Consider the Laplacian as an operator $\Delta\colon W^{2,p}(\Omega)\subset L^p(\Omega)\to L^p(\Omega)$ subject to homogeneous Robin boundary conditions, where $\Omega\subset \mathbf R^n$ is either bounded with smooth boundary or a halfspace and $1< p< \infty$. Is there any reference giving information about the invertibility of $\Delta$ in these spaces?


1 Answer 1


A careful exposition can be found in "Analytic Semigroups and Semilinear Initial Boundary Value Problems" by Kazuaki Taira.

He considers the mixed boundary value problem $$\begin{cases} A u=f & \mbox{in }\Omega,\\\ Bu:=a\left.\frac{\partial u}{\partial \nu}+bu\right|_{\partial\Omega}=\phi & \mbox{on }\partial\Omega,\end{cases}\qquad\qquad\qquad(*)$$ where $A$ is a second-order uniformly elliptic differential operator with $C^{\infty}$ coefficients and $\Omega\subset\mathbb R^n$ is a bounded domain with smooth boundary. The functions $a$, $b$ are assumed to be smooth and to satisfy some natural non-degeneracy conditions.

He shows that, for $1< p < \infty$ and $s > 1+1/p$, the mapping $$(A,B):H^{s,p}(\Omega)\to H^{s-2,p}(\Omega)\oplus B^{s-1-1/p,p}(\partial\Omega)$$ is an algebraic and topological isomorphism (Theorem 1, p. 4). Here $H^{s,p}(\Omega)$ is the standard Sobolev space and $B^{s-1/p,p}(\partial\Omega)$ is the Besov space of the traces (or boundary values) of functions $u\in H^{s,p}(\Omega)$.

This implies, for any $f\in H^{s-2,p}(\Omega)$ and $\phi\in B^{s-1-1/p,p}(\partial\Omega)$, the existence and uniqueness of a solution $u\in H^{s,p}(\Omega)$ to problem ($*$).

Edit. In your case $s=2$ and $\phi\equiv 0$ so one does not even have to worry about the fractional order Sobolev spaces and the Besov space of traces on the boundary.

  • $\begingroup$ Thank you very much for your reference. I just got the book, and noticed that the author is only dealing with bounded domains. Are there any references which deal with the unbounded-domain case, especially halfspace? $\endgroup$
    – Marc
    Dec 6, 2010 at 8:36
  • $\begingroup$ Unfortunately, I don't know such a reference. However, I believe that the proof from Taira's book (i.e. the reduction of the BVP to a problem for the corresponding pseudo-differential operator on the boundary and proving regularity estimates for the operator) should work mutatis mutandis in the case of halfspace. $\endgroup$ Dec 6, 2010 at 9:07

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.