2
$\begingroup$

$\newcommand{\cris}{\mathrm{cris}}$In my setting, $K/\mathbb Q_p$ is finite and unramified, and $V$ is a $2$-dimensional crystalline representation of $G_K$. Then we have $D_{\cris}(V)$, which is $2$-dim filtered $\phi$ module. I need to calculate a basis for $D_{\cris}(V)$, the matrix of $\phi$ and the filtration. (p-adic approximate would also be of help).

I wonder if there exists a effective way to calculate $D_{\cris}$ for general Galois representations? Since the only few examples (of calculating $D_{\cris}$) I know are characters and Tate module which are deeply dependent on the structure of the representations and the methods seems not generalizable.

To respond Sawin: The Galois representation I care about is quite general, it comes from a series of lifting $\mathbb Z/p^n$ representations. It's true that every Galois rep contains infinite data, but the data can be recovered by looking at mod $p^n$ representations step by step. So I wonder if there is a canonical way to use the reduced representations $\rho_n:G_K\to GL_n(\mathbb Z/p^n)$ to calculate the filtered $\phi$-module.

$\endgroup$
5
  • 4
    $\begingroup$ How is the representation $V$ described in your data? A Galois representation is a priori an infinite amount of data - what do you actually have that you want to calculate with? $\endgroup$
    – Will Sawin
    Commented Nov 14, 2023 at 13:45
  • 1
    $\begingroup$ I agree with the substance of Will’s comment - the question isn’t answerable until you tell us in what way the V in your problem is characterised. That said, a single element of Qp is in itself an “infinite amount of data”! $\endgroup$ Commented Nov 14, 2023 at 16:48
  • $\begingroup$ @WillSawin You are right, but you could see my last para in the new edition. $\endgroup$
    – Richard
    Commented Nov 16, 2023 at 2:41
  • $\begingroup$ But in what format is the data of the $\rho_n$ handed to you? Like, what is the input of the calculation you anticipate wanting/needing? $\endgroup$
    – tkr
    Commented Nov 16, 2023 at 3:53
  • $\begingroup$ Unless you know more about your representation this question is unanswerable. It’s trivial to check that if T is a Zp-lattice in a crystalline representation, then you can always find T’ which is non-crystalline and congruent to T modulo an arbitrarily high power of p. $\endgroup$ Commented Nov 16, 2023 at 6:34

0

You must log in to answer this question.