Modification of Crazy Function on Set of Measure Zero Still Yields Crazy Function

lebesgue-integralmeasure-theoryreal-analysis

I'm working on the following problem:

Let $f(x) = \frac{1}{x^{1/2}}$ if $x \in 0 <x < 1$ and $0$ otherwise. Let ${r_n}$ be an enumeration of the rationals and set $g(x) = \sum_1^\infty2^{-n}f(x-r_n)$. Show that $g$ is discontinuous at every point and unbounded on every interval. Further this remains so after any modification on a Lebesgue null set.

I'm stuck on the bolded part. I showed unboundedness everywhere and discontinuity everywhere. I "sense" the statement should be correct. I'm not sure what to say about it though.

Best Answer

Hint: For any interval, $(a,b)$, and any integer $M$, show that $\{x:g(x)>M\}\cap (a,b)$ has nonzero measure. Removing a null set does not change the measure of this set; in particular it will still be nonempty.

It will help to find rational number $r_n$ which lies in $(a,b)$ and consider the "spike" $2^{-n}f(x-r_n)$. As long as the spike is larger than $M$, than $g$ will be as well. You should be able to find a nontrivial interval where the spike is large.