The Student Room Group

Limits help

Hi everyone, I currently have this problem to prove.




I am aware of the product rule (that if there exist sequences An and Bn which both have limits, then the limit of AnBn is the product of the 2 limits). I know that An being null has a limit of 0. However, the question says that the sequence Bn is bounded. I know that a sequence which is bounded does not necessarily have a limit, e.g. (-1)^n, and so I don't think I can use the product rule here.

Any help would be very appreciated!

EDIT: Solved!
(edited 9 years ago)
Original post by Dannygem
Hi everyone, I currently have this problem to prove.




I am aware of the product rule (that if there exist sequences An and Bn which both have limits, then the limit of AnBn is the product of the 2 limts).
Although, the question says that the sequence Bn is bounded. However, I know that a sequence which is bounded does not necessarily have a limit, e.g. (-1)^n.

Any help would be very appreciated!

You could just blindly start writing down epsilons. Have you tried writing down what it means for (a_n) to be a null sequence, and for (b_n) to be bounded?
Reply 2
Original post by Smaug123
You could just blindly start writing down epsilons. Have you tried writing down what it means for (a_n) to be a null sequence, and for (b_n) to be bounded?


For An to be a null sequence it means that there exists a NeN+ such that n> N => l an ( - 0 ) l < E

For Bn to be bounded it means there exists UeN and LeN such that for all values of n we have L < Bn < U

That is correct right?
That's all i've got wrote down, not sure how to advance with the problem from there
(edited 9 years ago)
Original post by Dannygem
For An to be a null sequence it means that there exists a NeN+ such that n> N => l an ( - 0 ) l < E

For Bn to be bounded it means there exists UeN and LeN such that for all values of n we have L < Bn < U

That is correct right?
That's all i've got wrote down, not sure how to advance with the problem from there

Yes. Can you state what you need to prove, in epsilon-terms?
Reply 4
Original post by Smaug123
Yes. Can you state what you need to prove, in epsilon-terms?


We need to prove that lAnbn - 0 l < E?
Original post by Dannygem
We need to prove that lAnbn - 0 l < E?

For some E, yes. Let's fix an ϵ\epsilon. What you have so far is:

N: n>N, an<ϵ\exists N: \forall \ n > N, \ |a_n| < \epsilon
 n>N,L<bn<U\forall \ n > N, L < |b_n| < U.

We want to show something about anbn=anbn|a_n b_n| = |a_n| |b_n|; in particular, we want to bound it above by something arbitrarily small. Can you give me an upper bound?
Reply 6
Original post by Smaug123
For some E, yes. Let's fix an ϵ\epsilon. What you have so far is:

N: n>N, an<ϵ\exists N: \forall \ n > N, \ |a_n| < \epsilon
 n>N,L<bn<U\forall \ n > N, L < |b_n| < U.

We want to show something about anbn=anbn|a_n b_n| = |a_n| |b_n|; in particular, we want to bound it above by something arbitrarily small. Can you give me an upper bound?


So we have l AnBn l = l An l l Bn l < UE (as An is less than E and Bn is < U)?

But can't Bn be <0? and so if Bn < U that does not imply that l Bn l < U?
Original post by Dannygem
So we have l AnBn l = l An l l Bn l < UE (as An is less than E and Bn is < U)?

But can't Bn be <0? and so if Bn < U that does not imply that l Bn l < U?

That bound is correct if you work only with the modulus of bnb_n and defined U>all bnU > \text{all} \ |b_n|.
bnb_n can be less than 0, but bn|b_n| is always less than max(U,L)\text{max}(|U|, |L|). A more useful formulation of "b_n is bounded" is usually "bn<B|b_n| < B for some BB".
Reply 8
Original post by Smaug123
That bound is correct if you work only with the modulus of bnb_n and defined U>all bnU > \text{all} \ |b_n|.
bnb_n can be less than 0, but bn|b_n| is always less than max(U,L)\text{max}(|U|, |L|). A more useful formulation of "b_n is bounded" is usually "bn<B|b_n| < B for some BB".


Ah yes now I understand that!

So what i'd do is state what I know (as we discussed at the start, what a null sequence is and what a bound is, and what we're trying to prove)

Then say as l An l < E and l Bn l < B, then we can say l AnBn - 0 l = l An l l Bn l < EB

which implies l AnBn l < EB and since B is a constant we can re-write this as l AnBn - 0 l < E and so we conclude AnBn is a null sequence?
Original post by Smaug123
A more useful formulation of "b_n is bounded" is usually "bn<B|b_n| < B for some BB".
Without wanting to be too pedantic, this should be really be something like:

A more useful formulation of "b_n is bounded" is usually "we can find B such that bn<Bn|b_n| < B \, \forall n.

(i.e. make it clear that you can find a single B that works for ALL b_n).
Original post by Dannygem
Ah yes now I understand that!

So what i'd do is state what I know (as we discussed at the start, what a null sequence is and what a bound is, and what we're trying to prove)

Then say as l An l < E and l Bn l < B, then we can say l AnBn - 0 l = l An l l Bn l < EB

which implies l AnBn l < EB and since B is a constant we can re-write this as l AnBn - 0 l < E and so we conclude AnBn is a null sequence?

That's exactly the thought process which goes into the answer, yes :smile: for proper rigour, your "re-writing" step needs to happen near the beginning of the eventual proof (though of course you come up with it right at the end of the process):

Let BB be such that for all nn, bn<B|b_n| < B.
Fix arbitrary ϵ>0\epsilon > 0. Then we have, since (an)(a_n) is a null sequence, that there exists NN such that for all n>Nn>N, an<ϵB|a_n| < \frac{\epsilon}{B}. Let MM be such an NN.
Then for all n>Mn > M, have anbn=anbn<ϵB×B=ϵ|a_n b_n| = |a_n| |b_n| < \frac{\epsilon}{B} \times B = \epsilon; this means that anbn0a_n b_n \to 0 as nn \to \infty. That is, (anbn)(a_n b_n) is a null sequence.


Original post by DFranklin
Without wanting to be too pedantic, this should be really be something like:

A more useful formulation of "b_n is bounded" is usually "we can find B such that bn<Bn|b_n| < B \, \forall n.

Yes, sorry - that was sloppy of me. (Curse natural language, in which it's easy to mean something entirely different from what you said.)
(edited 9 years ago)
Reply 11
Thank you so much! Very appreciated!

:smile:

Quick Reply

Latest