Tautological relations and integrable systems

Basic Information

• Paper ID: 2210.07552
• Title: Tautological relations and integrable systems
• Authors: Alexandr Buryak, Sergey Shadrin
• Classification: math.AG math-ph math.MP
• Publication Time/Conference: Épijournal de Géométrie Algébrique, Volume 8 (2024), Article No. 12
• Paper Link: https://arxiv.org/abs/2210.07552

Abstract

This paper proposes a family of conjectural relations in the tautological cohomology of the moduli space $M_{g,n}$ of stable algebraic curves. A significant portion of these relations possess surprisingly simple form: the tautological classes involved are given by stable graphs that are trees, decorated only on half-edges with powers of psi classes. The authors prove that the proposed conjectural relations imply certain fundamental properties of the Dubrovin-Zhang (DZ) hierarchy and the double ramification (DR) hierarchy associated with F-cohomological field theory. These relations naturally extend a similar system of conjectural relations proposed by the first author with Guéré and Rossi in earlier work, which accounts for the normal Miura equivalence of DZ and DR hierarchies associated with arbitrary cohomological field theory. Finally, using a variant of the Liu-Pandharipande method, the authors prove all aforementioned relations for $n=1$ and arbitrary $g$, which may be of independent interest.

Research Background and Motivation

Problem Background

The core problem investigated in this paper concerns relations in the tautological cohomology of the moduli space $M_{g,n}$ of stable algebraic curves, and the deep connections between these relations and integrable systems. Specifically:

1. Relationship between moduli space geometry and integrable systems: Since Witten's conjecture (proved by Kontsevich), there has existed a profound connection between the geometry on $M_{g,n}$ and integrable systems. The conjecture states that the generating function of integrals $\int_{M_{g,n}} \prod_{i=1}^n \psi_i^{d_i}$ provides a solution to the KdV hierarchy.
2. DZ hierarchy and DR hierarchy: Dubrovin-Zhang constructed integrable hierarchies associated with cohomological field theory (DZ hierarchy), while Buryak later introduced the double ramification hierarchy (DR hierarchy). The relationship between these two hierarchies is an important open problem.
3. Polynomiality problem: The polynomiality of DZ hierarchy equations is a fundamental yet difficult problem, resolved only in the semisimple CohFT case.

Research Motivation

• Unified theoretical framework: Establish Miura equivalence between DZ and DR hierarchies
• Extension of existing results: Generalize the relations in BGR19 to the more general F-CohFT setting
• Resolution of fundamental problems: Prove polynomiality of the DZ hierarchy and related fundamental properties

Core Contributions

1. Proposed a new family of conjectural relations: In the tautological cohomology of $M_{g,n+m}$, proposed parametrized conjectural relations with simple form and profound geometric significance
2. Established connections with integrable systems: Proved that these conjectural relations imply fundamental properties of DZ and DR hierarchies associated with F-CohFT
3. Proved special cases:
   • Completely proved the case $n=1$ and arbitrary $g$
   • Completely proved the case $g=0$ and arbitrary $n$
4. Provided explicit geometric formulas: Gave explicit geometric formulas for the polynomial part of the DZ hierarchy
5. Simplified the relation system: Proved that for $m \geq 2$, the entire relation system is equivalent to finitely many relations of degree $2g+m-1$

Detailed Methodology

Task Definition

The main tasks investigated in this paper are:

• Input: The moduli space $M_{g,n}$ of stable algebraic curves and its tautological cohomology classes
• Output: Linear relations in tautological cohomology
• Constraints: Relations must be compatible with fundamental properties of integrable systems

Core Mathematical Framework

1. Construction of Tautological Cohomology Classes

For $m \geq 0$, define the class $B^m_{g,d} \in R^{\sum d_i}(M_{g,n+m})$:

$B^m_{g,d} := \sum_{T \in \text{SRT}^{(b,c,a)}_{g,n,m;\circ}} (-1)^{\deg(T)-1} e_*[T,d]$

where $T$ ranges over stable rooted trees satisfying specific conditions.

2. Formulation of Conjectural Relations

Conjecture 1 ($m \geq 2$): For $\sum d_i \geq 2g + m - 1$, we have $B^m_{g,d} = 0$.

Conjecture 2 ($m = 1$): $B^1_{g,d} = A^1_{g,d}$, where $A^1_{g,d}$ is defined by double ramification cycles.

Conjecture 3 ($m = 0$): $B^0_{g,d} = A^0_{g,d}$, which are the relations from BGR19.

3. Equivalent Characterization

Using generating polynomials $P_{g,n,m}(x_1, \ldots, x_n)$, Conjecture 1 is equivalent to: $\deg P_{g,n,m} \leq 2g + m - 2$

Technical Innovations

1. Classification of Stable Rooted Trees

Introduced the concepts of balanced trees, complete trees, and admissible trees, providing systematic combinatorial classification:

• Balanced trees: Every vertex except the root has at least one additional leg
• Complete trees: Satisfy specific hierarchical structure conditions
• Admissible trees: Satisfy degree constraint conditions

2. Generalization of Liu-Pandharipande Method

For the case $n=1$, used localization formulas on relative stable map moduli spaces:

$\epsilon_*(\mathcal{I}_g \cap [M_{g,1}(\mathbb{P}^1,1)]^{\text{vir}}) = \sum_{\text{components}} \frac{[\text{component}]^{\text{vir}}}{e_{C^*}(N^{\text{vir}})}$

3. Systematic Reduction

Proved that the entire relation system can be reduced to finitely many fundamental relations: $\tilde{B}^m_{g,d} = 0 \text{ for } d_i \geq 1, \sum d_i = 2g + m - 1$

Experimental Setup

Verification Methods

Since this is pure mathematical theoretical work, "experiments" are primarily theoretical verifications:

1. Direct calculations in special cases:
   • Explicit computations for low genus cases
   • Verification for small numbers of marked points
2. Consistency checks with known results:
   • Compatibility with Witten's conjecture
   • Consistency with existing DZ/DR hierarchy theory

Proof Strategy

• Case $n=1$: Use localization formulas for relative stable maps on $\mathbb{P}^1$
• Case $g=0$: Exploit the cohomological structure of $M_{0,n}$ and tensor product properties of CohFT

Experimental Results

Main Theorems

Theorem 2.2: Conjectures 1, 2, 3 hold for $n=1$.

Theorem 2.3: Conjectures 1, 2, 3 hold for $g=0$.

Application Results

Theorem 4.7: Conjecture 1 ($m=2$) implies polynomiality of the DZ hierarchy for arbitrary F-CohFT.

Theorem 4.10: Conjecture 2 implies that the DZ hierarchy of F-CohFT is polynomial and related to the DR hierarchy via Miura transformation.

Concrete Computational Examples

For $n=1, m \geq 2$, relations take the form: $\sum_{k \geq 1} (-1)^{k-1} \sum_{(g,d) \in S^{m,k}_{g,d}} \text{[decorated graph]} = 0$

where decorated graphs involve only powers of psi classes.

Related Work

Historical Development

1. Witten's Conjecture (1991): Established connection between moduli space geometry and KdV hierarchy
2. Dubrovin-Zhang Theory (2001): Constructed integrable hierarchies for general CohFT
3. Double Ramification Hierarchy (Buryak 2015): Provided new construction of integrable systems
4. BGR Relations (2019): Established tautological relations for DR/DZ equivalence

Position of This Work

This paper naturally extends the work of BGR19 in several directions:

• Extension from CohFT to F-CohFT
• Extension from $m=0$ to arbitrary $m \geq 0$
• Provides a more complete theoretical framework

Conclusions and Discussion

Main Conclusions

1. Existence of tautological relations: The proposed relations possess profound geometric and algebraic significance
2. Fundamental properties of integrable systems: These relations control core properties of DZ and DR hierarchies
3. Partial proof: The conjectures are completely proved in important special cases

Limitations

1. General case unresolved: For general $(g,n,m)$, the conjectures remain open
2. Computational complexity: Direct verification for high genus cases is technically challenging
3. Scope of applications: Primarily limited to algebraic geometry and mathematical physics

Future Directions

1. Complete proof: Seek proof methods for the general case
2. Generalized applications: Explore applications to other geometric and physical problems
3. Computational methods: Develop more efficient computational and verification methods

In-Depth Evaluation

Strengths

1. Theoretical depth: Connects deep structures in algebraic geometry, integrable systems, and mathematical physics
2. Methodological innovation: Cleverly combines combinatorial and geometric techniques
3. Result significance: Addresses fundamental problems in the field
4. Clear exposition: Rigorous mathematical expression with clear logical structure

Weaknesses

1. Conjectural nature: Core results remain conjectures requiring further proof
2. Technical threshold: Requires deep background in algebraic geometry and integrable systems
3. Computational verification: Lacks large-scale numerical verification

Impact

1. Theoretical contribution: Provides new perspectives for tautological cohomology and integrable systems theory
2. Methodological value: The generalization of Liu-Pandharipande method has independent value
3. Future research: Establishes foundation for subsequent research in the field

Applicable Scenarios

• Research on moduli space theory in algebraic geometry
• Integrable systems theory in mathematical physics
• Research on cohomological field theory and related geometric structures

References

This paper cites important literature in the field, including:

• Wit91 Witten's original work
• DZ01 Dubrovin-Zhang hierarchy theory
• Bur15 Establishment of double ramification hierarchy
• BGR19 Related tautological relations work
• LP11 Liu-Pandharipande method