The $G$-Noncommutative Minimal Model Program

This paper develops the $G$-equivariant noncommutative minimal model program by constructing quasi-convergent paths in spaces of Bridgeland stability conditions for $G$-equivariant coherent sheaves, utilizing induction techniques for finite groups and $\mathbb{T}$-stability conditions derived from small quantum cohomology for algebraic group actions.

The Gist

The Big Picture: Shaping the Universe with Math

Imagine you are an architect trying to simplify a chaotic, overgrown garden into a neat, beautiful park. In the world of mathematics, this "garden" is a complex geometric shape (an algebraic variety), and the "simplification" process is called the Minimal Model Program (MMP).

For decades, mathematicians have had a recipe for this: take a messy shape, cut out specific parts, and reshape it until you get a "minimal model" (the simplest version) or a "fiber space" (a shape that looks like a bundle of strings).

The Twist: This paper asks, "What if our garden isn't just a garden, but a garden that is being danced on by a group of people?"

• The "G": This stands for a Group (a set of symmetries or rules, like rotating a snowflake or shuffling a deck of cards).
• The Problem: When you try to simplify the garden, you must make sure you don't break the dance. The symmetries must be preserved at every step.

This paper builds a new mathematical toolkit to handle this "Group-Equivariant" (G-Equivariant) version of the problem, but with a very modern, "quantum" twist.

---

The Three Main Ingredients

To understand the paper, think of it as a recipe with three main ingredients:

1. The "Quantum Compass" (Stability Conditions)

In the old days, to simplify a shape, you just looked at it. But in this new "Noncommutative" world, you can't just look; you have to use a Stability Condition.

• The Analogy: Imagine a compass that doesn't point North, but points to "what is stable."
• In the world of quantum physics and advanced math, objects (like sheaves) can be unstable and wobble. A stability condition is a rule that tells you which objects are "solid" and which are "wobbly."
• The authors are creating a path (a journey) through a landscape of these rules. They want to walk a path where the objects slowly settle down into a stable, simplified form. This path is called a Quasi-Convergent Path.

2. The "Quantum Map" (Quantum Cohomology)

How do you know which path to walk? You need a map.

• The Analogy: Imagine the shape you are studying has a "quantum echo." If you shout at it, the echo comes back with hidden information about its shape and how it interacts with the universe. This echo is the Quantum Differential Equation.
• The paper uses this "echo" to calculate the exact coordinates for the compass (the stability condition). It's like using a GPS that doesn't just look at roads, but listens to the quantum vibrations of the universe to tell you the fastest route to the "minimal model."

3. The "Group Dance" (Equivariance)

This is the paper's unique contribution.

• The Analogy: Imagine you are simplifying a kaleidoscope. If you just break a piece off, the whole pattern shatters. You have to break it in a way that respects the symmetry of the kaleidoscope.
• The authors developed two main ways to do this:
  • Method A (For Finite Groups): If the group is a small, finite set of dancers (like a specific number of people), they use a technique called Induction. They take a solution for a single person (the non-equivariant case) and "copy-paste" it across the whole group, making sure everyone stays in sync.
  • Method B (For Continuous Groups): If the group is a continuous flow (like a spinning wheel or a torus), they introduce a new concept called T-Stability. Think of this as a "multi-dimensional compass" that can handle the spinning motion of the group without getting dizzy.

---

The Journey of the Paper

Here is how the authors navigate this landscape:

1. The Setup: They define the rules of the game. They explain how to handle shapes that have group symmetries and how to measure their "stability."
2. The Finite Group Trick: They show that if you already know how to simplify a shape for a single person, you can automatically figure out how to do it for a whole group of finite dancers. It's like learning to dance a waltz with one partner, and then realizing you can do it with a whole line of people by just following the same steps.
3. The Torus Solution: For shapes with continuous symmetry (like a sphere spinning), they invent a new type of stability condition (T-Stability). They prove that you can walk a path using the "Quantum Map" (Quantum Cohomology) to reach a stable state.
4. The Destination: They show that when you reach the end of this path, you don't just get a simpler shape; you get a specific "decomposition" of the shape.
   • The Metaphor: Imagine the complex shape is a giant Lego castle. The path they walk breaks the castle down into distinct, non-overlapping blocks. These blocks are the "Minimal Models."
   • Crucially, they prove that these blocks respect the group dance. If you rotate the castle, the blocks rotate together perfectly.

Why Does This Matter?

• Connecting Worlds: This paper connects three seemingly different worlds:
  1. Geometry: Simplifying shapes.
  2. Algebra: Breaking down complex categories into smaller pieces (Semiorthogonal Decompositions).
  3. Physics: Using quantum equations (Gromov-Witten invariants) to solve geometric problems.
• The "D-Equivalence" Conjecture: The paper suggests that if two shapes are "birationally equivalent" (you can turn one into the other by cutting and pasting), their "quantum DNA" (derived categories) is actually the same, even when symmetries are involved. This is a huge step toward proving that the universe's underlying structure is consistent, no matter how you look at it.

Summary in One Sentence

This paper builds a new mathematical bridge that allows us to simplify complex, symmetrical shapes by using "quantum echoes" to guide a path of stability, ensuring that the group symmetries (the "dance") are never broken during the process.

Technical Summary

1. Problem Statement and Motivation

The paper addresses the extension of the Noncommutative Minimal Model Program (NMMP), originally introduced by Halpern-Leistner [Hal24], to the equivariant setting.

• Classical Context: The classical Minimal Model Program (MMP) simplifies algebraic varieties via birational transformations (contractions, flips). Categorically, these transformations induce semiorthogonal decompositions (SODs) in the bounded derived category of coherent sheaves, $D^b(X)$.
• The NMMP Framework: Halpern-Leistner proposed a "noncommutative" approach where SODs are constructed via quasi-convergent paths in the space of Bridgeland stability conditions, $\text{Stab}(D^b(X))$. These paths are generated by the asymptotic behavior of solutions to quantum differential equations (QDEs) derived from quantum cohomology.
• The Gap: While the non-equivariant case is well-developed for specific varieties (e.g., projective spaces, blow-ups), the systematic incorporation of a group action $G$ (finite or reductive algebraic) remains largely unexplored. The authors aim to construct a G-equivariant NMMP (G-NMMP) that relates:
  1. SODs of the equivariant derived category $D^b_G(X)$.
  2. Equivariant quantum cohomology and QDEs.
  3. Quasi-convergent paths in the space of equivariant stability conditions.

2. Methodology and Framework

The authors develop the G-NMMP through two primary technical avenues, depending on the nature of the group $G$:

A. Finite Group Actions: Induction Techniques

For finite groups, the authors utilize induction and restriction functors between the non-equivariant derived category $D^b(X)$ and the equivariant category $D^b_G(X)$.

• Key Tool: They employ the induction theorem (building on [MMS09, QZ25, DHL25]) which establishes a closed embedding $\text{Res}^{-1}: \text{Stab}(D^b(X))^G \hookrightarrow \text{Stab}(D^b_G(X))$.
• Strategy: They lift known quasi-convergent paths from the non-equivariant setting (where solutions to QDEs are known) to the equivariant setting. This requires proving that the induced path preserves the spanning condition (ensuring asymptotic growth rates correspond to limit semistable objects) and geometricity.

B. Reductive Algebraic Group Actions: T-Stability Conditions

For connected reductive groups (e.g., torus actions), the standard stability space is insufficient because the group action introduces commuting auto-equivalences.

• New Concept: The authors introduce T-stability conditions on T-categories. A T-category is a triangulated category equipped with $m$ commuting auto-equivalences $T_i$.
• Definition: A T-stability condition $(\sigma, s)$ consists of a Bridgeland stability condition $\sigma$ and a tuple $s \in \mathbb{C}^m$ such that $T_i(\sigma) = s_i \cdot \sigma$. This generalizes "Gepner points" to a multi-parameter setting.
• Construction: They construct quasi-convergent paths in the space of T-stability conditions ($\text{TStab}$) using the small equivariant quantum cohomology and fundamental solutions of the equivariant QDE.

3. Key Contributions and Definitions

1. Proposal 3.11 (The G-NMMP Proposal):
   The central conjecture states that for a $G$-contraction $\pi: X \to Y$, there exists a quasi-convergent path $\sigma^G_t$ in $\text{Stab}(D^b_G(X))$ whose central charges are determined by the evaluation of the fundamental solution $\Phi^G_t$ of the G-truncated quantum differential equation:
   $$Z^G_t(E) = \text{ev}_s \int^{eq}_X \Phi^G_t(v^G(E))$$
   where $\text{ev}_s$ is an evaluation map dependent on a generic tuple $s$.

2. T-Stability Conditions (Definition 5.3):
   A rigorous definition of stability conditions compatible with commuting auto-equivalences, including a T-support property to ensure local finiteness and deformation properties (Proposition 5.8).

3. G-Point Sheaves (Definition 3.15):
   An equivariant analogue of skyscraper sheaves, defined as $P_x := \text{Ind}^G_{G_x}(\mathcal{O}^{G_x}_x)$. These are used to define geometric stability conditions in the equivariant setting (where all $P_x$ are stable with the same phase).

4. Spanning Condition (Definition 4.4):
   A crucial technical condition ensuring that the asymptotic growth of the central charge (derived from the QDE) corresponds to actual limit semistable objects in the derived category.

4. Main Results

Theorem 1.4: Lifting for Finite Groups

Let $X$ be a smooth projective variety with a finite group action $G$. If a quasi-convergent path $\sigma_t$ exists in the non-equivariant setting solving the NMMP proposal, then the induced path $\sigma^G_t = \text{Res}^{-1}(\sigma_t)$ in $\text{Stab}(D^b_G(X))$ solves the G-NMMP proposal.

• Implication: This recovers results for finite group actions on projective spaces and blow-up surfaces by lifting known non-equivariant solutions (e.g., from [Zul24, Kar24]).
• Geometricity: If the starting path is geometric, the induced equivariant path is also geometric.

Theorem 1.5: T-Stability for Projective Spaces

For projective space $\mathbb{P}^{m-1}$ with a torus $T = (\mathbb{C}^*)^m$ action, the authors construct explicit quasi-convergent paths in the space of T-stability conditions.

• Construction: Using the fundamental solution of the equivariant quantum differential equation and the Gamma conjecture II framework.
• Result: For any admissible parameter $\theta$, there exists a path $\sigma_t$ starting from a geometric stability condition (for $m=3$, i.e., $\mathbb{P}^2$) that yields a semiorthogonal decomposition:
  $$D^b_T(\mathbb{P}^{m-1}) = \langle E_1 \otimes \text{Rep}(T), \dots, E_n \otimes \text{Rep}(T) \rangle$$
  where $E_i$ are the standard exceptional objects.

Corollary 6.3: D-Equivalence Conjecture

Under the assumptions of the G-NMMP proposal, the authors prove that for finite groups, if two smooth projective varieties $X$ and $X'$ are $G$-birationally equivalent and have base-point free $G$-invariant canonical linear systems, then there is a canonical admissible embedding $D^b_G(X) \hookrightarrow D^b_G(X')$. If both are minimal models, they are derived equivalent ($D^b_G(X) \cong D^b_G(X')$).

Proposition 6.5: Dubrovin's Conjecture

The paper establishes a one-way implication of Dubrovin's conjecture in the equivariant setting: the existence of a full exceptional collection in $D^b_G(X)$ implies the existence of a specific asymptotic behavior in the solutions to the equivariant QDE.

5. Significance and Impact

• Unification: The paper successfully unifies birational geometry, derived categories, and quantum cohomology in the presence of group symmetries. It provides a categorical mechanism to understand how symmetries affect the "minimal model" structure of a variety.
• New Tools: The introduction of T-stability conditions provides a robust framework for studying stability in categories with commuting auto-equivalences, which is essential for torus actions and potentially other reductive group actions.
• Explicit Constructions: By leveraging induction and explicit QDE solutions, the authors provide concrete examples of equivariant SODs for projective spaces and blow-ups, moving beyond abstract existence proofs.
• Future Directions: The work lays the groundwork for studying the G-equivariant Dubrovin conjecture and the D-equivalence conjecture in higher dimensions and for more complex group actions, suggesting that equivariant parameters do not alter the asymptotic behavior of the QDE solutions but rather refine the decomposition structure.

In summary, Wu and Zhang provide a comprehensive framework for the G-NMMP, demonstrating that the deep correspondence between quantum differential equations and derived category structures holds true even when group symmetries are systematically incorporated.