1
$\begingroup$

Consider a family of maps $f_p\colon \ell^p(\mathbb Z,\mathbb R)\to \ell^p(\mathbb Z,\mathbb R)$, $p\ge 2$, where $$ f_q = f_p\big|_{\ell^q},\qquad \forall\; 2\le q\le p. $$ Moreover,

  • $f_p\colon \ell^p\to \ell^p$ is real analytic for any $p\ge 2$,
  • $f_p\colon \ell^p\to \ell^p$ is one-to-one for any $p\ge 2$,
  • $f_p\colon \ell^p\to \ell^p$ is a local diffeomorphism everywhere. More to the point, $d_xf_p\colon \ell^p\to \ell^p$ is a compact perturbation of the identity at any point $x\in\ell^p$.

Therefore, $f_p$, $p\ge2$, is a Fredholm map, that is its differential is a Fredholm operator at each point. However, we do not know whether $f_p$ is proper, i.e. whether preimages of compact sets are compact.

Furthermore, for $p=2$ we have the additional properties

  • $f_2$ is onto and hence is a diffeomorphism $f_2\colon \ell^2\to \ell^2$,
  • $\|f_2(x)\|_2 = \|x\|_2$.

Is $f_p$ onto also for $p > 2$?

Since $f_p\big|_{\ell^2} = f_2$ the range of $f_p$ contains $\ell^2$ and hence is dense in $\ell^p$ for any $p\ge 2$. The range is also open since $f_p$ is a local diffeomorphism. One may thus equivalently ask:

Is the range of $f_p$ closed for $p > 2$?

Any suggestions how to attack these kind of problems?

$\endgroup$
2
  • $\begingroup$ This is not known even in finite dimensions. $\endgroup$ Commented Sep 25, 2015 at 13:45
  • $\begingroup$ But in finite dimension you cannot have the property that when restricted to $\ell^2$ it is a diffeo? $\endgroup$ Commented Sep 25, 2015 at 18:02

0

You must log in to answer this question.

Browse other questions tagged .