2
$\begingroup$

The question of bounding the number of integer points on an elliptic curve $E/\mathbb{Q}$, shown to always be finite by Siegel, is an old question. There are various aspects to this problem. The aspect I am focusing on is the so-called repulsion principle of the integer points.

We first give some preliminaries. Suppose we have an elliptic curve $E/K$, $K$ a number field. Then the canonical height on $E$ admits a decomposition into local heights, given by

$$\displaystyle \hat{h}(P) = \frac{1}{[K : \mathbb{Q}]} \sum_{v \in M_K} d_v \lambda_v(P)$$

where $M_K$ is the set of places of $K$. A repulsion principle is then a statement of the form

$$\displaystyle \lambda_v(P_1 - P_2) \gg \min\{\lambda_v(P_1), \lambda_v(P_2)\},$$

where the exact statement depends on the place $v$, and of course depends on whether $v$ is an archimedean or non-archimedean place.

In this famous paper of Helfgott and Venkatesh, they proved that if $v$ is a place of potential good reduction for $E$, then we have

$$\displaystyle \lambda_v(P_1 - P_2) \geq \min\{\lambda_v(P_1), \lambda_v(P_2)\}.$$

If $v$ is a place of potential multiplicative (bad) reduction or if $v$ is archimedean, then one would need to partition the set of points into finitely many pieces, and some inequality of this sort holds on each piece; see Lemmas 3.1 - 3.3 in the aforementioned paper.

I am interested in a more basic version of this type of repulsion principle. Suppose we have an elliptic curve $E/\mathbb{Q}$, given by say a (minimal) short Weierstrass model:

$$\displaystyle E_{A,B} : y^2 = x^3 + Ax + B, \quad A,B \in \mathbb{Z},$$

and $p^4 | A \Rightarrow p^6 \nmid B$ for all primes $p$. Fixing this model, consider the set $E(\mathbb{Z})$ of integer points on this curve. More precisely, we have

$$E(\mathbb{Z}) = \{(x,y) \in \mathbb{Z}^2 : y^2 = x^3 + Ax + B\}.$$

Do we know of a statement of the form? That $E(\mathbb{Z})$ can be partitioned into $k = O(1)$ subsets $E_1 \cup \cdots \cup E_k$, such that whenever $(x_1, y_1), (x_2, y_2)$ are distinct elements of $E_j$, then

$$\displaystyle |x_1 - x_2| \geq f(A,B),$$

where $f(A,B)$ is some function that tends to infinity as $\max\{|A|,|B|\} \rightarrow \infty$?

$\endgroup$
2
  • 3
    $\begingroup$ The example $A=-1$, $B=T^2$ having the two points $(0,T)$ and $(1,T)$ shows that $k>1$ is needed. $\endgroup$ Commented Nov 26 at 9:00
  • 1
    $\begingroup$ Don't forget that Silverman noticed this sort of thing before us (and we credit them). See Gross-Silverman. $\endgroup$ Commented Dec 13 at 11:48

0

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.