2
$\begingroup$

I am considering a smooth vector field $A \colon \mathbb{R}^N \to \mathbb{R}^N$ and I define the magnetic Sobolev space $H_A^1$ as the closure of $C_0^\infty(\mathbb{R}^N)$ with respect to the norm $$ u \mapsto \int_{\mathbb{R}^N} \left| \left( -i\nabla -A \right)u \right|^2 + |u|^2. $$ My question is: suppose that I need to replace $\mathbb{R}^N$ by $\mathbb{R}_{+}^{N+1}=[0,+\infty) \times \mathbb{R}^N$. Does there exist a characterization of the trace space associated to $H_A^1$, if of course I identify $\partial \mathbb{R}_{+}^{N+1}$ with $\mathbb{R}^N$?

The answer should depend somehow on $A$, which I assume to be locally bounded but not necessarily globally bounded.

$\endgroup$
4
  • 1
    $\begingroup$ The space depends heavily on the behavior of $A$ at $\infty$. With no assumption on $A$ you cannot expect a precise answer concerning $H_A^1$. The curvature of $A$ may also introduce more subtle constraints. $\endgroup$ Jan 30, 2016 at 12:55
  • $\begingroup$ @LiviuNicolaescu Of course my $A$ should be divergent at infinity, since a constant $B=\operatorname{curl} A$ must be allowed. Do you have any reference? $\endgroup$
    – Siminore
    Jan 30, 2016 at 14:52
  • $\begingroup$ The answer depends on the behavior of $A$ at $\infty$. $\endgroup$ Jan 30, 2016 at 17:59
  • $\begingroup$ For instance $A(x_1,x_2,x_3)=(-x_2,x_1,0)$. $\endgroup$
    – Siminore
    Jan 30, 2016 at 19:15

1 Answer 1

2
$\begingroup$

If the magnetic field $\operatorname{curl} A$ is bounded, then the trace space is $$ \left\{ g \in L^2 (\mathbb{R}^{N - 1}, \mathbb{C}) ; \iint\limits_{\mathbb{R}^{N - 1} \times \mathbb{R}^{N - 1}} \frac{\lvert e^{-i \int_{0}^1 A ((1 -t)x + t y)) \,\mathrm{d} t \cdot (y - x)} g (y) - g (x)\rvert^2}{\lvert y - x \rvert^N} \,\mathrm{d}y \, \mathrm{d} y < +\infty \right\}; $$ with estimates depending only on $\lVert \operatorname{curl} A \rVert_{L^\infty}$. See our paper Nguyen, Van Schaftingen, Characterization of the traces on the boundary of functions in magnetic Sobolev spaces, arXiv:1905.01188.

$\endgroup$

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.