Let $\Omega\subset \mathbb R^N$ be open bounded with smooth boundary. Let $u\in SBV\cap L^\infty(\Omega)$ be given. We write $$ Du = \nabla u\lfloor \mathcal L^N + (u^+-u^-)\otimes \nu_u\mathcal H^{N-1}\lfloor S_u $$ where $S_u$ is the jump set of $u$.

Then define set $\mathcal F$ such that $$ \mathcal F:=\{v\in SBV(\Omega)\cap L^\infty(\Omega),\|\nabla v\|_{L^\infty}\leq C, T[v]=T[u],S_v\subset S_u\} $$ where $C>0$ is a constant.

Then, do I have that $v\in\mathcal F$ is $L^\infty$ uniformly bounded?

If not, how may I modify $\mathcal F$ so that my requirement is true? Maybe require more on jump set? Since I feel my conjecture may not be true since I give no requirement on how much it can jump for a function $v\in\mathcal F$.

Thank you!


Let $\Omega=[-1,1]$ and $u(x)=u(-x)=0$ for $x\in[2^{-2k+1},2^{-2k+2})$ and $u(x)=u(-x)=2^{-k}$ for $x\in[2^{-2k},2^{-2k+1})$, $k=1,\ldots,n,\ldots$.

Doesn't this $u$ belong to your space $SBV\cap L^\infty$ ? Because if no condition is imposed on the jumps, the set $\mathcal F$ clearly contains piecewise constant functions with the same jump set $S_u=\{\pm 2^{-n}:\ n=1,\ldots\}$ and arbitrarily large values.

  • $\begingroup$ I understand your point, and this is actually this is what I am worry about. I have no control of how "high" the function can jump. Do you think how may I impose such condition to control it? $\endgroup$ – JumpJump Aug 11 '15 at 11:16
  • $\begingroup$ I would try including (in the definition of $\mathcal F$) a condition that the singular (jump) part of $Dv$ has its (pointwise) norm dominated by that of the jump part of $Du$, or the weaker condition that $\int_{S_u} |v^+-v^-|\ d\mathcal{H}^{N-1}$ is bounded. $\endgroup$ – Jean Duchon Aug 11 '15 at 11:40
  • $\begingroup$ Does this weaker condition you mentioned has some l.s.c properties? $\endgroup$ – JumpJump Aug 11 '15 at 11:59
  • $\begingroup$ I think it does imply $L^\infty$-boundedness of $\mathcal F$. This needs a proof in general, but in 1 dimension it's obvious. The lower semicontinuity is another question. $\endgroup$ – Jean Duchon Aug 11 '15 at 12:23

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.