4

Question:

Given the $$f(x,y)=y-\displaystyle\frac{1}{x^2}$$ consider the set $S = \{(x,y) \in D: x > 0,\ f(x,y) > 0\}$ where $D$ is the domain of $f$. Sketch the set $S$ in the plane. Determine whether $S$ is closed,whether it is open, whether it is bounded and whether it is compact. Explain carefully


Attempt at answer:

The domain will be $x>0$ and $y>\displaystyle\frac{1}{x^2}$

A function is closed if it contains all of its boundary points

I’m not sure if I am correct in saying but since it is not a strictly greater than and equal than it does not include all of the boundary points Also $(1,1)$ is a boundary point in $S$ but it is not an element of the set as $1=\displaystyle\frac{1}{1^2}$

A function is open if all points are interior points. Not to sure how to show this one?

A function is bounded if you can draw a ball of radius $k$ such that the entire set is contained This is not possible as the function is infinite so will always have points outside the ball

A function is compact if it is closed and bounded. Since it is neither closed nor bounded it is not compact.

erin
  • 125
  • For $x<0$ the function is defined. Why the restriction $x>0$ when you are talking about its domain? – drhab Aug 16 '14 at 08:56
  • the question just asked to consider x>0 or am I confusing it? – erin Aug 16 '14 at 08:59
  • $D$ was referred to as the domain of $f$ and I do not see restrictions on it. $S$ was defined on base of $D$ with extra conditions among which $x>0$. – drhab Aug 16 '14 at 09:08

2 Answers2

3

Let me go through your attempted answer:

The domain will be $x>0$ and $y>\displaystyle\frac{1}{x^2}$

If you meant to say that $S=\{(x,y); x>0, y>\frac1{x^2}\}$, then this is correct.

If you meant $D$ by the word domain, then this is incorrect: $D$ is the set of all points for which $f(x,y)$ is defined.

A function is closed if it contains all of its boundary points

You meant to say that a set (not a *function) is closed if and only if it contains all boundary points. (But this is probably just a typo.)

Also $(1,1)$ is a boundary point in $S$ but it is not an element of the set as $1=\displaystyle\frac{1}{1^2}$

Yes, this is true. But maybe you should also explain why $(1,1)$ is a boundary point.

A function is open if all points are interior points. Not to sure how to show this one?

If you have $(x,y)\in S$ then $$y>\frac1{x^2}$$ and $x>0$.

Using continuity of the function $g(x)=1/x^2$ (or maybe using $h(x)=\frac1{x^2}-x$ would be even simpler), can you show that there exists $\varepsilon>0$ such that $$ |s-x|<\varepsilon, |t-y|<\varepsilon \qquad \implies \qquad s>\frac1{t^2}?$$ This basically says that points close enough to $(x,y)\in S$ are still in $S$. Showing this is sufficient to get that each $(x,y)\in S$ is an interior point.

A function is bounded if you can draw a ball of radius $k$ such that the entire set is contained

You should write set instead of function.

This is not possible as the function is infinite so will always have points outside the ball.

I get your point, but perhaps you should explain it more in detail. For example the points $P_n=(n,n)$ for $n>1$ all belong to $S$. Is it clear that for any ball $B(0,r)$ around the origin one of this points will be outside the given ball?

A function is compact if it is closed and bounded. Since it is neither closed nor bounded it is not compact.

Again set instead of function. Otherwise ok.

  • $P_n \in S$ if $n > 1$ – Darth Geek Aug 16 '14 at 09:14
  • Thanks so much for all your help – erin Aug 16 '14 at 09:15
  • @drhab Thanks I have corrected it. – Martin Sleziak Aug 16 '14 at 09:16
  • 1
    @DarthGeek Thanks I have corrected this. – Martin Sleziak Aug 16 '14 at 09:16
  • I am not sure how to explain why (1,1) is a boundary point. – erin Aug 16 '14 at 09:20
  • also i showing that is open I understand your summary but I dont understand the first part of it and where you get the equations from eg the f(x) and g(x) – erin Aug 16 '14 at 09:22
  • @erin That depends on your definition of boundary point. If your definition is that $P$ is a boundary point of $S$ if and only if any neighborhood of $P$ contains points both from the set $S$ and its complement $S$, then you could simply try to show that you can find such points close to $(1,1)$. For example $(1+\varepsilon,1+\varepsilon)$ and $(1-\varepsilon,1-\varepsilon)$. Or you might say that existence of such points is clear from the picture. (It all depends on how rigorously do you want/are supposed to write the proof.) – Martin Sleziak Aug 16 '14 at 09:23
  • 1
    @erin I have changed notation to $g(x)$ and $h(x)$. (It was my mistake - I should not have used $f$, since another function is already denoted by $f$.) I have suggested $h(x)=\frac1{x^2}-x$ because $(x,y)\in S$ $\leftrightarrow$ $x>0$ and $y>h(x)$. – Martin Sleziak Aug 16 '14 at 09:25
  • yes that is the definition that we are using for boundary pt. this is probably a stupid question but why do the two values have to be the same eg (1+e,1+e) – erin Aug 16 '14 at 09:27
  • 1
    @erin They don't have to be the same. I have simply used the same values, because from the picture it is obvious that one of them is in $S$ and the other isn't. You can as well take $(1,1+\varepsilon)$ and $(1,1-\varepsilon)$. (This might be even better when doing the computation showing whether the point is in $S$.) I should have also written that $\varepsilon>0$, but this is probably clear from the context. – Martin Sleziak Aug 16 '14 at 09:33
  • Thanks that has cleared it up a lot :) – erin Aug 16 '14 at 09:39
1

What you wrote is the domain of $f$ restricted to $S$. The actual domain of $f$ is $\mathbb{R}^2 - \{(0,y): y\in \mathbb{R}\}$.

As you said, $(1,1)$ is a boundary point of $S$ that doesn't belong in $S$ so $S$ is not closed.

To show that all it's points are interior, consider this:

Let $(x_0,y_0)\in S$ then $x_0 > 0$ and $x_0^2y_0 > 1$. Let $d = \min\lbrace\sqrt{(x_0-x)^2+(y_0-y)^2}:x^2y=1\rbrace$ Consider $k = \min\{x_0,d\}$

Then $B\left((x_0,y_0),k\right)\subseteq S$. Therefore all points in $S$ are interior.

Your proof that $S$ is not bounded seems a bit dodgy. Perhaps this helps. $c:=(1,2)\in S$. So if $S$ is bounded there is a $k\in R$ so that $S \subseteq B(c,k)$ But $(2k+1,2)\in S$ while $(2k+1,2)\notin B(c,k)$. Therefore, $S$ is not bounded.

Indeed, if $S$ is not closed or not bounded it is not compact.

Darth Geek
  • 12,296