Abstract
In complex algebraic geometry, the problem of enumerating plane elliptic curves of given degree with fixed complex structure has been solved by R. Pandharipande using Gromov-Witten theory. In this article, we treat the tropical analogue of this problem: the determination of the number of tropical elliptic plane curves of degree $d$ and fixed "tropical $j$-invariant" interpolating an appropriate number of points in general position. We show that this number is independent of the position of the points and the value of the $j$-invariant and that it coincides with the number of complex elliptic curves. The result can be used to simplify Mikhalkin's algorithm to count curves via lattice paths in the case of rational plane curves.
Introduction: The Tropical Enumerative Problem
In enumerative geometry, a fundamental goal is to determine how many geometric objects of a specific type satisfy a set of constraints. For elliptic curves (curves of genus 1), a key constraint is the $j$-invariant, which characterizes the curve's complex structure. While R. Pandharipande solved this for classical complex geometry, tropical geometry offers a new perspective by "skeletonizing" these complex curves into piecewise-linear combinatorial graphs.
This paper addresses whether the tropical version of this counting problem yields the same result as the classical one. Specifically, the authors investigate if the tropical count $E_{trop}(d)$ is independent of the choice of points and the value of the $j$-invariant, and if it matches the complex count $E(d, j)$.
Methodology: Tropical Moduli and Limiting Cases
The authors construct the moduli space of tropical elliptic curves as a weighted polyhedral complex. This framework allows them to use linear algebra—specifically the determinants of evaluation maps—to define the "multiplicity" of each curve. A tropical elliptic curve is defined as a graph with exactly one cycle, and the tropical $j$-invariant is defined as the integer length of this cycle.
To prove their correspondence theorem, the authors analyze two extreme scenarios for the $j$-invariant:
- Very Large $j$: When the cycle is very long, the curve behaves like a rational (genus 0) curve with a "contracted" edge. This relates the elliptic count to the known count of rational curves ($N_{trop}(d)$).
- Very Small $j$: When the cycle shrinks toward zero, the count relates to specific combinatorial features of rational curves, leading to a new formula for lattice path counting.
Additionally, the authors introduce a factor of $1/2$ in certain weights to account for automorphisms (symmetries) in the tropical curves, suggesting the moduli space behaves like a "tropical orbifold."
Key Insights: The Magic Ratio and Universal Invariance
A primary finding of the paper is the confirmation of the "Magic Ratio": $E_{trop}(d) = \binom{d-1}{2} N_{trop}(d)$. This mirrors the classical result where the number of elliptic curves is a simple function of the number of rational curves of the same degree.
Furthermore, the authors discover a property of "Universal Invariance." In complex geometry, the $j$-invariant has exceptional values ($j=0$ and $j=1728$) where counts differ due to extra symmetries. In the tropical world, however, the count remains invariant for all $j \ge 0$, with no such exceptional values.
Results and Algorithmic Improvements
The paper establishes a formal Correspondence Theorem, proving that for $j \notin \{0, 1728\}$, the tropical and complex counts coincide exactly. They demonstrate that the number of tropical elliptic curves is independent of the specific configuration of the $3d-1$ points, provided they are in general position.
Beyond theoretical matching, the work provides a significant practical contribution: a modified version of Mikhalkin’s lattice path algorithm. By analyzing the "very small $j$" limit, the authors derived a computationally "cheaper" algorithm that ignores paths with only steps of size 1, significantly reducing the complexity of calculating $N(d)$.
Discussion and Future Directions
While the paper proves that the counts match, it does not yet provide a direct, one-to-one bijection between specific tropical and complex curves. Future work might explore why the tropical world lacks the exceptional $j$ values found in the complex world. Additionally, while many concepts here apply to other toric surfaces, some specific results—like the optimized lattice path count—rely on the specific geometry of the projective plane and may not generalize universally. The formalization of "tropical orbifolds" also remains an open area for more general theoretical development.
Summary updated