1

Let $a,b$ be nonzero and different real numbers. Let $f: [0,1] \to R $ with

$$ f(x) = \left\{ \begin{array}{lr} ax & : x \in \mathbb{Q}\\ bx & : x \notin \mathbb{Q} \end{array} \right. $$

Let $D$ be the set of discontinuities of $f$. I want to find out what $D$ is but I am having difficulties. Do I have to consider cases on $a$ and $b$ ?

hossayni
  • 401
ILoveMath
  • 10,999

4 Answers4

5

Break into two cases, the case where $a=b$ and the case where $a\neq b$. When $a\neq b$ visualize the function as two lines (with holes in them) intersecting at the origin. Remember that every irrational point is a limit point of rationals and every irrational point is a limit point of rationals.

Seth
  • 9,563
2

For all $x_0\in [0,1]$, and if $f$ is continous at $x_0$, then it needs two conditions, 1)$\lim_{x\to x_o}f(x)$ must exists 2)$\lim_{x\to x_o}f(x)=f(x_0)$.

First we check whether $\lim_{x\to x_o}f(x)$ exists, this is equivalent to $$\lim_{x\to x_o, x\in \mathbb{Q}}f(x)=\lim_{x\to x_o, x\in \mathbb{R}\setminus\mathbb{Q}}f(x)$$ Hence we must have $ax_0=bx_0\iff (a-b)x_0=0$.

If $a\neq b$, this can only be satisfied by $x_0=0$. Hence $[0,1]\setminus\{0\}\subset D$. Then we check whether it's continuous at $0$. Indeed it is, since $\lim_{x\to x_o, x\in \mathbb{Q}}f(x)=\lim_{x\to x_o, x\in \mathbb{R}\setminus\mathbb{Q}}f(x)=0=f(0)$. So the function is continuous at $0$ and $D=[0,1]\setminus\{0\}$.

If $a=b$, then $f(x)=ax \quad x\in [0,1]$. So $f$ is continous at $[0,1]$, $D=\emptyset$.

John
  • 13,738
2

Peace be upon you

The answer is $(0,1]$, because the two lines diverges by passing from 0. The discontinuity can be written mathematically as you write the discontinuity of Dirichlet function.

hossayni
  • 401
1

It is a modification of the Dirichlet function see : http://en.wikipedia.org/wiki/Nowhere_continuous_function if $a\neq b$ is discontinuous at every point, except at the origin, where it is continuous.

  • 1
    Not at every point... – Dustan Levenstein Oct 10 '14 at 22:42
  • 2
    Continuous at the origin – Seth Oct 10 '14 at 22:50
  • I do not think so, the value at the origin is zero, regardless of a and b, but it is not continuous, or please show me, it will be enlightening! – Wolphram jonny Oct 10 '14 at 22:53
  • 1
    @julianfernandez take any sequence converging to $0$, call it $x_n$. What does $f(x_n)$ converge to? – DanZimm Oct 29 '14 at 00:45
  • @DanZimm you are correct, thanks for the enlightenment! I still have trouble making sense of it. Continuity at only-one-point sounds like an oxymoron to my mind. Could it be that the classical definition of continuity fails to capture the concept of continuity for this pathological case? Or is it just me? I think I am going to post my own question about this. – Wolphram jonny Oct 29 '14 at 14:03
  • @julianfernandez if you post a question I'll gladly expound on this - otherwise my quick answer is it depends on what you consider the "classical definition," I personally think about it as "small changes in domain cause relatively small changes in the range," which seems misleading at first but after some thought makes sense. For this sort of view point continuity at just one point makes sense. – DanZimm Oct 29 '14 at 15:19
  • @DanZimm I posted my question here: http://math.stackexchange.com/questions/999320/why-a-function-continuous-at-only-one-point-is-not-an-oxymoron – Wolphram jonny Oct 31 '14 at 02:47