Let $f:[a,b] \to \mathbb{R}$ be bounded.

Is there always a Riemann integrable function $g:[a,b] \to \mathbb{R}$ such that for any partition $P$, we have $U(g,P) = U(f,P)$?

Of course, if $f$ is Riemann integrable, we can simply take $g \equiv f$, but that's not generally true.

The motivation to ask this question was the function $f(x) = \mathbf{1}_{\mathbb{Q} \cap [0,1]}(x)$ defined on $[0,1]$. For any partition $P$, $f$ shares its upper sum with $f \equiv 1$ on $[0,1]$, which happens to be integrable.

| cite | improve this question | | | | |

[This is not a complete answer.]

I guess that a good candidate is the upper semicontinuous envelope $$ g(x) := \lim_{r\to 0+} \left( \sup_{y\in B_r(x)} f(y) \right). $$ Indeed, $g\geq f$ and, for every interval $[c,d]\subset[a,b]$ one has $$ \sup_{[c,d]} g = \sup_{[c,d]} f, $$ so that $U(g,P) = U(f,P)$ for every partition $P$.

The function $g$ is upper semicontinuous; on the other hand, an upper semicontinuous function can have a set of discontinuities of positive measure, hence need not be Riemann integrable. Such a function could be used to construct a counterexample.

In conclusion, I think that the answer to your question is negative.

| cite | improve this answer | | | | |

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.