A classification algorithm for reflexive simplices

Basic Information

• Paper ID: 2510.09131
• Title: A classification algorithm for reflexive simplices
• Author: Marco Ghirlanda
• Classification: math.CO (Combinatorics), math.AG (Algebraic Geometry)
• Publication Date: October 10, 2025
• Paper Link: https://arxiv.org/abs/2510.09131

Abstract

This paper presents a universal classification algorithm for reflexive simplices that determines all reflexive simplices in five and six dimensions. From the perspective of algebraic geometry, this amounts to a classification of five and six-dimensional Gorenstein pseudo-weighted projective spaces. As a byproduct of the methodology, we obtain explicit formulas for the Picard group and Gorenstein index of arbitrary pseudo-weighted projective spaces.

Research Background and Motivation

Problem Background

Reflexive polytopes are an important mathematical concept, referring to lattice polytopes that contain the origin in their interior and whose duals are also lattice polytopes. This concept first appeared in Batyrev's work, which determined all unimodularly equivalent classes of reflexive polygons. Kreuzer and Skarke extended this classification to three and four dimensions.

Research Motivation

1. Importance of Classification: Complete classification of reflexive polytopes in five or higher dimensions appears computationally infeasible, but reflexive simplices as a special case are more tractable
2. Existing Results: Current classifications include 5 reflexive triangles, 48 reflexive 3-simplices, and 1561 reflexive 4-simplices
3. Geometric Applications: Reflexive simplices are directly related to Gorenstein pseudo-weighted projective spaces in toric geometry

Limitations of Existing Methods

• Complete classification of high-dimensional reflexive polytopes is computationally intractable
• Lack of effective algorithms for handling high-dimensional reflexive simplices
• Absence of explicit formulas for the Picard group and Gorenstein index of Gorenstein pseudo-weighted projective spaces

Core Contributions

1. Proposes Universal Classification Algorithm 4.17: Systematically classifies reflexive simplices in arbitrary dimensions
2. Completes High-Dimensional Classification: Determines 220,794 reflexive simplices in five dimensions and 309,019,970 in six dimensions
3. Establishes Geometric Correspondence: Transforms the classification of reflexive simplices into classification of Gorenstein pseudo-weighted projective spaces
4. Derives Explicit Formulas: Provides explicit expressions for the Picard group and Gorenstein index of pseudo-weighted projective spaces (Theorem 3.3)
5. Efficient Implementation: Five-dimensional classification completes in under 1 minute on a mid-range computer; six-dimensional requires approximately 20 days

Methodology Details

Task Definition

Given dimension n, find all unimodularly equivalent n-dimensional reflexive simplices, equivalent to classifying all n-dimensional Gorenstein pseudo-weighted projective spaces.

Core Technical Framework

1. Degree Matrix Representation

Uses degree matrices as Gale dual encoding of lattice simplices, representing simplices as matrices whose columns belong to a finite abelian group Z × Γ. This representation has two main advantages:

• Unimodular equivalence of lattice simplices corresponds to automorphisms of Z × Γ
• The reflexivity condition of simplices transforms into explicit conditions on degree matrices

2. Automorphism Group Generators (Theorem 2.1)

For finitely generated abelian group G = Z^k ⊕ Z/μ₁Z ⊕ ... ⊕ Z/μᵣZ, its automorphism group Aut(G) is generated by:

• ψᵢ: coordinate sign changes
• ψᵢ,ᵤ: unit multiplication on torsion part
• αᵢ,ⱼ, βᵢ,ⱼ, γᵢ,ⱼ, δᵢ,ⱼ: various additive transformations

3. Reflexivity Condition Decomposition (Proposition 4.1)

Decomposes the reflexivity condition into:

• A condition involving only free rows: L | Σᵢwᵢ
• Independent conditions for each torsion row: Mⱼ | (Σᵢwᵢ)/L
• Linear relations for the final coordinate: ηₙⱼ = -(η₀ⱼ + ... + ηₙ₋₁,ⱼ)

Algorithm Architecture

Core Steps of Algorithm 4.17:

1. Weight Vector Computation: According to Remark 4.6, compute all Gorenstein weight vectors via unit fraction decomposition
2. Torsion Vector Enumeration: For each weight vector and each μ ∈ Z≥2, compute all minimal Gorenstein torsion vectors using Procedure 4.10
3. Degree Matrix Combination: Appropriately combine weight vectors with torsion vectors using Procedure 4.13
4. Isomorphism Class Representatives: Use Procedure 4.16 to select unique representatives for each isomorphism class

Technical Innovations

1. Early Pruning Strategy: Utilize Lemma 4.8 to rapidly exclude impossible (a,b) pairs, significantly improving algorithm performance
2. Minimality Discrimination: Lemma 4.9 provides effective conditions for determining whether torsion vectors are minimal
3. Normal Form: The normal form introduced in Definition 4.14 ensures unique representation of isomorphism classes

Experimental Setup

Computational Environment

• Mid-range computer with 16-thread parallel computation
• Five-dimensional computation: less than 1 minute
• Six-dimensional computation: approximately 20 days

Data Availability

Complete classification data is available at 11, including explicit representations of all five and six-dimensional reflexive simplices.

Experimental Results

Main Classification Results

Classification 1.1: Under unimodular equivalence,

• Five-dimensional reflexive simplices: 220,794
• Six-dimensional reflexive simplices: 309,019,970

Corollary 1.2: Under isomorphism,

• Five-dimensional Gorenstein pseudo-weighted projective spaces: 220,794
• Six-dimensional Gorenstein pseudo-weighted projective spaces: 309,019,970

Theoretical Results

Theorem 3.3: For pseudo-weighted projective space Z, provides explicit formulas for Picard group and Gorenstein index:

• Pic(Z) = ⟨(LM, 0)⟩
• ι(Z) = lcm((LM)/gcd(LM, Σᵢwᵢ), μⱼ/gcd(μⱼ, Σᵢηᵢⱼ'))

Algorithm Efficiency Verification

The algorithm demonstrates good scalability in practical computation, with effective pruning strategies significantly reducing the search space.

Related Work

Historical Development

1. Batyrev (1994): Determined classification of all reflexive polygons
2. Kreuzer & Skarke (1998, 2000): Extended to three and four-dimensional reflexive polytopes
3. Schöller & Skarke (2019): Computed weight systems of five-dimensional reflexive polytopes

Uniqueness of This Work's Contribution

• First systematic treatment of high-dimensional reflexive simplices classification
• Establishes deep connections with toric geometry
• Provides practical computational algorithms and explicit theoretical formulas

Conclusions and Discussion

Main Conclusions

1. Successfully developed a universal classification algorithm for reflexive simplices
2. Completed full classification of five and six-dimensional reflexive simplices
3. Established one-to-one correspondence between reflexive simplices and Gorenstein pseudo-weighted projective spaces
4. Derived computational formulas for Picard group and Gorenstein index

Limitations

1. Computational Complexity: Six-dimensional computation requires 20 days; higher dimensions may be impractical
2. Memory Requirements: Storage and processing of large datasets impose high computational resource demands
3. Algorithm Optimization: Certain pruning strategies may have room for improvement

Future Directions

1. Algorithm optimization to handle higher dimensions
2. Exploration of applications of reflexive simplices in mirror symmetry
3. Investigation of asymptotic behavior and statistical properties of classification results

In-Depth Evaluation

Strengths

1. Theoretical Depth: Skillfully combines combinatorial geometry with algebraic geometry
2. Computational Innovation: Clever application of degree matrix methods and automorphism group theory
3. Practical Value: Provides complete classification data and computational tools
4. Rigor: All major results have complete mathematical proofs

Weaknesses

1. Scalability Limitations: Algorithm complexity grows exponentially with dimension
2. Limited Scope: Primarily restricted to reflexive simplices; not applicable to general reflexive polytopes
3. Implementation Details: Certain technical details could be described more thoroughly

Impact

1. Academic Contribution: Provides important tools for toric geometry and combinatorial geometry
2. Computational Geometry: Offers new approaches for classifying high-dimensional geometric objects
3. Practical Applications: Potential applications in mirror symmetry and string theory

Applicable Scenarios

• Research on toric varieties in algebraic geometry
• Polytope classification problems in combinatorial geometry
• Mirror symmetry research in mathematical physics
• High-dimensional object processing in computational geometry

References

1 Victor V. Batyrev, "Dual polyhedra and mirror symmetry for Calabi–Yau hypersurfaces in toric varieties" 2 Maximilian Kreuzer and Harald Skarke, "Classification of reflexive polyhedra in three dimensions" 3 Maximilian Kreuzer and Harald Skarke, "Complete classification of reflexive polyhedra in four dimensions"

This paper achieves breakthrough progress on the important mathematical problem of reflexive simplices classification, providing not only deep theoretical insights but also practically feasible computational algorithms. The innovation of its methodology and completeness of its results make it a significant contribution to the field.