Analysis of Convergence for Series a) and b)

Click For Summary
SUMMARY

The series a) $\displaystyle \sum\frac{1}{n!}\left(\frac{n}{e}\right)^n$ diverges, as shown by applying Stirling's formula and comparing it to the divergent series $\displaystyle \sum\frac{1}{\sqrt{2\pi n}}$. In contrast, series b) $\displaystyle\sum\frac{(-1)^n}{n!}\left(\frac{n}{e}\right)^n$ converges, validated by Leibniz's test for alternating series. The discussion also highlights a common mistake regarding the comparison of integrals and sums, emphasizing the importance of correctly applying inequalities in convergence tests.

PREREQUISITES
  • Understanding of series convergence and divergence
  • Familiarity with Stirling's formula
  • Knowledge of Leibniz's test for alternating series
  • Basic calculus concepts, including integrals and logarithms
NEXT STEPS
  • Study the application of Stirling's formula in series analysis
  • Learn more about convergence tests, specifically the Ratio Test and Root Test
  • Explore advanced topics in asymptotic analysis
  • Investigate the properties of alternating series and their convergence criteria
USEFUL FOR

Mathematicians, students studying calculus or real analysis, and anyone interested in series convergence and divergence techniques.

alexmahone
Messages
303
Reaction score
0
Do the following series converge or diverge?

a) $\displaystyle \sum\frac{1}{n!}\left(\frac{n}{e}\right)^n$

b) $\displaystyle\sum\frac{(-1)^n}{n!}\left(\frac{n}{e}\right)^n$

My attempt:

a) $\displaystyle\frac{1}{n!}\left(\frac{n}{e}\right)^n\approx\frac{1}{\sqrt{2\pi n}}$ (Stirling’s formula)

$\displaystyle \sum\frac{1}{n!}\left(\frac{n}{e}\right)^n\approx \sum\frac{1}{\sqrt{2\pi n}}$

Since $\displaystyle\sum\frac{1}{\sqrt{2\pi n}}$ diverges, $\displaystyle \sum\frac{1}{n!}\left(\frac{n}{e}\right)$ also diverges.

b) $\displaystyle\frac{(-1)^n}{n!}\left(\frac{n}{e}\right)^n\approx\frac{(-1)^n}{\sqrt{2\pi n}}$ (Stirling’s formula)

$\displaystyle \sum\frac{(-1)^n}{n!}\left(\frac{n}{e}\right)^n\approx \sum\frac{(-1)^n}{\sqrt{2\pi n}}$

By Leibniz's test for alternating series, $\displaystyle\sum\frac{(-1)^n}{\sqrt{2\pi n}}$ converges.

So, $\displaystyle\sum\frac{(-1)^n}{n!}\left(\frac{n}{e}\right)^n$ also converges.
 
Last edited:
Physics news on Phys.org
Is my solution correct?
 
chisigma said:
Setting $\displaystyle a_{n}= \frac{1}{n!}\ (\frac{n}{e})^{n}$ You have...

$\displaystyle \ln a_{n}= n\ (\ln n -\ln e) -\ln n!= n\ (\ln n-1) - \sum_{k=1}^{n} \ln k$ (1)

Now You have...

$\displaystyle n\ (\ln n-1)= \int_{1}^{n} \ln x\ dx > \sum_{k=1}^{n} \ln k$ (2)

... so that is...

$\displaystyle \lim_{n \rightarrow \infty} \ln a_{n} >0 \implies \lim_{n \rightarrow \infty} a_{n} >1$ (3)

... and the necessary condition for convergence is not verified...

Kind regards

$\chi$ $\sigma$

But what's wrong with my solution?
 
My previous post was wrong because I incorrectly wrote $\displaystyle \int_{1}^{n} \ln x\ dx > \sum_{k=1}^{n} \ln k$ and is $\displaystyle \int_{1}^{n} \ln x\ dx < \sum_{k=1}^{n} \ln k$...

Very sorry!... any thanks is absolutely a nonsense!...

Kind regards

$\chi$ $\sigma$
 
Last edited:
chisigma said:
Writing $\displaystyle \lim_{n \rightarrow \infty} \frac{\sqrt{2\ \pi\ n}}{n!}\ (\frac{n}{e})^{n}=1$ doesn't mean $\displaystyle \frac{1}{n!}\ (\frac{n}{e})^{n}= \frac{1}{\sqrt{2\ \pi\ n}}$...

Kind regards

$\chi$ $\sigma$

But aren't they approximately equal for large $n$?

---------- Post added at 03:27 AM ---------- Previous post was at 03:09 AM ----------

chisigma said:
My previous post was wrong because I incorrectly wrote $\displaystyle \int_{1}^{n} \ln x\ dx > \sum_{k=1}^{n} \ln k$ and is $\displaystyle \int_{1}^{n} \ln x\ dx < \sum_{k=1}^{n} \ln k$...

Very sorry!... any thanks is absolutely a nonsense!...

Kind regards

$\chi$ $\sigma$

Does that mean my solution is right?
 
Alexmahone said:
But aren't they approximately equal for large $n$?

---------- Post added at 03:27 AM ---------- Previous post was at 03:09 AM ----------



Does that mean my solution is right?

Rewriting my previous post we have...

$\displaystyle \int_{1}^{n} \ln x\ dx - \sum_{k=1}^{n} \ln k = \sum_{k=1}^{n} \int_{k}^{k+1} (ln x - \ln (k+1))\ dx \sim - \sum_{k=1}^{n} \frac{1}{k+1}$

... so that is $\displaystyle \lim_{n \rightarrow \infty} a_{n}=0$ and Your solution is probably correct...

Kind regards

$\chi$ $\sigma$
 

Similar threads

  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 11 ·
Replies
11
Views
3K
  • · Replies 16 ·
Replies
16
Views
4K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 15 ·
Replies
15
Views
3K
  • · Replies 6 ·
Replies
6
Views
4K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 5 ·
Replies
5
Views
3K
  • · Replies 3 ·
Replies
3
Views
2K