2

$\forall \epsilon \in (0,1) \ \exists N \in \mathbb{N^+} :\ \forall n\ge N:\ |x_n-a|\le 2\varepsilon \Leftrightarrow \forall\varepsilon_1>0,\exists N \in \mathbb{N^+},n\ge N,|x_n-a|<\varepsilon_1$

how to manipulate $N-\varepsilon $ language to prove above statement?

Noa Even
  • 2,801
  • 1
    For $\varepsilon\geq 1$ just choose the same $N$ as for $\varepsilon = 1/2$. – Jakobian Aug 10 '19 at 16:57
  • Basically, you need an arbitrarely small $\varepsilon_1$. By taking $\varepsilon_1=2\cdot \varepsilon$, arbitrarely small $\varepsilon$ leads to arbitrarely small $\varepsilon_1$ and vice-versa. – rtybase Aug 10 '19 at 16:57
  • 1
    Off topic and perhaps grouchy, but what is it with these abbreviated titles? How to prove this if and only what? A title should give us an idea of what the question will be and if you cut off the question mid-sentence, then the title "How to prove $A \iff$" makes as much sense as "Why is Brie Larson the". Why is Brie Larson the what? – fleablood Aug 10 '19 at 17:42

3 Answers3

0

A general understanding of the statements we're dealing with is a good thing:

The first statement is telling: given any real number $\epsilon$ between zero and one, we can find (that's the core of the statement) a natural number $N$ so that the distance $|x_{n\geq N} - a|$ between the sequence's elements $x_{n\geq N}$ starting at $n=N$ and the limit $a$ oft the sequence $x_n$ is smaller than the double of the given real number $\epsilon$.

The second statement is telling: given any real number $\epsilon_1$ greater than zero, we can find a natural number $N$ so that the distance $|x_{n\geq N} - a|$ between the sequence's elements $x_{n\geq N}$ starting at $n=N$ and the limit $a$ of the sequence $x_n$ is smaller than the given real number $\epsilon_1$.


What needs to be done in general

If you want to prove the equivalence, then:

(a) assume that the first statement is true and prove the second statement

(b) assume the second statement is true and prove the first statement.


The (a) direction or the $\implies$ direction

(a) assume for any given $0<\epsilon<1$ we can find a natural number $N_0$ so that $|x_{n\geq N_0} - a|<2\epsilon$, that is assume the first statement is true. To prove the second statement, we need to find a number $N_1$ for any given $\epsilon_1 >0$, so that $|x_{n\geq N_1} - a|<\epsilon_1$.

First we prove the following implication

$\forall\epsilon \in (0,1),\exists N_0 \in \mathrm{N^+},|x_{n\ge N_0}-a|\le 2\epsilon \implies \forall\epsilon_1>0,\exists N_1 \in \mathrm{N^+},|x_{n\ge N_1}-a|<\epsilon_1$

What is the premise? What is the assumption?

Assume we can find a number $N_0$ so that the condition or the inequality $|x_{n\geq N_0} - a|<2\epsilon$ is true for any given $0<\epsilon<1$.

What is the conclusion or entailment?

The task is to find or show that we can find a number $N_1$ so that the inequality or the condition $|x_{n\geq N_1} - a|<\epsilon_1$ is true for any given $\epsilon_1 > 0$.

How do we mathematicize or how do we use the assumption to get to the conclusion?

Now if we're given an $0<\epsilon_1^x<2$, then it follows directly from the assumption, that $|x_{n\geq N_1^x} - a|<\epsilon_1^x$ and we can find an $N_1^x$ (in this particular case of $0<\epsilon_1^x<2$ we can just choose or set $N_1^x=N_0$) so that the condition $|x_{n\geq N_1^x} - a|<\epsilon_1^x$ for any given $0<\epsilon_1^x<2$ is satisfied.

In the case we're given any $\epsilon_1^y \geq 2$ and asked to find an $N_1^y$ so that $|x_{n\geq N_1^y} - a|<\epsilon_1^y$ is true, we note that $|x_{n\geq N_1^y} - a|<\epsilon_1^x<\epsilon_1^y$. This shows that we can find such an $N_1^y$ that makes $|x_{n\geq N_1^y} - a|<\epsilon_1^y$ a true condition.

With being able to find a number $N_1^x$ so that $|x_{n\geq N_1^x} - a|<\epsilon_1^x$ is true and a number $N_1^y$ so that $|x_{n\geq N_1^y} - a|<\epsilon_1^y$ is also true for any given $0<\epsilon_1^x<2$ and for any given $\epsilon_1^y \geq 2$, we have proved now that we can find an $N_1$ so that $|x_{n\geq N_1} - a|<\epsilon_1$ is true for any given $\epsilon_1>0$.


The (b) direction or the $\impliedby$ direction

(b) assume the second statement is true and imply the first statement. Prove the following implication:

$\forall\epsilon \in (0,1),\exists N_0 \in \mathrm{N^+},|x_{n\ge N_0}-a|\le 2\epsilon \impliedby \forall\epsilon_1>0,\exists N_1 \in \mathrm{N^+},|x_{n\ge N_1}-a|<\epsilon_1$

If $\forall\epsilon_1>0,\exists N_1 \in \mathrm{N^+},|x_{n\ge N_1}-a|<\epsilon_1$, then clearly $\forall\epsilon \in (0,1),\exists N_0 \in \mathrm{N^+},|x_{n\ge N_0}-a|\le 2\epsilon$, because $2\epsilon > 0$.


After $\implies$ and $\impliedby$, then you have shown the equivalence $\iff$, because showing that $(A \implies B)$ and showing $ (A \impliedby B)$ for two statements $A$ and $B$ is equivalent to saying that $A \iff B$.

Ahmed Hossam
  • 1,151
  • Your logic is not correct here. – Ted Shifrin Aug 10 '19 at 17:40
  • You're right. Thank you. I'll correct it. – Ahmed Hossam Aug 10 '19 at 17:50
  • I don't know why you are writing $\epsilon_1^x$ and similar instead of $\epsilon_1$. I don't see any reason to compare $\epsilon_1^x$ and $\epsilon_1^y$. Where did you learn this sort of notation? – Axion004 Aug 11 '19 at 04:41
  • Because we need to go from any given $0<2\epsilon<2$ to any given $\epsilon_1>0$ there are two possibilities for a given $\epsilon_1$. Either a given $\epsilon_1$ is between $0$ and $2$, which we will label $\epsilon_1^x$ ($0 < \epsilon_1^x < 2$) or $2 \leq \epsilon_1^y < \infty$. In this context $x$ and $y$ are just labels, that tell us from which intervall $\epsilon_1$ was picked – Ahmed Hossam Aug 11 '19 at 10:55
0

Given the first statement, we wish to prove the second. We start with $\epsilon_1>0$ and wish to find $N$ so that $|a_n-a|<\epsilon_1$ for all $n\ge N$. Set $\epsilon = \epsilon_1/2$. The first statement provides us an $N$ so that $$|a_n-a|<2\epsilon = 2\cdot\frac{\epsilon_1}2 = \epsilon_1$$ for all $n\ge N$. So that $N$ will work. [You can call the $N$ from the first statement $N_0$, if that makes it clearer to you.]

Can you do the other direction?

Ted Shifrin
  • 115,160
  • should I work left $\implies$ right ,and right $\implies$ left? – nevermind_15 Aug 10 '19 at 17:54
  • I showed you the proof of (first statement) $\implies$ (second statement). Make sure you understand it carefully. Then try to do the reverse direction. Most times you have to prove an $\iff$ statement, you should prove the two separate implications, yes. – Ted Shifrin Aug 10 '19 at 17:57
0

Can I do like this?:

$$\begin{aligned} left\Rightarrow right:& \forall\varepsilon \in (0,1),\exists N \in \mathrm{N^+},n\ge N,|x_n-a|\le 2\varepsilon \\& \frac{\varepsilon_{1}}{2}=\varepsilon,\Rightarrow \forall \varepsilon= \frac{\varepsilon_{1}}{2} \in (0,\frac{1}{2} ] \Leftrightarrow \forall \varepsilon_1 \in (0,1] ,\exists N_2 \in \mathrm{N^+},n\ge N_2,|x_n-a|\le \varepsilon_1 \\& \varepsilon=\frac{2}{\pi}\arctan(\varepsilon_2) \in (\frac{1}{2},1) \\&\Leftrightarrow \forall\varepsilon_2 \in(1,+\infty),\exists N_3 \in \mathrm{N^+},n\ge N_3,|x_n-a|\le 2\varepsilon= \frac{4}{\pi}\arctan(\varepsilon_2)\le \varepsilon_2 \\&\varepsilon_3=\left\{\begin{aligned} & \varepsilon_1=2\varepsilon & \varepsilon \in (0,\frac{1}{2 } ], \\& \varepsilon_2=\tan(\frac{\pi}{2}\varepsilon) & \varepsilon \in (\frac{1}{2} ,1). \end{aligned}\right., N_4=\max\{N_3,N_2\} \\&\Rightarrow \forall \varepsilon_3 >0 , \exists N_4 \in \mathrm{N^+},n\ge N_4 ,|x_n-a|\le \varepsilon_3 \Leftrightarrow \lim_{n\to\infty}{x_n}=a \\right \Rightarrow left:&\forall \varepsilon >0 , \exists N \in \mathrm{N^+},n\ge N ,|x_n-a|\le \varepsilon, 2\varepsilon_1=\varepsilon >0, \varepsilon_1=\frac{\varepsilon}{2}>0 \\& \forall \varepsilon_1>0 ,\exists N_1 \in \mathrm{N^+},n\ge N_1, |x_n-a|\le 2\varepsilon_1. \\& \varepsilon_2 \in (0,1) \subset (0,+ \infty) \ni \varepsilon_1 \Rightarrow \forall \varepsilon_2 \in(0,1) ,\exists N_2 \in \mathrm{N^+},n\ge N_2 ,|x_n-a|\le 2\varepsilon_2. \\&\square \end{aligned}$$