0
$\begingroup$

I am supposed to determine if function:

$$g(t)=\left\{\begin{matrix} \frac{1}{p+q}, t \in \left ( 0,1 \right )\cap \mathbb{Q}, p,q \in \mathbb{N}, t=\frac{p}{q}, GCD\left ( p,q \right )=1\\ 0, t \in\left ( 0,1 \right )\setminus \mathbb{Q} \end{matrix}\right.$$

  1. Is continous in $t\in \left ( 0,1 \right )\setminus \mathbb{Q}$
  2. Is integrable on $\left ( 0,1 \right )$ and if so, determine the value of integral

My solution:

  1. $\left ( \forall \varepsilon> 0 \right )\left ( \exists \sigma > 0 \right ) $ such as interval $\left ( x_{0}-\sigma , x_{0}+\sigma \right ) $ does not contain any number in the form of $\frac{1}{p+q} $, because numbers in those form are finite. So, for $t\in \left ( x_{0}-\sigma , x_{0}+\sigma \right ) $ is $g(x)=0$ and therefore function is continous in $t \in\left ( 0,1 \right )\setminus \mathbb{Q}$.
  2. $P(n)=\frac{1}{n}$, $L\left (P,g \right )=0$

Can anyone help me, how to continue, or if my solution is so far correct?

$\endgroup$
5
  • $\begingroup$ I'm sorry, but your proof of (1) is completely false. The set of numbers of the form $\frac 1{p+q}$ is infinite. And they are the values of $g(t)$, not of $t$, but your argument claims they restrict $t$. $g(t) = 1/(p+q)$ for any rational value of $t$, and there are rational numbers in every open interval. Thus every neighborhood of a point in $(0,1) \setminus \Bbb Q$ will contain points where $g(t) \ne 0$. To prove this, you need to show that if the interval is small enough, the value of $p+q$ is large (just showing $q$ is large is sufficient). $\endgroup$
    – Paul Sinclair
    Nov 2, 2019 at 3:47
  • $\begingroup$ As for integrability, the lower sum for any partition is clearly $0$, so you will need to show that the upper sum goes to $0$ as the partition refines. $\endgroup$
    – Paul Sinclair
    Nov 2, 2019 at 3:59
  • $\begingroup$ @PaulSinclair can you elaborate, how I show that upper sum goes to 0? $\endgroup$
    – Peter F.
    Nov 2, 2019 at 8:25
  • 1
    $\begingroup$ if $\epsilon > 0$, then there is a $q$ such that $1/q < \epsilon$, and there are only finitely many rational numbers in $(0,1)$ whose reduced denominator is $< q$. Each of these can be contained in partition intervals as small as you want. $\endgroup$
    – Paul Sinclair
    Nov 2, 2019 at 15:55
  • $\begingroup$ @PaulSinclair thank you $\endgroup$
    – Peter F.
    Nov 2, 2019 at 19:52

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.