RC-positivity, comparison theorems and prescribed Hermitian-Yang-Mills tensors I

This paper solves the prescribed Hermitian-Yang-Mills tensor problem on compact Kähler manifolds by proving the existence and uniqueness of a smooth Hermitian metric for any given positive definite tensor under specific positivity conditions, utilizing a new comparison theorem to derive quantitative Chern number inequalities for holomorphic vector bundles and Fano manifolds.

The Gist

Imagine you are an architect trying to build a perfect structure on a curved piece of land. In the world of mathematics, specifically geometry, this "land" is a complex shape called a manifold, and the "structure" is a vector bundle (think of it as a collection of tiny, flexible sheets or fibers attached to every point of the land).

For decades, mathematicians have been trying to figure out how to make these structures "balanced" or "stable." This paper, written by Mingwei Wang, Xiaokui Yang, and the legendary Shing-Tung Yau, solves a major puzzle about how to prescribe exactly how these structures should behave.

Here is the breakdown of their discovery in simple terms:

1. The Big Problem: The "Shape" of Curvature

Imagine you have a rubber sheet (the manifold) and you are stretching a fabric over it (the vector bundle). The fabric has a natural "curvature" or tension, much like a drum skin.

• The Old Rule: Previously, mathematicians knew that if the fabric was naturally stable (mathematically "stable"), there was one specific way to stretch it so the tension was perfectly even everywhere. This was like finding a "Goldilocks" tension that wasn't too tight or too loose.
• The New Challenge: The authors asked a bolder question: What if we don't just want "even" tension? What if we want the tension to match a specific, pre-determined pattern we designed? Can we force the fabric to have exactly the curvature we want, as long as that pattern is "positive" (meaning it doesn't have weird, negative dips)?

2. The Main Discovery: The "Prescribed Tension" Theorem

The paper says YES.

If you start with a fabric that is already "positively curved" (it's already in a good, stable state), you can reshape it to match any positive curvature pattern you desire.

• The Metaphor: Imagine you have a lump of clay that is already smooth and round. The authors proved that you can mold this clay into any other smooth, round shape you want, provided the new shape doesn't have any sharp, inward-pointing spikes (which would be "negative").
• The Result: They found a unique way to stretch the fabric so that its internal tension matches your blueprint perfectly. There is only one way to do it, and it works every time.

3. The Secret Weapon: The "Comparison Theorem"

How did they prove this? They used a clever trick called a Comparison Theorem.

• The Analogy: Imagine you have two rubber bands. One is your "standard" band (the starting point), and the other is your "target" band (the one you want to create).
• The authors proved a rule: If your target band is "less tense" than your standard band in a specific mathematical sense, then your target band must be "smaller" or "tighter" in a physical sense.
• This rule acts like a guardrail. It prevents the fabric from stretching infinitely or collapsing. It ensures that as you try to mold the fabric, it stays within a safe, manageable range, allowing you to find the perfect fit without the math breaking down.

4. Why Does This Matter? (The Applications)

Why should a non-mathematician care about stretching imaginary fabrics?

• Quantifying the Universe: The authors used this new tool to create new "rules of the road" for geometry. They derived new inequalities (mathematical limits) that tell us how much "stuff" (like energy or mass) can fit into a certain space.
• Fano Manifolds: These are special, fancy shapes that appear in string theory and physics. The paper shows that on these shapes, you can always find a perfect "balance" of forces. This helps physicists understand the fundamental structure of the universe.
• Chern Numbers: These are like "ID tags" for shapes. The paper gives us a better way to calculate these tags, which helps us distinguish between different types of geometric shapes.

5. The "Line Bundle" Twist

The paper also looks at a simpler version of this problem: a single string (a line bundle) instead of a whole sheet.

• They showed that if you try to force the string to have a "negative" or "bad" tension pattern, the math breaks. You might get two different answers, or no answer at all.
• The Lesson: Nature (or the math) insists on positivity. You can only mold the fabric into shapes that are fundamentally "good" (positive). If you try to force a "bad" shape, the system rejects it.

Summary

Think of this paper as a master guide for geometric sculpting.

• Before: We knew how to make a sculpture that was "balanced."
• Now: We know how to sculpt a piece of geometric clay to match any specific, positive design we dream up, and we know exactly how to do it without the clay falling apart.

This work bridges the gap between abstract algebra (the rules of the shapes) and analysis (the actual stretching and bending), providing a powerful new tool for understanding the geometry of our universe.

Technical Summary

Here is a detailed technical summary of the paper "RC-positivity, Comparison Theorems and Prescribed Hermitian-Yang-Mills Tensors I" by Mingwei Wang, Xiaokui Yang, and Shing-Tung Yau.

1. Problem Statement

The paper addresses the prescribed Hermitian-Yang-Mills (HYM) tensor problem for holomorphic vector bundles over compact Kähler manifolds.

• Context: Given a compact Kähler manifold $(M, \omega_g)$ and a holomorphic vector bundle $E$ of rank $r$, the classical Hermitian-Einstein equation seeks a metric $h$ such that the contraction of the Chern curvature $\Lambda_{\omega_g} \sqrt{-1} R_h$ is proportional to the metric itself ($\lambda \cdot h$). This is central to the Donaldson-Uhlenbeck-Yau theorem, linking stability to the existence of such metrics.
• The Generalization: The authors generalize this by asking: Given an arbitrary smooth, positive definite Hermitian tensor $P \in \Gamma(M, E^* \otimes E^*)$, does there exist a unique smooth Hermitian metric $h$ on $E$ such that:
  $$\Lambda_{\omega_g} \sqrt{-1} R_h = P$$
• The Obstacle: Unlike the Calabi-Yau equation (prescribing Ricci curvature) or the standard HYM equation (prescribing a scalar multiple of the metric), the prescribed tensor $P$ is not necessarily a scalar multiple of the metric. The existence of solutions depends critically on the curvature properties of a background metric.

2. Methodology

The proof strategy diverges significantly from the continuity method used in the Calabi-Yau theorem or the heat flow methods used in the Donaldson-Uhlenbeck-Yau theorem. The authors employ a topological degree argument (openness and closedness) combined with a novel comparison theorem.

A. The Comparison Theorem (Uniqueness)

The core of the uniqueness argument is Theorem 1.2, a comparison principle for HYM tensors.

• Hypothesis: Let $h$ and $h_0$ be two metrics. If the HYM tensor of the background metric $h_0$ is positive definite ($\Lambda \sqrt{-1} R_{h_0} > 0$) and the HYM tensor of $h$ satisfies $\Lambda \sqrt{-1} R_h \leq \Lambda \sqrt{-1} R_{h_0}$, then $h \leq h_0$.
• Technique: The proof utilizes the section $H = h \cdot h_0^{-1}$ and analyzes the maximum eigenvalue of $H$. By constructing a specific test section involving powers of $((\Lambda + \epsilon)I - H)^{-2n}$ and using integration by parts with the Bochner-Kodaira formula, the authors derive a contradiction if the maximum eigenvalue exceeds 1. This establishes a $C^0$-estimate (metric bounds) based on curvature constraints.

B. The Existence Argument (Openness and Closedness)

The authors define a map $G: \text{Herm}^+(E) \to \text{Herm}(E)$ by $G(h) = \Lambda_{\omega_g} \sqrt{-1} R_h$. They aim to show $G$ is a bijection onto the space of positive definite tensors.

1. Openness: Established via the Implicit Function Theorem. The linearization of the map $G$ at a reference metric $h_1$ is shown to be an invertible, self-adjoint elliptic operator:
   $$\mathcal{L}(\Psi) = \Delta_{\partial, h_1} \Psi + \Omega_1 \cdot \Psi$$
   where $\Omega_1$ is the HYM tensor of $h_1$. The positive definiteness of $\Omega_1$ ensures the operator is coercive and invertible.
2. Closedness: Requires uniform $C^k$-estimates for a sequence of metrics $h_m$ where $G(h_m) \to P$.
   • $C^0$-estimate: Provided directly by the Comparison Theorem (Theorem 1.2).
   • $C^1$-estimate: Derived via a sophisticated Bochner-type calculation involving the tensor $T = \partial_{h_0} H \cdot H^{-1}$. The authors establish a differential inequality for $|T|^2$ and $\text{tr}_E H$ to bound the gradient.
   • Higher-order estimates: Once $C^1$ bounds are established, standard elliptic regularity theory ($W^{k,p}$ and $C^{k,\alpha}$ estimates) yields smooth convergence.

3. Key Contributions and Results

Main Theorem (Theorem 1.1)

Let $E$ be a holomorphic vector bundle over a compact Kähler manifold $(M, \omega_g)$. If there exists a smooth Hermitian metric $h_0$ such that its HYM tensor $\Lambda_{\omega_g} \sqrt{-1} R_{h_0}$ is positive definite, then for any Hermitian positive definite tensor $P$, there exists a unique smooth Hermitian metric $h$ satisfying:
$$\Lambda_{\omega_g} \sqrt{-1} R_h = P$$

Corollaries and Applications

1. Fano Manifolds (Corollary 1.4 & 1.5): On a Fano manifold, where the anticanonical bundle is ample, there exist Kähler metrics with positive Ricci curvature. Consequently, for any positive definite tensor $P$ (including the metric $g$ itself), there exists a unique metric $h$ on the tangent bundle $T^{1,0}M$ satisfying the prescribed equation.
   • Special Case: If $\text{Ric}(\omega_g) = c_0^{-1} \omega_g$, the solution is simply $h = c_0 g$.
2. Quantitative Chern Number Inequalities (Theorem 1.6): The authors derive a refined inequality for Chern numbers. If a metric $h_0$ exists such that $a \cdot h_0 \leq \Lambda \sqrt{-1} R_{h_0} \leq b \cdot h_0$, then:
   $$\int_M ((r-1)c_1(E)^2 - 2r c_2(E)) \wedge \omega_g^{n-2} \leq \frac{r(r-1)(b-a)^2}{8\pi^2 n^2} \int_M \omega_g^n$$
   This generalizes the classical Bogomolov-Gieseker inequality (which corresponds to the case $a=b$, i.e., the Hermitian-Einstein case).
3. Line Bundles and RC-positivity (Section 7): The paper clarifies the connection to RC-positivity (introduced by Yang). It shows that the condition $P > 0$ is necessary for uniqueness; if $P$ is not positive definite, multiple solutions or no solutions may exist. This links the vector bundle problem to the Kazdan-Warner problem for scalar curvature on line bundles.

4. Significance

• New Paradigm for HYM Equations: This work provides a new framework for solving non-linear geometric PDEs on vector bundles that does not rely on stability conditions (like slope stability) or heat flows. Instead, it relies on the existence of a "good" background metric (positive HYM tensor) and comparison principles.
• Bridge to RC-positivity: The results deepen the understanding of RC-positivity, a curvature condition introduced by Yang that generalizes positivity of the tangent bundle. The paper suggests that RC-positivity is the natural curvature condition for the existence of prescribed HYM tensors, analogous to how positive Ricci curvature relates to the Calabi conjecture.
• Alternative Proof of DUY: The authors note that their comparison theorem and existence result provide an alternative proof of the Donaldson-Uhlenbeck-Yau theorem, bypassing the traditional stability-to-existence flow arguments.
• Geometric Inequalities: The quantitative Chern number inequalities offer a new tool for studying the topology of vector bundles and Fano manifolds, providing bounds that depend on the "spread" ($b-a$) of the curvature eigenvalues rather than just the existence of a specific Einstein metric.

In summary, this paper solves a fundamental existence and uniqueness problem for a broad class of geometric PDEs on vector bundles, establishing a powerful comparison theorem as the linchpin for both uniqueness and a priori estimates, with immediate applications to Chern class topology and the geometry of Fano manifolds.