0

A function $h:\mathbb{Q} \rightarrow \mathbb{R}$, with $h(x) = 0$ for $|x|<\sqrt2$ and $h(x) = 1$ for $|x| > \sqrt2$ is continuous for all $x$ in $\mathbb{Q}$.

It states in the solution that for $x_0$ in $\mathbb{Q}$ and $\delta =$ min{|$x_0$ + $\sqrt2$|,|$x_0$ - $\sqrt2|$} it follows that $h(x) - h(x_0) = 0$ and therefore < $\epsilon$. Thus continuous.

I don't understand how to arrive at that solution. Can someone help me out? I am new to analysis. This might be simple but I just don't get it.

2 Answers2

0

(Remark: if you haven't sketched the function $h$ -- on paper, or in your head if your visual imagination is up to it -- then you absolutely should do that first. Don't try to do this sort of thing just by pushing symbols around.)

Obviously the function $h_0\,:\,\mathbb{R}\rightarrow\mathbb{R}$ with "the same definition" as $h$ isn't continuous. So what's different when you're mapping from $\mathbb{Q}$? Only the fact that the points $\pm\sqrt{2}$ aren't in $\mathbb{Q}$, and those are the points in $\mathbb{R}$ where $h_0$ fails to be continuous.

So, suppose you have some point $x_0$ that _isn't $\pm\sqrt2$. You need to show that $h$ is continuous there, which means you need to find an interval around $x_0$ in which $h$ is well-behaved.

Usually "well-behaved" would be defined in terms of how much you're allowing the value of $h$ to change -- "for all $\epsilon>0$, there is some $\delta>0$", etc. -- but in this particular case $h$ is constant on any interval that doesn't include $\pm\sqrt2$.

And so that's the idea. Given $x_0$, we know it doesn't equal $\pm\sqrt2$ because those values aren't rational. So we can find an interval around $x_0$ small enough to stay away from those points. How small does that interval need to be? Its radius needs to be smaller than the distance from $x_0$ to either $+\sqrt2$ or $-\sqrt2$. That's where the definition of $\delta$ comes from. And once we have an interval around $x_0$ that avoids those points, we know that $h$ is constant there.

  • Oh I see, so there is not really an algebraic way to get to this but rather just use, that between every rational and irrational we can find another rational and therefore the interval from x_0 to this number is always smaller then the interval of x_0 to sqrt(2). And thus we stay out of the „dangerzone“ of sqrt(2) and the function therefor stays constant. Ok! – Rick Dänzer Feb 25 '24 at 17:50
0

As $\sqrt{2}$ is not a rational number, we are not asked to analise continuity around it. So let's pick $x_0$ as proposed. Without loss of generality let's just consider the case $x_0 > \sqrt{2}$. Let $\delta = x_0 - \sqrt{2}$, then for any $\epsilon>0$ and any $x_1$ such that $|x_1-x_0|<d$ with $d<|x_0 - \sqrt{2}|$. Then we have that $x_1>\sqrt{2}$ and thus $h(x_1)=1$ so $|h(x_1)-h(x_0)|=|1-1|=0 < \epsilon$.

Mefitico
  • 1,845
  • Thank you. Why do we know there exists such an x_1? And why is (x_1 - x_0) negative? Only then x_1 > sqrt(2). Otherwise i get that x_1 < 2*x_0 - sqrt(2). Where we could conclude nothing, right? So how do I know wether x_1 > x_0 or not? – Rick Dänzer Feb 25 '24 at 17:17
  • @RickDänzer the "without loss of generality" means you basically would have to write this argument down 3 times for each case: $x_0 > \sqrt{2}$ is done, $x_0 < -\sqrt{2}$ and $-\sqrt{2} < x_0 < \sqrt{2}$ are the other cases. As I am assuming the case $x_0 > \sqrt{2}$. etween any two distinct real numbers there always exists a rational and an irrational number. There is an elementary proof for that but is not so easy to come up with it by oneself. This is what ensure the proposed $x_1$ exists. – Mefitico Feb 26 '24 at 13:06