2
$\begingroup$

Let $X$, $Y$ be reflexive Banach spaces, and let $\imath:X\hookrightarrow Y$ be a bounded inclusion with dense image. Then for any domain $\Omega\subset\Bbb R^n$ we may define $C_{\mathrm c}^\infty(\Omega;X)$ to be the (Fréchet)-smooth functions $\Omega\to X$ with compact support, with the usual inductive limit topology, and likewise for $C_{\mathrm c}^\infty(\Omega;Y)$. Then $\imath$ clearly induces a continuous inclusion $$\imath_*:C_{\mathrm c}^\infty(\Omega;X)\hookrightarrow C_{\mathrm c}^\infty(\Omega;Y)$$ however, it is not at all clear to me if $\imath_*$ will have dense image. I suspect it won't, though I have yet to find a counterexample. So, does it have dense image, and if so, why?

$\endgroup$

1 Answer 1

4
$\begingroup$

$C^\infty_c(\Omega, X) = C^\infty_c(\Omega,\mathbb R){\bar\otimes} X$ completed inductive tensor product which agrees with the projective tensor product since $C^\infty_c(\Omega,\mathbb R)$ is nuclear. Do the same for $Y$. Finite rank tensors are dense in $C^\infty_c(\Omega,\mathbb R){\bar\otimes} Y$ and these can be approximated by finite rank tensors in $C^\infty_c(\Omega,\mathbb R){\bar\otimes} X$.

$\endgroup$
2
  • 2
    $\begingroup$ One must be careful with the inductive tensor product. Even for nuclear spaces it does not necessarily coincide with the projective tensor product. This however is not needed for the argument. $\endgroup$ Commented Jul 17, 2017 at 7:18
  • $\begingroup$ Incidentally, could a similar technique be used to show that $$C_{\mathrm c}^\infty(\Omega)'\bar{\otimes}X=\mathcal{L}(C_{\mathrm c}^\infty(\Omega),X)$$ for any Banach space $X$, assuming such a thing is true? $\endgroup$ Commented Jul 18, 2017 at 21:45

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .