🔢 mathematics

About Wess-Zumino-Witten equation and Harder-Narasimhan potentials

This paper identifies algebraic obstructions to the existence of approximate solutions to the Wess-Zumino-Witten equation for polarized families of complex projective manifolds, introduces a generalized Monge-Ampère equation utilizing Harder-Narasimhan filtrations to provide optimal approximations that minimize the Yang-Mills functional, and applies these results to prove an asymptotic converse to the Andreotti-Grauert theorem.

Original authors: Siarhei Finski

Published 2026-05-05

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Siarhei Finski

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are an architect trying to build a perfect, smooth dome over a complex, multi-layered landscape. In the world of mathematics, this "dome" is a special kind of shape (a metric) that makes the landscape perfectly balanced. The paper you are reading is about what happens when you try to build this dome over a specific type of landscape: a fiber bundle.

Think of a fiber bundle like a long, winding train. The "base" is the track (a surface), and the "fibers" are the individual train cars attached to it. The author, Siarhei Finski, is asking: Can we build a perfectly balanced dome over this entire train system?

Here is the breakdown of the paper's journey, using simple analogies:

1. The Problem: The Perfect Dome vs. The Bumpy Reality

In some cases, you can build a perfectly smooth, balanced dome. In mathematics, this is called a solution to the "Wess-Zumino-Witten (WZW) equation."

The Goal: Find a shape where the "curvature" (how much it bends) is perfectly even everywhere.
The Obstacle: Sometimes, the train cars (the fibers) are so mismatched or "unstable" that a perfect, smooth dome is mathematically impossible to build. It's like trying to stretch a perfect sheet of plastic over a pile of jagged rocks; it just won't lie flat without tearing or wrinkling.

The paper asks two big questions:

Why can't we build the perfect dome? (What are the algebraic "blockers"?)
If we can't build the perfect one, what is the best possible approximation we can make?

2. The "Harder-Narasimhan" Filter: Sorting the Train Cars

To understand why the dome fails, the author looks at the "train cars" (the fibers) and sorts them. He uses a mathematical tool called the Harder-Narasimhan filtration.

The Analogy: Imagine the train cars are made of different materials. Some are heavy and stable (like steel), others are light and wobbly (like cardboard). The Harder-Narasimhan filtration is a way of sorting these cars from "most stable" to "least stable."
The Result: This sorting process creates a "map" or a "profile" of the train. If the train is perfectly uniform (all cars are the same), the map is a single point. If the cars are mixed, the map is a spread-out curve.

3. The Main Discovery: The "Best Guess" Dome

The paper proves a remarkable thing: Even if a perfect dome is impossible, you can always build a "near-perfect" one.

The "Best Guess": The author constructs a sequence of shapes that get closer and closer to the perfect balance. These shapes don't solve the original equation perfectly, but they minimize the "wobble" (the error) as much as physically possible.
The Secret Ingredient: To build this "best guess" dome, the author uses a technique called Geometric Quantization.
- The Metaphor: Imagine you are trying to paint a smooth gradient on a wall, but you only have a coarse brush. Instead of trying to paint it in one go, you take a million tiny dots (pixels) and arrange them perfectly. As you zoom out, the dots look like a smooth, perfect gradient.
- The author uses high-level algebra (looking at the train cars from a very high "zoom" level) to create these "dots," and then zooms back out to create the smooth, approximate dome.

4. The "Harder-Narasimhan Potential": The Speedometer

The paper introduces a new concept called the Harder-Narasimhan potential.

The Analogy: Think of the train moving along the track. The "potential" is like a speedometer that tells you how fast the train should be going at any point to stay balanced, based on the mix of heavy and light cars.
The author shows that this "speedometer" reading is exactly what you need to plug into a new, slightly modified equation to get your "best guess" dome.

5. The Big Payoff: Connecting Two Worlds

The paper connects two different areas of math that usually don't talk to each other:

Algebra: The study of shapes and equations (like the stability of the train cars).
Geometry: The study of smooth curves and surfaces (like the dome).

The Conclusion:

If the train cars are perfectly balanced (mathematically "semistable"), you can build a perfect dome.
If they are not, you can't build a perfect dome, but you can build the closest possible approximation.
The paper gives a precise formula to calculate exactly how far you are from perfection based on the "mix" of the train cars.

6. A Special Case: The "Andreotti-Grauert" Converse

The paper also solves a specific puzzle proposed by the late mathematician Jean-Pierre Demailly.

The Puzzle: If you know that certain "holes" (cohomology) in the structure disappear as you look at it from further away, does that mean you can build a specific type of smooth shape?
The Answer: Yes, but only if the "track" (the base) is a simple curve (like a line or a circle). The author proves that in this specific setting, the disappearance of the holes guarantees the existence of the "best guess" dome.

Summary

In short, this paper is about making the best of a bad situation. When the perfect mathematical solution doesn't exist because the underlying structure is too messy, the author provides a recipe to build the absolute best approximation possible. He does this by sorting the messy parts, measuring their "instability," and using a high-tech "pixelation" technique to construct a shape that is as smooth and balanced as the laws of mathematics allow.

Technical Summary: Wess-Zumino-Witten Equation and Harder-Narasimhan Potentials

Problem Statement
The paper investigates the existence of solutions to the Wess-Zumino-Witten (WZW) equation on a polarized family of complex projective manifolds $\pi: X \to B$ . Specifically, for a relatively ample line bundle $L$ with first Chern class $[\alpha]$ , the equation is given by:
$\alpha^{n+1} \wedge \pi^*\omega_B^{m-1} = 0$
where $\omega_B$ is a Kähler form on the base $B$ . While this equation is well-understood when $B$ has a boundary (relating to Mabuchi geodesics), the case where $B$ is a compact manifold without boundary presents a different theory. The paper addresses two central questions:

Algebraic Obstructions: What are the algebraic obstructions governing the existence of approximate solutions to the WZW equation?
Optimal Alternatives: When exact or approximate solutions do not exist, what are the "best" alternatives? The paper seeks to construct sequences of forms that minimize the associated WZW functional.

The study is motivated by the generalization of the Hermite-Einstein equation on vector bundles. In the specific case where $X = \mathbb{P}(F^*)$ for a vector bundle $F$ over $B$ , the solvability of the WZW equation is equivalent to the existence of a Hermite-Einstein metric on $F$ , linking it to the Kobayashi-Hitchin correspondence. This work aims to extend this correspondence to fibrations not necessarily associated with vector bundles.

Methodology
The approach relies heavily on geometric quantization and the asymptotic analysis of direct image sheaves. The methodology proceeds through the following steps:

Harder-Narasimhan Filtrations and Potentials: The author utilizes the Harder-Narasimhan (HN) filtration of the direct image sheaves $E_k = R^0\pi_* L^{\otimes k}$ . A key insight is that these filtrations induce a bounded, submultiplicative filtration on the section ring $\bigoplus E_k$ (away from a negligible set).
Weighted Bergman Kernels: The paper defines a "Harder-Narasimhan potential," $HN_k(\omega)$ , derived from a weighted Bergman kernel associated with the HN filtration. As $k \to \infty$ , these potentials converge to a limiting function $HN(\omega)$ , which represents the speed of the geodesic ray associated with the HN filtration.
Approximate Critical Structures: Drawing on results by Atiyah-Bott, Daskalopoulos-Wentworth, and Sibley, the construction utilizes $L^1$ -approximate critical Hermitian structures on the vector bundles $E_k$ . These structures are adapted to the HN filtration and serve as the starting point for constructing geodesic rays.
Dequantization: The core technical contribution involves the "dequantization" of these geodesic rays. By constructing specific Hermitian metrics on $E_k$ that respect the algebraic structure of the ring bundle $\bigoplus E_k$ (via symmetric powers and quotient metrics), the author derives a sequence of $(1,1)$ -forms $\omega_\epsilon$ on $X$ .
Auxiliary Monge-Ampère Equation: The paper introduces an auxiliary equation:
$\omega^{n+1} \wedge \pi^*\omega_B^{m-1} = HN(\omega) \cdot \omega^n \wedge \pi^*\omega_B^m \cdot (n+1)$
Unlike the WZW equation, this equation admits approximate solutions for any polarized family.

Key Contributions and Results

Theorem 1.1 (Algebraic Characterization): The paper establishes a precise formula linking the infimum of the WZW functional to an algebraic quantity derived from the HN slopes. Specifically:
$\text{WZW}(c_1(L) - t\pi^*[\omega_B], \omega_B) = \int_{\mathbb{R}} |x - t| d\eta^{HN}(x) \cdot \int_X c_1(L)^n \pi^*[\omega_B]^m \cdot (n+1)$
Here, $\eta^{HN}$ is the weak limit of the probability measures formed by the normalized HN slopes of $E_k$ . This result provides a differential-geometric characterization of the HN measure and identifies the algebraic obstructions (the spread of $\eta^{HN}$ ) to the existence of solutions.
Theorem 1.7 (Existence of Approximate Solutions): The main technical achievement is proving that the auxiliary Monge-Ampère equation admits approximate solutions for any $\epsilon > 0$ . These solutions, $\omega_\epsilon$ , minimize the WZW functional and serve as the closest counterparts to true solutions when the latter do not exist.
Corollary 1.3 (Generalized Kobayashi-Hitchin Correspondence): The paper proves that the existence of approximate solutions to the WZW equation is equivalent to the semistability of the underlying algebraic data (specifically, that $\eta^{HN}$ is a Dirac mass). When specialized to projectivized vector bundles, this recovers a version of the Kobayashi-Hitchin correspondence.
Corollary 1.5 (Asymptotic Cohomology): As an application, the paper proves an asymptotic converse to the Andreotti-Grauert theorem (conjectured by Demailly) for fibered settings where the base is a curve. It shows that the sequence of forms minimizing the WZW functional saturates the lower bound given by the asymptotic cohomology.
Connection to Hessian Quotient Equations: The paper discusses the relationship between the WZW equation and Hessian quotient equations, noting that while the WZW equation is degenerate (non-elliptic linearization), the approximate solutions can be related to solutions of the J-equation (a type of Hessian quotient equation) when the base is a curve.

Significance and Claims
The paper claims to resolve the question of algebraic obstructions for the WZW equation in the boundaryless setting, providing a complete characterization via the Harder-Narasimhan measure. It asserts that the constructed approximate solutions are optimal in the sense that they minimize the associated Yang-Mills functional.

The significance lies in:

Generalization: Extending the Kobayashi-Hitchin correspondence beyond vector bundles to general fibrations.
New Equation: Introducing and solving an auxiliary Monge-Ampère type equation that generalizes the WZW equation by incorporating the weighted Bergman kernel of the HN filtration.
Methodological Innovation: Demonstrating that geometric quantization, specifically the dequantization of geodesic rays associated with $L^1$ -approximate critical structures, can be used to construct solutions to degenerate geometric PDEs where traditional continuity methods or geometric flows may not apply directly.

The author notes that the results are most robust when the base $B$ is a curve (dimension 1), where the relationship between the WZW functional and absolute Monge-Ampère functionals is direct. For higher-dimensional bases, the techniques for studying the WZW functional break down outside the relatively ample setting, and the existence of explicit solutions to the auxiliary equation remains an open question regarding regularity. The paper does not claim to resolve the general Demailly conjecture for all cases, but rather provides a proof for the fibered case over curves.