1
$\begingroup$

I am currently trying to prove the multiplicative limit law:

let $(a_n)^{\infty}_{n=m}, (b_n)^{\infty}_{n=m}$ be convergent sequences of real numbers, and $X, Y$ be the real numbers $X = \lim_{n\to \infty}a_n$ and $Y = \lim_{n\to \infty}b_n$. $$ \lim_{n \to \infty}a_nb_n = \left(\lim_{n\to \infty}a_n\right) \cdot \left(\lim_{n\to \infty}b_n\right) $$

Since both $(a_n)^{\infty}_{n=m}$ and $(b_n)^{\infty}_{n=m}$ are convergent to X and Y respectively, We know that $|a_n - X| \leq \epsilon'$ and $|b_n - Y| \leq \delta$.

We also know, by some lemma we proved earlier in the book, that $|a - b| \leq \epsilon \land |c - d| \leq \delta \implies |ac - bd| \leq \epsilon \cdot |c| + \delta \cdot |a| + \epsilon \delta$.

This is perfect, as I can use it to show that $|a_nb_n - XY| \leq \epsilon$ for some arbitary $\epsilon > 0$, as long as I show that there exists $\epsilon' * |Y| \leq \frac{\epsilon}{3}$ and that there exists some $0 < \delta < 1$ such that $\delta \cdot (|X| + \epsilon') \leq \frac{2}{3}\epsilon$

I could prove the first part using the Archimedean property of the reals, but I am not so sure about the second part. The second part feels like it should work since we can choose an arbitrarily small $\delta$, but I can't prove that it does. Am I doing something wrong? is it possible to change this proof a bit to make it work?

$\endgroup$
4
  • 1
    $\begingroup$ What you've written is nonsense, the sequences $a_n$ and $b_n$ might not converge even if the product does. Take $\{1,0,1,0,\ldots\}$ and $\{0,1,0,1,0,\ldots\}$. $\endgroup$ Commented Aug 15, 2020 at 12:57
  • $\begingroup$ @TSF Right, I forgot to declare that they converge to a real number, but that is assumed as part of the exercise. I'll edit that in. $\endgroup$
    – javajav
    Commented Aug 15, 2020 at 12:59
  • 1
    $\begingroup$ Looks like a proof to me. Just state $$\delta = \min \left(1, \frac{2 \epsilon}{3(|X|+ \epsilon')}\right)$$ $\endgroup$
    – aschepler
    Commented Aug 15, 2020 at 13:12
  • $\begingroup$ hmm, so I guess I overthought it, as usual ^^'. Thank you very much! If you'll post it as an answer I will mark it as the solution. $\endgroup$
    – javajav
    Commented Aug 15, 2020 at 13:19

1 Answer 1

1
$\begingroup$

If $a_n \to a, b_n \to b$ then there is some $M$ such that $|a|,|b_n| \le M$.

Then $|a_nb_n -ab| = |a_nb_n -a b_n + a b_n -ab| \le |a_n-a| |b_n| + |a| |b_n-b| \le M (|a-a_n|+ |b-b_n|)$.

Now choose $N$ big enough so that $|a-a_n|, |b-b_n| < {\epsilon \over 2 M}$.

$\endgroup$
4
  • $\begingroup$ how did you get to $|a_nb_n -a b_n + a b_n -ab| \leq |a_n - a| |b| + |a| |b_n - b|$? specifically the $|a_n - a| |b|$ bit, because that can't be inferred from the triangle inequality, from what I know. $\endgroup$
    – javajav
    Commented Aug 15, 2020 at 14:53
  • $\begingroup$ It was a minor omission, I should have had $b_n$ instead of $b$. Rather than grinding through detail, it might be better to look at the general idea behind the proof so that minor typos. don't block progress. $\endgroup$
    – copper.hat
    Commented Aug 15, 2020 at 15:35
  • $\begingroup$ I am sorry if that's what it looked like. It's just that I wanted to see if my understanding of what you wrote was complete, or that I maybe missed some crucial detail (in your proof, or even worse, in the book/material). Thank you for taking your time to write this! $\endgroup$
    – javajav
    Commented Aug 15, 2020 at 17:40
  • $\begingroup$ Good luck, hope it helps! $\endgroup$
    – copper.hat
    Commented Aug 15, 2020 at 19:56

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .