1

Let $(a_{n,m})_{n\in\mathbb N,\,m\in \mathbb N}$ be a double sequence taking real values. We may consider different types of limit:

  1. $\lim_{n\to\infty,\,m\to\infty} a_{n,m} = l_1\in\mathbb R\,$ if for every $\epsilon>0$ there exists $N_\epsilon\in\mathbb N$ such that $\min(n,m)>N_\epsilon\Rightarrow|a_{n,m}-l_1|<\epsilon\,$;
  2. $\lim_{||(n,m)||\to\infty} a_{n,m} = l_2\in\mathbb R\,$ if for every $\epsilon>0$ there exists $N_\epsilon\in\mathbb N$ such that $\max(n,m)>N_\epsilon\Rightarrow|a_{n,m}-l_2|<\epsilon\,$;
  3. $\lim_{n\to\infty}(\lim_{m\to\infty} a_{n,m}) = l_3\in\mathbb R$ if for every $\epsilon>0$
    • for every $n\in\mathbb N$ there exist $L_n\in\mathbb R$ and $M_{\epsilon,n}\in\mathbb N$ such that $m>M_{\epsilon,n}\Rightarrow|a_{n,m}-L_n|<\epsilon\,$
    • there exist $N_\epsilon\in\mathbb N$ such that $n>N_\epsilon\Rightarrow|L_n-l_3|<\epsilon$ ;
  4. $\lim_{m\to\infty}(\lim_{n\to\infty} a_{n,m}) = l_4\in\mathbb R$ if for every $\epsilon>0$
    • for every $m\geq M_\epsilon$ there exist $\Lambda_m\in\mathbb R$ and $N_{\epsilon,m}\in\mathbb N$ such that $n>N_{\epsilon,m}\Rightarrow|a_{n,m}-\Lambda_m|<\epsilon\,$
    • there exist $M_\epsilon\in\mathbb N$ such that $m>M_\epsilon\Rightarrow|\Lambda_m-l_4|<\epsilon$ ;

In What is the definition of double sequence $a_{mn}$ being convergent to $l$? it is shown that convergence 1. does not ensure that the sequence $(a_{n,m})_{n,m\in\mathbb N}$ is bounded.

Convergence 2. should be enough to guarantee that $(a_{n,m})_{n,m\in\mathbb N}$ is bounded.

My question is : convergence 3. together with convergence 4. (with $l_3\neq l_4$ in general) suffice to guarantee $(a_{n,m})_{n,m\in\mathbb N}$ bounded?

Notice that in the definitions of double limits 3. and 4. I have assumed convergence of the first limit for every value of the second index, in order to avoid problems similar to those appearing here What is the definition of double sequence $a_{mn}$ being convergent to $l$?

tituf
  • 837

1 Answers1

2

Let

$$a_{n,m}=\begin{cases} n,&\text{if }n=m\\ 0,&\text{otherwise;} \end{cases}$$

then each row limit is $0$ and each column limit is $0$, but the double sequence is unbounded.

$$\begin{array}{cccccc|c} \color{red}0&0&0&0&0&\ldots&0\\ 0&\color{red}1&0&0&0&\ldots&0\\ 0&0&\color{red}2&0&0&\ldots&0\\ 0&0&0&\color{red}3&0&\ldots&0\\ 0&0&0&0&\color{red}4&\ldots&0\\ \vdots&\vdots&\vdots&\vdots&\vdots&\color{red}\ddots&\vdots\\\hline 0&0&0&0&0&\ldots&0 \end{array}$$

Brian M. Scott
  • 631,399
  • Thank you for your answer. If I have a double sequence such that limits $l_3$ and $l_4$ exist and are finite but $l_3\neq l_4$, what can be a strategy to prove that the sequence is bounded? (knowing that limit 2. cannot exists) – tituf Mar 26 '21 at 23:38
  • 1
    @tituf: You’re welcome. You’d need to know more about the double sequence. Note that we could change all of the zeroes below the diagonal to $1$ and get $\ell_3=0\ne 1=\ell_4$ without really changing the nature of the problem presented by examples like this one. At the moment I don’t see any condition that would ensure boundedness that isn’t just some restatement of boundedness. – Brian M. Scott Mar 26 '21 at 23:52