Geometric Stability Conditions in Algebraic Geometry

• Geometric stability conditions are rigorously defined pairs, combining a heart of a t-structure with a central charge, to formalize object stability in triangulated categories.
• They play a key role in constructing and classifying moduli spaces via wall–chamber decompositions and the analysis of discriminant functions like the Le Potier curve.
• This framework bridges algebraic geometry, representation theory, and mathematical physics, with emerging applications in machine learning and dynamical systems.

Geometric stability conditions constitute a central concept in modern algebraic geometry, homological mirror symmetry, and mathematical physics. These conditions formalize a notion of “stability” for objects in triangulated categories, particularly derived categories of coherent sheaves or Fukaya categories, by encoding geometric and algebraic constraints through the pairing of a central charge and a t-structure. The geometric subclass—so named because it reflects the structure of underlying varieties and their moduli—serves as a crucial organizing principle for wall-crossing phenomena, moduli construction, representation theory, and topological quantum phases.

1. Foundational Definition and Structure

Let $X$ be a smooth projective variety over $\mathbb{C}$, and let $D^b(X)$ denote its bounded derived category of coherent sheaves. A Bridgeland stability condition on $D^b(X)$ is a pair $\sigma=(\mathcal{A},Z)$, where $\mathcal{A}$ is the heart of a bounded t-structure on $D^b(X)$, and the central charge $Z:K(D^b(X))\to\mathbb{C}$ is a group homomorphism. This data must satisfy two principal axioms:

• Every nonzero object $E\in\mathcal{A}$ has $Z(E)\in\mathbb{R}_{> 0} e^{i\pi\phi(E)}$ for a unique phase $\phi(E)\in(0,1]$.
• Harder–Narasimhan filtrations exist in $\mathcal{A}$.

Additionally, the support property requires a quadratic form $Q$ on the numerical $K$-lattice $K_{\mathrm{num}}(X)$, negative-definite on $\ker Z$ and nonnegative on semistables (Dell, 2023).

A stability condition is called geometric if all skyscraper sheaves $\mathcal{O}_x$ are $\sigma$-stable (necessarily of the same phase), aligning the notion of stability with the geometry of the underlying variety (Dell, 2023). The set $\Gstab(X)$ of geometric stability conditions forms a robust and often connected submanifold of the full stability space.

2. Construction and Classification: Surfaces and Curves

Surfaces. The construction on smooth projective surfaces proceeds via a double tilt: first, defining a torsion pair with respect to the classical slope $\mu_H$, and then tilting again at a relative slope in the new heart. The associated central charge reads: $$Z_{H,B,\alpha,\beta}(E) = (\alpha - i\beta) H^2 \ch_0(E) + (B + i H)\cdot \ch_1(E) - \ch_2(E)$$ Geometric stability arises for parameter regions where $\alpha > \Phi_{X,H,B}(\beta)$, with $\Phi$ the Le Potier function, expressing a sharp discriminant bound (Dell, 2023). On projective surfaces, $\Gstab(X)$ is always connected and can be realized as an explicit connected open submanifold in $\NS_{\mathbb{C}}(X)\times\mathbb{C}$.

Curves. On smooth projective curves of genus $g\geq 1$, all locally finite stability conditions are geometric and can be described up to $\mathbb{C}$-action by

$$Z_{(\alpha,\beta)}(r,d) = -d+(\beta + i\alpha) r,\quad \alpha>0$$

with heart Coh($C$), yielding $\Stab(C)\cong \mathbb{C}\times\mathbb{H}$. For singular irreducible curves (with at worst Gorenstein singularities), the connected component generated by geometric stability conditions remains closed and characterized by analogous data; the stability manifold is governed by the structure of the curve’s $K_0$ group and duality properties (Liu, 14 Sep 2025).

3. Symmetry, Inheritance, and Special Quotients

A key feature is the functorial invariance of geometric stability conditions under group actions and autoequivalences:

• Finite Abelian Quotients: If $G$ acts freely on $X$ and $Y=X/G$, there is an analytic isomorphism between $G$-invariant geometric stability conditions on $X$ and $\widehat{G}$-invariant geometric conditions on $Y$, where $\widehat{G}$ is the group of irreducible representations. The induced geometric chambers on the quotient inherit richness from the cover, and special classes (Beauville-type, bielliptic surfaces) exhibit full connected components of geometric conditions even with non-finite Albanese morphism (Dell, 2023).
• Autoequivalences and Fourier–Mukai Transforms: On Weierstrass elliptic surfaces, the relative Fourier–Mukai transform permutes geometric chambers within the stability manifold. Exact relations between stability data before and after the transform are given in terms of explicit transformations of the (central charge, heart) pairs, preserving support and chamber structure (Lo et al., 2022).
• Spherical Twists and Chamber Expansion: Spherical twists by torsion sheaves, especially on K3 surfaces, can produce stability conditions lying outside the classical geometric chamber but within the same connected component, corresponding to chambers associated with non-nef divisors (Collins et al., 2024).

4. Wall–Chamber Structure and the Le Potier Function

Geometric stability conditions display intricate wall–chamber decompositions in parameter spaces, governed by discriminant functions, Chern class data, and the Le Potier curve. For surfaces such as $\mathbb{P}^2$, the Le Potier function describes the sharp locus where Chern classes of Gieseker-semistable sheaves accumulate. The geometric chamber comprises those stability conditions with parameters lying above the Le Potier curve; boundaries correspond to loci where exceptional bundles or skyscrapers change phase or lose stability (Li, 2016, Dell, 2023).

The Le Potier function for a surface $(X,H)$ is

$$\Phi_{X,H}(x) = \limsup_{\mu_H(F)\to x} \frac{\ch_2(F)}{H^2\ch_0(F)}$$

where the supremum is over all semistable sheaves of slope approaching $x$. For surfaces that are finite abelian quotients of varieties with finite Albanese (including Beauville-type surfaces), $\Phi_{X,H}(x) = \frac{1}{2}x^2$ and is continuous—a fact that refutes certain earlier conjectures predicting discontinuities at $\mu=0$ (Dell, 2023).

5. Connections to Moduli, Representation Theory, and Mathematical Physics

Geometric stability conditions underlie the construction of modul spaces of semistable objects (sheaves, complexes, modules) in a broad spectrum of settings:

• Moduli Spaces: Wall–crossing phenomena governed by geometric chambers yield birational transformations and explicit descriptions of moduli spaces, particularly in the context of threefolds and blowups, where perverse t-structures provide the link between perverse and geometric stability (Zhang, 31 Mar 2025, Li, 2015).
• Representation Theory: For derived categories of quiver modules, geometric models (e.g., “total stability” via polygons) provide explicit stability functions making all indecomposables stable, resolving conjectures in the theory of Dynkin quivers (2208.00073).
• Homological Mirror Symmetry: The category of Lagrangians in symplectic geometry admits geometric stability structures compatible with wall-crossing under mirror operations, such as the Atiyah flop, effecting chamber transitions in derived Fukaya categories (Fan et al., 2017).
• Quantum Theory: The geometric stability hypothesis for fractional Chern insulators asserts that robust many-body quantum phases arise when the single-particle band geometry (Berry curvature, Fubini–Study metric) approximates uniform Kähler geometry. Explicit inequalities quantify when the geometric setting supports a robust spectral gap (Jackson et al., 2014).

6. Applications and Implications Beyond Classical Contexts

Beyond algebraic geometry and representations, geometric stability notions appear as organizing principles in diverse contexts:

• Geometric Invariant Theory (GIT): Stability conditions generalize to stacks and schemes with group actions via central charges defined on the stack of graded points; these formal frameworks subsume K-stability of polarized varieties and are connected to moment map equations and differential geometry (e.g., deformed Hermitian Yang–Mills metrics, Z-critical connections) (Dervan, 2022, Dervan et al., 2020).
• Machine Learning and Data Analysis: “Geometric stability” for learned representations measures the reliability of the pairwise geometry of data embeddings under perturbations—quantified by rank correlations of dissimilarity matrices—differentiating it from pure similarity notions. Geometric stability acts as a sensitive indicator of drift, robustness, and controllability in neural representations, underpinning monitoring and model selection (Raju, 14 Jan 2026).
• Dynamical Systems: Geometric criteria (e.g., the steepness property, positivity of stiffness tensors) define local transversality conditions ensuring long-term stability of nearly-integrable Hamiltonian systems and spatially modulated phases in physics, with explicit (semi)-algebraic characterizations facilitating constructive verification (Barbieri, 2024, Domokos et al., 2013, Garg et al., 2018).

7. Principal Developments and Current Research Frontiers

The theory of geometric stability conditions exhibits a range of current developments:

• Analytic and topological descriptions of the stability manifolds for complex surfaces and higher-dimensional varieties, with explicit computations of chamber structures and boundary phenomena.
• Establishment of functorial correspondences and invariance results (under group actions, autoequivalences, and perverse decompositions) delineating how geometric loci are preserved or transformed.
• Counterexamples to conjectures on discriminant continuity and predictions of non-geometricity in exotic settings (e.g., non-finite Albanese), highlighting the flexible but subtle interplay between geometry and derived categories.
• Extensions to non-commutative and Fukaya categories, representation spaces of quantum groups, and applications in mathematical physics and machine learning, demonstrating the unifying power of geometric stability.

Ongoing questions include the global topology of stability manifolds (e.g., contractibility, wall–chamber decompositions), explicit moduli construction via geometric stability data in threefold and higher settings, and the generalization to wild and affine categories. Further applications in physics and computer science are emerging, establishing geometric stability as a cross-disciplinary principle.