Let $f : \mathbb{R}\times(0, 1) \to \mathbb{R}$ be a measurable function such that

a) $f(t,\,\cdot\,)\in L^1(0,1)$ for a.e. $t$,

b) there exists $C > 0$ such that for every interval $(a,\, b) \subset (0,\, 1)$ and for every $s \leq t$ there holds $$ \bigg| \int^b_a f(t,x)- f(s,x)\,\, dx\bigg|\leq C(b-a)(t-s).\qquad (1)$$

Prove that, up to redefining $f$ on a subset of $\mathbb{R}×(0, 1)$ of zero measure, there holds $$|f(t, x) − f(s, x)| \leq C(t − s)\qquad\qquad (2)$$ for every $s \leq t$ and every $x \in (0, \, 1)$.

My solution.

Remark: here I use the notation $\mu$ to denote both the one-dimensional and the two-dimensional lebesgue measure.

From a) there exists a subset $B$ of $\mathbb{R}$ s.t. $f(t,\,\cdot\,)\in L^1(0,1)$ for avery $t\in B$ (with $\mu(\mathbb{R}\setminus B)=0$). In $(1)$ I divide by $b-a$ and I take the limit $b\to a$ obtaining, by lebesgue differentiation theorem, $$|f(t, a) − f(s, a)| \leq C(t − s)\quad \text{ for } a \in A\subset (0,\,1)$$ that is $(2)$ (where $\mu((0,1)\setminus A)=0$). I claim that the subset of $\mathbb{R}×(0, 1)$ of zero (bidimensional) measure is $B \times A$. In order to prove that $\mu((\mathbb{R}×(0, 1)) \setminus (B \times A))=0$ I define the partition $X_n=[n,n+1)$ of the real line. Since $$\mu((B \times A) \cap (X_n \times (0,1)))=\mu((B \cap X_n ) \times A))=1\cdot 1=1$$ we get $\mu((X_n \times (0,1))\setminus (B \times A) )=0$. From additivity of the measure we get $$ \mu((\mathbb{R}×(0, 1)) \setminus (B \times A))=\mu\bigg(\bigcup_n\,\big((X_n \times (0,1))\setminus (B \times A)\big)\bigg)=0.$$

Questions and doubts:

  • It is al correct ?

  • I am also interested in different kind of proof.


The claim is wrong. Let $a=-b<0$ and $f(t,x):=e^t x$. Then all integrals $\int_a^b f(t,x)=0$, but there is no estimate of the form $|f(t,x)-f(s,x)|\leq C|t-s|$ valid for all $x\in[a,b]$.

  • $\begingroup$ Thanks for your answer. Then the exercise is wrong but where is the error in my proof? $\endgroup$ – Ef_Ci Jul 1 '18 at 15:07
  • $\begingroup$ Ad if we suppose also that $ \big| \int^b_a f(t,x)- f(s,x)\,\, dx\big| \neq 0$ ? $\endgroup$ – Ef_Ci Jul 1 '18 at 15:59

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.