Hello all,

could someone point me to a reference that ties the smoothness of the solution $u$ to the classical elliptic problem

$\nabla \cdot ( q \nabla u ) = f \;,\; x \in \Omega$

$u = g \;,\; x \in \Gamma = \partial \Omega$

to the smoothness of $f$, $q$ and $g$?

$\Omega$ is a convex polygonal domain in $\Re^d$ with $d \in \{2,3\}$. The boundary $\Gamma$ is piecewise linear (can have corners, e.g., if $\Omega$ is the unit square).

I am particularly interested in the (minimal) smoothness requirements for the forcing and boundary data $f$ and $g$, such that $u \in {\cal H}^2(\Omega)$ (not just locally).

I went through Evans' book on PDEs but he assumes homogeneous boundaries and proves only local smoothness $u \in {\cal H}_{\rm loc}^s(\Omega)$ based on assumptions on the forcing $f$. My $g$ is generally nonzero.

Also, would the smoothness theory for the BVP above extend to a Helmholtz problem with a pure Neumann BC?

Thanks for any good pointers!

Kind regards, -- Mihai

  • $\begingroup$ Is $\Gamma$ the boundary of $\Omega$? If so, what kind of regularity are you assuming for $\Gamma$? $\endgroup$ Mar 1, 2011 at 0:06
  • 3
    $\begingroup$ have you looked at Gilbarg and Trudinger? $\endgroup$ Mar 1, 2011 at 0:11
  • $\begingroup$ @Yakov Shlapentokh-Rothman: I updated the problem definition. @Willie Wong: thanks for the pointer, I will check it out today. $\endgroup$
    – Mihai
    Mar 1, 2011 at 17:26
  • 1
    $\begingroup$ For an example of things that can go wrong since $\Gamma$ is not smooth, see mathoverflow.net/questions/38054/… $\endgroup$ Mar 2, 2011 at 0:30
  • $\begingroup$ Gilbarg and Trudinger assume smooth boundaries as far as I see. I am going through Grisvard right now to see what I can figure out from there. $\endgroup$
    – Mihai
    Mar 2, 2011 at 17:33

2 Answers 2


Grisvard's book is a standard reference for elliptic problems in domains with corners.


See http://en.wikipedia.org/wiki/Elliptic_regularity and references therein.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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