2
$\begingroup$

I am self-studying the book Understanding Analysis by S. Abbott and I tried to solve the exercise 2.6.6 (c) without relying on Bolzano-Weierstrass Theorem as they did here. Such exercise requires the following definition.

Lets call a sequence quasi-incrasing if for all $\varepsilon>0$ there exists an $N$ such that whenever $n>m\geq N$ it follows that $a_n>a_m-\varepsilon.$

Now, I want to prove that if $(a_n)$ is a quasi-increasing and bounded sequence, then $(a_n)$ must converge. My attempt of proof goes as follows.

Suppose, in order to derive a contradiction, that $(a_n)$ does not converge. Then by the Cauchy Criterion, $(a_n)$ is not a Cauchy sequence. Hence, there must be an $\varepsilon_0$ such that for all $N \in \mathbb{N}$ there exists $n,m \geq N$ such that $|a_m-a_n|\geq \varepsilon_0.$ (*)

On the other hand, as $(a_n)$ is a quasi-increasing sequence there is an $N \in \mathbb{N}$ such that for all $n>m\geq N$, it follows that $a_n>a_m-\frac{\varepsilon_0}{2}.$ (**)

Also, the boundedness of $(a_n)$ guarantees the existence of an $M>0$ such that $|a_n| \leq M$ for all $n \in \mathbb{N}.$

Now, by (*) we can choose $n,m \geq N$ such that $|a_{n}-a_{m}| \geq \varepsilon_0.$ WLOG, assume $n>m$ (observe that it cannot happen that $n=m$). Then we have the inequality $a_n>a_m-\frac{\varepsilon_0}{2}.$

Now, since $|a_m-a_n|\geq \varepsilon_0>\frac{\varepsilon_0}{2}$ and $\frac{\varepsilon_0}{2}>a_m-a_n,$ then it must be satisfied the inequality $$a_n>a_m+\frac{\varepsilon_0}{2}.$$

Choose $n_1=m$ and $n_2=n.$

Suppose that we have choosen $n_1<\cdots<n_k$ such that for all $j=1,\ldots,k-1$ it follows the inequality $$a_{n_{j+1}}>a_{n_j}+\frac{\varepsilon_0}{2}.$$

Now choose $N'=n_k+1.$ Observe that $N'>n_k>n_1> N.$ Again, by (*) we can choose $n',m'\geq N'$ such that $|a_{m'}-a_{n'}| \geq \varepsilon_0.$ WLOG assume $n'>m'.$

We have now the two following conclussions:

(1) Since we have $n'>m'\geq N'>N$, then by (**) we must have $a_{n'}>a_{m'}-\frac{\varepsilon_0}{2}>a_{m'}-\varepsilon_0,$ and hence $\varepsilon_0>a_{m'}-a_{n'}.$ This, together with the fact that $|a_{m'}-a_{n'}| \geq \varepsilon_0$ allows us to conclude that $$a_{n'}\geq a_{m'}+\varepsilon_0.$$

(2) Since $m'\geq N'>n_k>N,$ then by (**) we have that $a_{m'}>a_{n_k}-\frac{\varepsilon_0}{2}.$

Then, by (1) and (2) we should obtain $$a_{n'} \geq a_{m'}+\varepsilon_0>\left(a_{n_k}-\frac{\varepsilon_0}{2}\right)+\varepsilon_0)=a_{n_k}+\frac{\varepsilon_0}{2}.$$

Choose now $n_{k+1}=n'>N'>n_k.$

We have then constructed inductively a subsequence $(a_{n_k})$ of $(a_n)$ that satisfies $$a_{n_{k+1}}>a_{n_k}+\frac{\varepsilon_0}{2} \quad \text{(***)}$$ for all $k \in \mathbb{N}$.

Lastly, choose a number $K \in \mathbb{N}$ such that $K\frac{\varepsilon_0}{2}\geq M-a_{n_1}.$ So, by the condition (***) applied several times we obtain $$a_{n_{K+1}}>a_{n_K}+\frac{\varepsilon_0}{2}> a_{n_{K-1}}+2\frac{\varepsilon_0}{2}> \cdots > a_{n_2}+(K-1)\frac{\varepsilon_0}{2} >a_{n_1}+K\frac{\varepsilon_0}{2}\geq M.$$

This contradicts the fact that $|a_{n_{K+1}}|\leq M.$

Then we have the desired conclussion.

I hope you can check this proof and let me now if there is a mistake. Thanks in advance.

$\endgroup$

0

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.