Abstract
One of the big questions in the area of curves over finite fields concerns the distribution of the numbers of points: Which numbers occur as the number of points on a curve of genus $g$? The same question can be asked of various subclasses of curves. In this article we classify the possibilities for the number of points on genus 4 hyperelliptic supersingular curves over finite fields of order $2^n$, $n$ odd.
Introduction
In the study of algebraic geometry over finite fields, a fundamental inquiry involves determining the possible number of rational points a curve can possess. For a curve of genus $g$ over a field with $q$ elements, the Hasse-Weil bound (refined by Serre) establishes a general range: $|N - (q+1)| \leq g \lfloor 2\sqrt{q} \rfloor$. However, this bound is often quite loose for specific subclasses of curves.
Supersingular curves represent a particularly interesting subclass due to the high degree of symmetry in their Jacobians. While the possible point counts for supersingular curves of genus 1, 2, and 3 have been previously established, genus 4 remained an open problem. This research fills that gap by providing a complete classification for hyperelliptic supersingular curves in characteristic 2, specifically over fields of order $q = 2^n$ where $n$ is odd.
Methodology
The research employs two distinct mathematical frameworks to derive the point-counting constraints:
1. Quadratic Forms and Character Sums
The primary approach involves representing the point-counting problem as a character sum. By analyzing the trace of the polynomial defining the curve, the authors transform the problem into the study of quadratic forms over a field of characteristic 2. Specifically, they utilize the properties of linearized polynomials to demonstrate that the rank of these forms must follow specific patterns. This rank is intrinsically linked to the dimension of the radical (the kernel) of a bilinear form, which directly limits the possible number of points.
2. Theory of Abelian Varieties
An alternative, more geometric proof is provided using the Jacobian of the curve. By examining the characteristic polynomial of the Frobenius endomorphism, the authors relate the curve's points to the decomposition of its Jacobian into simpler components—specifically, supersingular elliptic curves. This method relies on the hypothesis that a certain algebraic element acts as an endomorphism on the Jacobian, leading to the same restrictions on point counts as the quadratic form method.
Key Insights
The Normal Form
A crucial step in making the character sum calculations tractable was the identification of a "normal form" for these curves. The paper demonstrates that any supersingular hyperelliptic curve of genus 4 in characteristic 2 can be simplified to the following equation: $$y^2 + y = fx^9 + ax^5 + bx^3 + cx + d$$
The Odd Dimension Constraint
In the quadratic form analysis, the authors prove that the dimension of the radical of the associated bilinear form must be odd (specifically 1, 3, or 5). Because the point count $N$ is tied to $2^w$ (where $w$ is this dimension), the resulting values are restricted to specific powers of 2, significantly narrowing the range of possibilities compared to the general Hasse-Weil bound.
The "Missing" Value
Perhaps the most striking discovery is the exclusion of certain values. While general theory might suggest that $N - (q+1)$ could be any multiple of $\sqrt{2q}$ within the Hasse-Weil limit, this research proves that the value $\pm 3\sqrt{2q}$ is mathematically impossible for this specific class of curves.
Main Results
The central contribution of this work is Theorem 4, which states that for a genus 4 hyperelliptic supersingular curve over $\mathbb{F}_q$ (where $q=2^n$ and $n$ is odd), the number of rational points $N$ must satisfy:
$$N - (q + 1) \in \{0, \pm \sqrt{2q}, \pm 2\sqrt{2q}, \pm 4\sqrt{2q}\}$$
To validate these findings, the authors provide specific examples of curves for $n=11$, demonstrating that each of the predicted point counts is indeed achievable. This confirms that the derived set of possibilities is tight and complete.
Conclusion and Future Work
This classification provides a definitive answer for genus 4 hyperelliptic supersingular curves in characteristic 2 with $n$ odd. However, several avenues for future research remain:
- Characteristic and Parity: The behavior of these curves in characteristics other than 2, or for fields where the exponent $n$ is even, may differ significantly.
- Non-Hyperelliptic Curves: While many supersingular curves are hyperelliptic, not all are. Classifying the point counts for non-hyperelliptic genus 4 curves remains an open challenge.
- The Endomorphism Hypothesis: The second proof path relies on an assumption regarding the Frobenius endomorphism. While consistent with the first proof, a formal verification of this hypothesis for all such curves would strengthen the geometric understanding of these varieties.
Summary updated