1

Let $g(x) :\mathbb{R}\to\mathbb{R}$ be a non-decreasing function in the sense that $g(y)\geq g(x) \; \forall y \geq x$. Can $g(x)$ be nowhere continuous? My guess is that any such function that is non-decreasing must have at most countably infinite number of discontinuities. Otherwise, it will cease to be discontinuous due to non-decreasing nature. However, I couldn't come up with a rigorous proof of this. Am I right in my guess?

My attempt was to assume that for each $x_0 \in \mathbb{R}$ there exists a fixed $\varepsilon_{x_0}>0$ such that for all $x \in (x_0,\infty)$ we have the relation that $f(x)>f(x_0)+\varepsilon_{x_0}$ by non-decreasing nature because for each arbitrary $\delta$ neighborhood of $x_0$ there is at least one $x'$ in it that satisfies the relation (hence all $x>x'$[non-decreasing] works but $x'$ can be made arbitrarily close to $x_0$). My intuition is that such a strictly increasing relation can only be forced for countable number of points. Is there a nice way to show it? Or I am on the wrong track?

Nvm: This proof strategy is wrong. Take the Floor function.

Someone
  • 2,865
  • 1
    'My guess is that any such function that is non-decreasing must have at most countably infinite number of discontinuities.' I seem to remember that the sum of an uncountable number of strictly positive real numbers is always unbounded, which would disprove the existence of such a function. I.e. such a function would have 'jumped' up an uncountable amount times in the interval (0,1). – Alexander Geldhof Apr 25 '20 at 10:28
  • Ah, my bad. I didn't know this was a standard question. It came up to me at some other area. I tried to search for the whole Real function but the results didn't match the one linked as duplicate.

    And yes, that works! Uncountable sums of strictly positive but small real numbers will become unbounded very easily and the relation won't make sense anymore. Thank you.

    – Someone Apr 25 '20 at 10:36

0 Answers0