/(\sqrt{n}+1), then that expression pretty much looks like 1/n and thus the series is approx (-1)^n/n and is thus conditionally convergent like (-1)^n/n. But I then realized perhaps you can't apply the alternating series test since in the original series, the abs value of it is not decreasing for every term.E.g.1/(100+root(100))=1/110 < 1/101 Anyone have any idea on how to tackle this, or am I right from my original attempt?
 
Last edited:
Physics news on Phys.org
freshman2013 said:

Homework Statement


Here are some series I'm completely stuck on.
1.sqrt(n)*(1-cos(1/n))

2. a series in which if n is odd, then an is 1/(n+\sqrt[]{n}) while if n is even, then an is -1/n


Homework Equations





The Attempt at a Solution


For 1., I tried integral test which seemed impossible to integrate, and then I tried comparison test but i can't find anything to compare to

For 2. I thought that as n approaches infinity for 1/(n+\sqrt[]{n}), which equals (1/\sqrt{n})/(\sqrt{n}+1), then that expression pretty much looks like 1/n and thus the series is approx (-1)^n/n and is thus conditionally convergent like (-1)^n/n. But I then realized perhaps you can't apply the alternating series test since in the original series, the abs value of it is not decreasing for every term.E.g.1/(100+root(100))=1/110 < 1/101 Anyone have any idea on how to tackle this, or am I right from my original attempt?

Let's start with #1.

Your series is: ##\sum_{n=1}^{∞} \sqrt{n} - \sqrt{n}cos(\frac{1}{n})##

I know this may seem misleading, but this series appears to converge.

Hint: ##|cos(\frac{1}{n})| ≤ 1##
 
Zondrina said:
Let's start with #1.
I know this may seem misleading, but this series appears to converge.

Hint: ##|cos(\frac{1}{n})| ≤ 1##
Yes, it converges, but you need to use a stronger fact about cos of small angles. Freshman2013, do you know the Taylor expansion of cos?
 
For the first problem, this trig identity may be useful: ##2\sin^2(x) = 1 - \cos(2x)##.
 
For problem 2, try adding ##a_n## and ##a_{n+1}##, where ##n## is odd.
 
so for the second one, could I do this:
the first term 1/(n+\sqrt{n}) for all odd n, I rewrite as a separate series 1/(2n-1+\sqrt{2n-1}) and the 2nd term -1/n for all even n I rewrite as -1/(2n)

I combine those two series and I get (1-\sqrt{2n-1})/(4n^2-2n+2n\sqrt{2n-1}) and test for that(which is convergent). And I do the same except with 1/2n instead -1/2n to check for absolute convergence.
 
freshman2013 said:
And I do the same except with 1/2n instead -1/2n to check for absolute convergence.
For absolute convergence, you can do something simpler. I claim that if there is some subsequence ##(n_k)_{k=1}^{\infty}## such that ##\sum_{k=1}^{\infty} |a_{n_k}|## diverges, then ##\sum_{n=1}^{\infty} |a_n|## must also diverge. If you can prove this, then you can apply it immediately to your series.
 

Similar threads

  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 17 ·
Replies
17
Views
2K
  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
Replies
14
Views
2K