I am reading Ciarlets book on elliptic boundary value problems and the finite element method. He defines the Sobolev spaces $H^m_0(\Omega)$ as the completion of the space $\mathcal{D}(\Omega)$ of compactly supported infinitely differentiable real valued functions with respect to the Sobolev norm $$ \| u\|_{m, \Omega} = \left( \int_\Omega \sum_{|\alpha| \leq m} |D^\alpha u|^2 \right)^{1/2} $$

However, under certain requirements on the boundary $\Gamma$ of the domain $\Omega$, the spaces for $ m = 1 $ and $m = 2$ can be characterised as

$$ H^1_0(\Omega) = \left\{ u \in H^1(\Omega) \mid u = 0 \text{ on } \Gamma\right\} \\ H^2_0(\Omega) = \left\{ u \in H^2(\Omega) \mid u = \partial_{n} u = 0 \text{ on } \Gamma\right\} $$ where $n$ is the outward pointing unit normal to $\Gamma$. What is the fundamental difference between these characterisations, and the definition involving the completion of the space? Under what circumstances can I use the characterisation at the bottom?

Thanks for all help in advance! :)


They are the same, you can find the proof in Adam's book, he considered these two spaces and he denoted them by $H^{m,p}$ and $W^{m,p}$, then he proved the equivalence between these two definitions.


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.