# Analysis Divergent Sequences

MathSquareRoo

## Homework Statement

Prove that the given sequence diverges to infinity.
{an} = (-n^4+n^3+n)/(2n+7)

## Homework Equations

Diverges definition

## The Attempt at a Solution

So far I have:

Let M>0 and let N= something.

I'm having a hard time figuring out what N should equal for the proof. I know I have to make the numerator smaller and denominator larger (I think), but I get confused. Can someone explain it please?

Homework Helper
Gold Member

## Homework Statement

Prove that the given sequence diverges to infinity.
{an} = (-n^4+n^3+n)/(2n+7)

## Homework Equations

Diverges definition

## The Attempt at a Solution

So far I have:

Let M>0 and let N= something.

I'm having a hard time figuring out what N should equal for the proof. I know I have to make the numerator smaller and denominator larger (I think), but I get confused. Can someone explain it please?

The first thing you must notice is with that minus sign in the numerator on the ##n^4## term, this isn't going to go to ##+\infty##. It will go to ##-\infty##. Let's call your fraction ##f(n)## for the moment. Start by writing the careful definition of the statement$$\lim_{n\rightarrow \infty}f(n) = -\infty$$Then you will know what you must prove.

[Edit, added later]: Alternatively you could show ##|f(n)|\rightarrow \infty## and observe that ##f(n)<0## for large ##n##.

Last edited:
MathSquareRoo
Yes, I understand that it will diverge to negative infinity. I still need help understanding how to get the N for the proof. Can you help?

Homework Helper
Gold Member
Yes, I understand that it will diverge to negative infinity. I still need help understanding how to get the N for the proof. Can you help?

I don't know if you saw may later edit, which is the method I would use. Given ##M>0##, to find ##N## you generally use an exploratory argument which is kind of working backwards with estimates. So start with$$\left | \frac{-n^4+n^3+n}{2n+7}\right |\ge ...$$and underestimate it to get something workable. I would start with the old calculus trick of dividing numerator and denominator by the highest power of ##n##. Then work on making the denominator larger and the numerator smaller to underestimate it with something that still goes to infinity.

: Edited it twice. I had it right the first time.

Last edited:
Homework Helper
For sufficiently large n, the highest degree term will dominate. For sufficiently large n, (-n^4+n^3+n)/(2n+7) will be approximately -(n^3)/2.

MathSquareRoo
I'm sorry, I still don't understand how to write the proof for this divergent sequence. I have a similar problem done correctly I believe. Can someone help me write a proof like the following one, but for this sequence above??

Here is a similar lim, I think I have correctly proved.

{an}=(n^2+1)/(n-2)

Proof: Let M>0 and let N=2M. Then n>N implies 1/2n>M which implies (n^2+1)/(n-2)<n^2/2n=1/2n>M. Hence, lim n--> {an}=+∞.

Homework Helper
Gold Member
I don't know if you saw may later edit, which is the method I would use. Given ##M>0##, to find ##N## you generally use an exploratory argument which is kind of working backwards with estimates. So start with$$\left | \frac{-n^4+n^3+n}{2n+7}\right |\ge ...$$and underestimate it to get something workable. I would start with the old calculus trick of dividing numerator and denominator by the highest power of ##n##. Then work on making the denominator larger and the numerator smaller to underestimate it with something that still goes to infinity.

: Edited it twice. I had it right the first time.

I'm sorry, I still don't understand how to write the proof for this divergent sequence. I have a similar problem done correctly I believe. Can someone help me write a proof like the following one, but for this sequence above??

Here is a similar lim, I think I have correctly proved.

{an}=(n^2+1)/(n-2)

Proof: Let M>0 and let N=2M. Then n>N implies 1/2n>M which implies (n^2+1)/(n-2)<n^2/2n=1/2n>M. Hence, lim n--> {an}=+∞.

Did you try what I suggested? If so, how far have you gotten? Show us.

MathSquareRoo
I get (-n^3/2). Then I"m not sure how to solve for the N that I need.

Homework Helper
Gold Member
I get (-n^3/2). Then I"m not sure how to solve for the N that I need.

I have no idea where you got that but, as an example of solving for ##N##, you are given an ##M## and you want ##n^{3/2} > M## for ##n > N##. So solve the inequality ##n^{3/2} > M## for ##n## to see how large ##n## needs to be. Then choose ##N## so that if ##n>N## you can guarantee ##n## is as large as it needs to be to make it work.

MathSquareRoo
I'm sorry, I meant (-n^3)/2. So do I set (-n^3)/2 > M and solve for n?

Homework Helper
Gold Member
I'm sorry, I meant (-n^3)/2. So do I set (-n^3)/2 > M and solve for n?

Yes, same idea. You want the absolute value though: ##\frac{n^3}{2}>M##. So how big does ##n## need to be? What ##N## will work?

MathSquareRoo
N>(2M)^1/3

Is that correct?

That is how you determine how large N needs to be with an exploratory argument. Now, at this point, we have, if ##n>N## $$\left | \frac{-n^4+n^3+n}{2n+7}\right |\ge ...\hbox{ your steps here }...\ge n^{\frac 3 2}> M$$
To write up the argument you would phrase it like this: Given ##M>0## if ##n>(2M)^\frac 1 3## then$$\left | \frac{-n^4+n^3+n}{2n+7}\right |\ge ...\hbox{ your steps here }...\ge n^{\frac 3 2}> M$$Notice in this last paragraph, it looks like you magically came up with ##n>(2M)^\frac 1 3##, but it came from the exploratory argument.