While working on regularity theory for variational formulation of PDEs I often find ''Let $\Omega \subset \mathbb R^n$ and $u$ be sufficiently regular so that Green's identity $$ \int_\Omega (-\Delta u)v = \int_\Omega \nabla u \nabla v - \int_{\partial \Omega}v \partial_n u$$ holds''. My question is: How regular does $\Omega$ has to be so that the above statement holds for $u \in H^2 (\Omega)$?

Thanks so much in advanced,



$\partial\Omega$ and $u$ have to be smooth enough in order for the integral $\int_{\partial\Omega}v\partial_n u$ to make sense.

What suffices is: Lipschitz continuity for the boundary and $\partial_n u\in L^2(\partial\Omega)$, for $u$. Even with weaker assumptions it is OK, for example $\partial_n u\in H^{-1/2}(\partial\Omega)$.

  • $\begingroup$ The reference I had in mind (PDEs by Avner Friedman) is not reader friendly book. Let me check and see if I can find anything friendlier. $\endgroup$ – Yiorgos S. Smyrlis Jan 5 '14 at 16:43
  • $\begingroup$ I looked but could not find references to regularity. I found the theory of adjoint equations, which amounts to the same thing, but it was all about smooth functions. $\endgroup$ – D G Jan 5 '14 at 17:01
  • $\begingroup$ See Theoretical Numerical Analysis: A Functional Analysis Framework By Kendall Atkinson, Weimin Han - Chapter 8. $\endgroup$ – Yiorgos S. Smyrlis Jan 5 '14 at 18:24
  • $\begingroup$ Found it. Thank you so much for your effort. $\endgroup$ – D G Jan 5 '14 at 19:33

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.