Proving the infimum of A:=(m,n) is m

550 Views Asked by At

I'm just practicing doing proofs that involve supremums and infimums to better understand them for my elementary analysis class. Could someone please correct any errors in my proof. I found this proof here. The proof in question is example 2. I'm just wondering if I could prove the second part of the proof using contradiction.

$A=(m,n)$ implies $\{x\in R: m\lt x\lt n\}$ In addition, we can see that there is no minimum or maximum since this is an open interval.

Now, in order to prove that $\inf(A)=m$, I need to 1) Show m is a lower bound: this is done by using the definition of what the set A is. By definition: $m\lt x:\forall x\in A$

Thus, m is a lower bound of A.

2) By contradiction: assume $m_0$ is the largest lower bound of A which implies that $m_0\gt m$ and $m_0\lt x:\forall x\in$ A. Then, $m\lt m_0 \lt x \lt n$ $\Rightarrow$ $m_0 \in A$. Since $m_0\in A \Rightarrow \frac{m+m_0}{2}\in A \Rightarrow \exists$ x such that $x\lt m_0$. Which is a direct contradiction to $m_0\lt x$. Thus, m is the largest lower bound.

Since these two conditions were satisfied, $\inf(A)=m$

As always, thank you for your time.

1

There are 1 best solutions below

0
On BEST ANSWER

It's easier to just do definitions.

$A=(m,n)=\{x\in \mathbb R | m <x <n\} $.

1) for all $x \in A $, $x <m $ by definition. So $m$ is lower bound.

2) If $r>m $ then $m < \min (n,r) $ and by the archimedian principle there is an $y$ so that $m<y <\min(n,r)$. (If you want you can set $y=\min (\frac {m+r}2,\frac {m+n}2) $... but you don't have to.

Then $m<y <\min (n,r)\le r $ so $r$ is not an lower bound. So $m $ is greatest lower bound.

If you want to do a prove by contradiction, you can do 2 differently.

2') Suppose $m $ is not greatest lower bound. Then there is an lower bound $m_0; m <m_0$ (it doesn't have to be the greatest lower bound).

Let $y $ be so $m < y <m_0$. The for any $x\in A $ we have: $m_0 \le x; m <x < n ; m < y <m_0$. So $m <y <n $. So $y \in A $. This contradicts $m_0$ is lower bound of $A $.

So $m $ is greastest lower bound.

A third way and fourth use the theorem: if $x \le y \le x+\epsilon $ for all $\epsilon > 0$ then $y=x $. (Have you proven that? If you haven't you should. It's a good exercise.)

2'') $m $ is a lower bound. So if $m_0$ is greatest lower bound then $m \le m_0$. Let $(n-m) > \epsilon >0$. Then $m <m+\epsilon < n $ so $m+\epsilon \in A $. So as $m_0 $ is lower bound $m_0 \le m+\epsilon$. So $m\le m_0\le m+\epsilon $.

So $m=m_0$

2''') let $m_0$ be greatest lower bound. $m_0 <n $ as $m_0 <x <n $ for all $x\in A $. Let $\epsilon >0$. $m_0 - \epsilon < \inf A=m_0$ so $m_0-\epsilon \not \in A $ and $m_0-\epsilon <m_0 <n$. So $m_0-\epsilon \le m $.

And as $m $ is lower bound $m\le m_0$. So $m_0-\epsilon\le m\le m_0$.

So $m=m_0$.