Are these 100 series enough?

nasi112

Full Member
Joined
Aug 23, 2020
Messages
677

Stuck at number 4. I don't understand his answer. Would the answer change if we started at n = 0?
 
You are more likely to get help if you provide more information instead of expecting us to watch 6 hour video.
Here what would be helpful:
 
It was a typo. I meant at n = 2. The guy in the video changed n = 1619.

n=21(lnn)lnn\sum_{n=2}^{\infty}\frac{1}{(\ln n)^{\ln n}} n=16191(lnn)lnn\sum_{n=1619}^{\infty}\frac{1}{(\ln n)^{\ln n}}
Do these two series have the same answer?
 
Do these two series have the same answer?
They do have the same boolean answer, i.e., they both either converge or both diverge:
n=2xn=n=21618xn+n=1619xn\sum_{n=2}^\infty x_n = \sum_{n=2}^{1618} x_n + \sum_{n=1619}^\infty x_n
 
Stuck at number 4. I don't understand his answer.
Do these two series have the same answer?

I haven't bothered to watch the video. Could you at least tell us what the question is?

In particular, is it about values of sums, or about convergence of series?

Assuming it's the latter, do you see why the initial value would not matter?
 
I haven't bothered to watch the video. Could you at least tell us what the question is?

In particular, is it about values of sums, or about convergence of series?

Assuming it's the latter, do you see why the initial value would not matter?
I watched some of it, and I believe the question is about convergence, not about the value of the sum (yes, the sum does converge). But I still don't understand what kind of help the OP needs.
 
I found his answer pretty clear after the first 5 minutes: which part of it are you having difficulty with?
The discovery of 1n2\frac{1}{n^2} to use the comparaison test is hard to find. An average student would never find that. My initial thought was to use 1lnn\frac{1}{\ln n}. Apparently it didn't work. After watching the video multiple times, I understood what the guy did. For an average student, is there another way to solve the problem?
 
The discovery of 1n2\frac{1}{n^2} to use the comparaison test is hard to find. An average student would never find that. My initial thought was to use 1lnn\frac{1}{\ln n}. Apparently it didn't work. After watching the video multiple times, I understood what the guy did. For an average student, is there another way to solve the problem?
Not all problems are made to be easy for an average student. I agree that this particular one isn't trivial, and, to be honest, I had no clue how to solve it until I watched the video.
 
I still haven't looked, now because the goal is to see if a student who hasn't see the video could work it out.

I first thought of the root test, and if my scratchings are correct, then that seems to work (with a little L'Hopital thrown in).

But I agree that I wouldn't expect every such problem to be relatively easy, or for it to be at all obvious what method to try. You just try a few things and see what helps.
 
Are you suggesting limn1(lnn)lnnlnn\displaystyle \lim_{n \to \infty}\sqrt[\ln n]{\frac{1}{(\ln n)^{\ln n}}}?
 
No. What does the root test say?

 
No. What does the root test say?

That what I did. I can't tell the difference.
 
Here is a (hopefully) self-contained summary of the video segment:
Question: does this infinite series converge?:
n=21(lnn)lnn\sum_{n=2}^\infty \frac{1}{(\ln n)^{\ln n}}Answer: yes, the series converges.
Proof: for large enough nn we have an=(lnn)lnn>n2a_n={(\ln n)^{\ln n}} > n^2:
an=elnan=elnnlnlnn=(elnn)lnlnn=nlnlnna_n = e^{\ln a_n} = e^{\ln n \ln\ln n} = \left(e^{\ln n} \right)^{\ln\ln n} = n^{\ln\ln n}so all we need is to have is: lnlnn>2lnn>e2n>ee21618,2\ln\ln n > 2 \equiv \ln n > e^2 \equiv n > e^{e^2} \approx 1618,2
 
Usually. But sometimes, when you don't see that any of the familiar methods exactly fits, you can try using methods in slightly different ways. And they may work.
I understood what the guy did. For an average student, is there another way to solve the problem?
This is what you were asking for: an alternate method. What he does in the video is to stretch the comparison test a bit, by trying something that is not obvious, and eventually making it work; my claim is that you can (again, if I did it right) stretch the root test a bit an make it work here, even though it won't be as direct as in the sort of problem you mention. In both cases, the theorem is still true, even though it takes extra work to apply it.

I wouldn't call either approach easy; but sometimes life works that way.
 
I asked to solve the problem with another method that suits an average student. Even if your approach doesn't fit for me you can write it. I may understand it because I know the root test and I know the L'hopital. I'm still confused of how you will do it. Can you please start your approach to see what I didn't see?
 
Top