2

I am new to analysis and I have no clue how to solve this limit. This is an exam problem from my analysis 1 course, there are one or two similar ones on the exam.

$$\lim_{n\to \infty}\frac{1}{\sqrt[4]{{n^4}+n+2}}+\cdots+\frac{1}{\sqrt[4]{{n^4}+5n-1}}$$

The only thing I tryed was this silly idea to rewrite it as one single fraction and apply Stolz-Cesaro theorem, but it got way too messy so I doubt that is the way.

I can't find explanations generally on these limits of sequences of the type $\frac{1}{f(x_n)}+\cdots+\frac{1}{f(x_{n+k})}$ (I hope this is a good representation). Should series be involved in solving these kinds of limits ?

EDIT: The limit is supposed to be solved only with the knowledge prior to derivatives and integrals.

Thanks in advance

powerline
  • 499

6 Answers6

2

Hint Assuming that the sum means $$\lim_{n\to \infty}\frac{1}{\sqrt[4]{{n^4}+n+2}}+\frac{1}{\sqrt[4]{{n^4}+n+3}}+\frac{1}{\sqrt[4]{{n^4}+n+4}}+\cdots+\frac{1}{\sqrt[4]{{n^4}+5n-1}}$$

Use that $$\frac{1}{\sqrt[4]{{n^4}+5n-1}}+\frac{1}{\sqrt[4]{{n^4}+5n-1}}+\frac{1}{\sqrt[4]{{n^4}+5n-1}}+\cdots+\frac{1}{\sqrt[4]{{n^4}+5n-1}} \leq \\ \leq \frac{1}{\sqrt[4]{{n^4}+n+2}}+\frac{1}{\sqrt[4]{{n^4}+n+3}}+\frac{1}{\sqrt[4]{{n^4}+n+4}}+\cdots+\frac{1}{\sqrt[4]{{n^4}+5n-1}} \leq \\ \leq \frac{1}{\sqrt[4]{{n^4}+n+2}}+\frac{1}{\sqrt[4]{{n^4}+n+2}}+\frac{1}{\sqrt[4]{{n^4}+n+2}}+\cdots+\frac{1}{\sqrt[4]{{n^4}+n+2}}$$

N. S.
  • 132,525
  • Thanks for your contribution. I guess the squeeze is the best option. I am going to try to finish it. P.S. Could you post in the edit the next step, how would it look like the limit evaluation of the smallest(or largest) sum in the ordering? – powerline Jan 25 '20 at 04:56
2

$$L=\lim_{n \rightarrow \infty} \sum_{k=1}^{4n-1} \frac{1}{(n^4+n+k)^{1/4}}$$ $$L=\lim_{n \rightarrow \infty} \sum_{k=1}^{4n-1}\frac{1}{n} \frac{1}{(1+\frac{1}{n^3}+\frac{k}{n^4})^{1/4}}= \int_{0}^{4} dx=4.$$ Here $1/n^3$, $k/n^3$ can be neglected. Also term $1/(n^4+n+1)^{1/4} \rightarrow 0$ as $n \rightarrow \infty$.

Z Ahmed
  • 43,235
2

We have

$$\sum_{k=1}^{4n-2}\frac{1}{\sqrt[4]{n^4+n+k+1}} = \frac 1n \sum_{k=1}^{4n-2}\frac{1}{\sqrt[4]{1+\frac{n+k+1}{n^4}}}$$

Hence,

$$\underbrace{\frac{4n-2}{n\sqrt[4]{1+\frac{5}{n^3}}}}_{\stackrel{n\to\infty}{\longrightarrow}4} < \frac 1n \sum_{k=1}^{4n-2}\frac{1}{\sqrt[4]{1+\frac{n+k+1}{n^4}}} < \underbrace{\frac{4n-2}{n\sqrt[4]{1+\frac 1{n^3}}}}_{\stackrel{n\to\infty}{\longrightarrow}4}$$

  • I see, the squeeze theorem is to be deployed. I fail to understand how did you get the smallest and the largest expressions in the ordering. Could you elaborate ? Thanks for a nice answer – powerline Jan 25 '20 at 05:20
  • 1
    @GrigoriPerelman : I just saw i missed a factor on the LHS. The estimation comes as follows: Each summand is of the form $\frac{1}{\sqrt[4]{1+\frac{n+k+1}{n^4}}}$. For each summand we have $$\frac{1}{\sqrt[4]{1+\frac{5}{n^3}}}=\frac{1}{\sqrt[4]{1+\frac{5n}{n^4}}}< \frac{1}{\sqrt[4]{1+\frac{n+k+1}{n^4}}} < \frac{1}{\sqrt[4]{1+\frac{n}{n^4}}} = \frac{1}{\sqrt[4]{1+\frac{1}{n^3}}}$$ – trancelocation Jan 25 '20 at 05:30
  • I see, but I miserably fail to understand why do you have 4n-2 in the numerators on the LHS and RHS in the parent answer ? – powerline Jan 25 '20 at 05:39
  • Because there are $4n-2$ summands altogether from $k=1$ to $k=4n-2$. – trancelocation Jan 25 '20 at 05:40
  • Let me ask you one more thing. Did you get the n^4+n+1+k = n^4+5n-1-(4n-2)+k this way ? Is this an usual rule to determine that term in the series ? – powerline Jan 25 '20 at 05:59
  • I did it the other way round. I thought if I let $k$ start at $1$, what should be the maximum $k_{max}$ to reach $5n-1$. So, you get $n+1+k_{max}= 5n-1 \Leftrightarrow k_{max} = 5n-1 -(n+1) = 4n-2$. So, in a sense, it is not so much a rule but a logical way to determine the number of summands which is needed to calculate the sum. – trancelocation Jan 25 '20 at 06:04
  • I meant how did you get the bolded part n^4+n+1+k , the one in the denominator. Not the number above sigma, I get that process. I am looking for a firm rule how to determine the varying term right of the sigma. I hope you get me – powerline Jan 25 '20 at 06:08
  • Ah. From your description of the problem I concluded - since no other specific rule for the increase of the denominator was given - that it was meant to be an increase $1$ by $1$. Fact is, that the problem itself was not properly stated because the rule of the general term was not explicitly stated. So, basically it was a badly stated exam problem and I can imagine that this has created some confusion among students. – trancelocation Jan 25 '20 at 06:14
1

To figure out how many terms are in the sum $$\lim_{n\to \infty}\frac{1}{\sqrt[4]{{n^4}+n+2}}+\frac{1}{\sqrt[4]{{n^4}+n+3}}+\frac{1}{\sqrt[4]{{n^4}+n+4}}+\cdots+\frac{1}{\sqrt[4]{{n^4}+5n-1}}$$

take the $n$ dependence of the last term, and subtract from it the $n$ dependence of the first term. Since we should also count the first term itself we add $1$ to this difference and get the number of terms:

$$({n^4}+5n-1)-({n^4}+n+2)+1=4n-2$$

This leads to the sum

$$\lim_{n\to \infty}\sum_{i=1}^{4n-2}\frac{1}{\sqrt[4]{{n^4}+n+1+i}}$$

Note that in the limit $n\to\infty$ we have for all $i$:

$$\frac{n}{\sqrt[4]{{n^4}+n+1+i}}\to 1$$

This means that in the limit, each summand can be equivalently replaced by $1/n$, and summing this leading contribution $4n-2$ times, we get

$$\frac{4n-2}{n}\to 4$$

in the strict limit.

Kagaratsch
  • 2,239
  • Thanks a lot for you contribution, but I forgot to emphasize that the limit is supposed to be solved only with the tools prior to derivatives and integrals. I made an edit in the post so that is clear now. – powerline Jan 25 '20 at 04:43
  • 1
    @GrigoriPerelman see my updated answer. – Kagaratsch Jan 25 '20 at 04:53
  • "the coefficient of n in the denominator goes from 1⋅n to 5⋅n, which means that roghly 4n terms are being summed". Could you elaborate on how did you conclude that roughly 4n terms are being summed, i.e. how did you find out that uper boundary of the sigma ? Thanks for the answer, btw I am not familiar with the O-sign notation, so I guess I fail to understand your argument fully. – powerline Jan 25 '20 at 05:00
  • 1
    @GrigoriPerelman see updated answer, the big O notation is not really necessary here, so I removed it. – Kagaratsch Jan 25 '20 at 05:13
  • Ok, now I get it. Great answer, one thing though. You said "Note that the leading behavior in the limit n→∞ is: ..." I see that, but I am not sure if my proffesor would acknowledge that argument, as it's a purely observation/approximation not referring to any rule or property. How would you clarify that the expression is ≈1/n ? – powerline Jan 25 '20 at 05:27
  • 1
    @GrigoriPerelman I rewrote it a little bit. Basically, the problem is reduced from taking a sum, to showing the limit on each summand. Taking into account the possible values of $i$ you can fill in the details. – Kagaratsch Jan 25 '20 at 05:30
1

If $n\ge 1$ then $$4-\frac {2}{n}=\frac {4n-2}{n}=\sum_{j=n+1}^{5n-2}\frac {1}{n}>$$ $$>\sum_{j=n+1}^{5n-2}\frac {1}{(n^4+j)^{1/4}}>$$ $$>\sum_{j=n+1}^{5n-2}\frac {1}{n+1}=$$ $$=\frac {4n-2}{n+1}=4-\frac {6}{n+1}$$ because if $n\ge 1$ and $n+1\le j\le 5n-2$ then $$0<n^4< n^4+j\le$$ $$\le n^4 +5n-2 <$$ $$< n^4+4n+6n+4n+1\le$$ $$\le n^4+4n^3+6n^2+4n+1=(n+1)^4.$$

-1

Also note that $\frac{1}{n+1}<\frac{1}{n^4+n+k} <\frac{1}{n}$ so $$ \sum_{k=1}^{4n-1} \frac{1}{n+1}<S_n=\sum_{k=1}^{4n-1} \frac{1}{(n^4+n+k)^{1/4}} <\sum_{k=1}^{4n-1} \frac{1}{n}.$$ So, we note that $$ \lim_{n \rightarrow \infty} \frac{4n-1}{n+1}=4= \lim_{n \rightarrow \infty} \frac{4n-1}{n}= \lim_{n \rightarrow \infty} S_n$$

When $n \rightarrow \infty$ the first term of the sequence $1/(n^4+n+2)^{1/4} \rightarrow 0$ it can well be neglected.

Note that $$\frac{1}{(n^4+n+k)^{1/4}} > \frac{1}{n+1}$$ can be checked to be true for all $1<k\le 4n-1$

Z Ahmed
  • 43,235