🔢 mathematics

On the existence of minimizer on a log Fano cone singularity

This paper proves that for a log Fano cone singularity over an uncountable algebraically closed field, the invariant $\delta(X,\Delta;\nu_0)$ admits a minimum among $\mathbb{T}$ -invariant valuations, a result established via a generic limit argument that fails without the log Fano cone structure.

Original authors: Donghyeon Kim

Published 2026-08-25

📖 3 min read🧠 Deep dive

CC BY 4.0

Original authors: Donghyeon Kim

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

In the vast landscape of modern mathematics, there is a field dedicated to understanding the shape and structure of spaces that exist in higher dimensions. These are not spaces we can walk through, but abstract geometric worlds defined by equations. Within these worlds, mathematicians often look for special points where the geometry becomes sharp or "singular," much like the tip of a cone or the corner of a pyramid. For decades, researchers have been trying to understand how these singularities behave, particularly when they are part of a larger system that possesses a kind of symmetry. The central question has been whether there is always a single, perfect way to measure the "stability" of these points—a specific mathematical value that represents the most efficient or balanced state. Finding this perfect measurement is crucial because it acts as a key to unlocking deeper theorems about how these complex shapes can be broken down and studied, a process known as stable degeneration. Without knowing that such a perfect measurement exists, the entire framework for analyzing these geometric structures remains incomplete.

A recent paper by Donghyeon Kim addresses this precise problem within a specific type of geometric setting known as a log Fano cone singularity. The author proves that in these specific, symmetric environments, there is indeed a single, optimal valuation—a mathematical tool used to measure the severity of a singularity—that minimizes a particular invariant called the local delta-invariant. In simpler terms, the paper demonstrates that if you are looking at a geometric shape with a sharp point that follows certain symmetry rules, there is always one specific way to measure that point that yields the lowest possible value. This finding confirms that the search for this "best" measurement is not a futile chase; the target actually exists. The proof relies on a sophisticated technique involving limits, where the author constructs a sequence of approximations to show that the minimum value is eventually reached by a specific, well-behaved mathematical object.

However, the paper also draws a sharp boundary around where this success applies. It explicitly shows that if you remove the specific symmetry conditions of the log Fano cone structure, the guarantee of finding a minimizer disappears. The author provides a concrete counterexample in a two-dimensional setting where, without the cone structure, no single valuation can achieve the minimum value. In this scenario, the values keep getting lower and lower, approaching a limit but never actually landing on a specific solution. This distinction is vital because it clarifies that the existence of the minimizer is not a universal law for all geometric singularities, but rather a special property that depends heavily on the underlying symmetry of the space.

The significance of this work lies in its ability to settle a long-standing question for a major class of geometric objects. By proving that a minimizer exists for log Fano cone singularities, the paper provides a solid foundation for further research into K-stability, a theory that connects geometry to physics and algebra. The author's approach involves carefully descending complex data from a large, uncountable field of numbers down to a smaller, countable field, allowing for a rigorous construction of the limiting object. This method ensures that the result is not just a theoretical possibility but a mathematical certainty within the defined scope. The paper concludes by noting that while the minimizer exists, questions remain about the specific algebraic properties of the object that achieves this minimum, inviting future mathematicians to explore the deeper nature of these optimal solutions.

Technical Summary: On the Existence of Minimizer on a Log Fano Cone Singularity

Problem Statement
This paper addresses the existence of a minimizer for the local $\delta$ -invariant, denoted $\delta(X, \Delta; \nu_0)$ , within the context of log Fano cone singularities. The invariant is defined for a log Fano cone singularity $x \in (X, \Delta, T)$ over an uncountable algebraically closed field of characteristic 0, where $\nu_0$ is a $T$ -invariant valuation centered at $x$ with finite log discrepancy $A_{X,\Delta}(\nu_0) < \infty$ . The value is given by:
$\delta(X, \Delta; \nu_0) := \inf_{\nu \in \text{Val}^*_{T, X, \ni x}} \frac{A_{X,\Delta}(\nu)}{S(\nu_0; \nu)}$
where $S(\nu_0; \nu)$ is the local $S$ -invariant. While the existence of minimizers for related invariants (such as log canonical thresholds, normalized local volumes, and global $\alpha$ - and $\delta$ -invariants) has been established in prior works (e.g., [JM12], [Blu18], [BJ20]), the existence of a minimizer for the local $\delta$ -invariant in the specific setting of log Fano cone singularities, particularly when $\nu_0$ is an arbitrary $T$ -invariant valuation (not necessarily a Reeb vector weight), remained an open question. The paper notes that if the log Fano cone structure is discarded, a minimizer may not exist, even in dimension 2.

Methodology
The proof employs the generic limit argument, a technique previously utilized in [dFM09], [dFEM10], [Blu18], and [BJ20]. The author adapts the approach from [BJ20] but faces a specific technical hurdle: the lack of a direct counterpart to [BJ20, Proposition 6.3] in the local cone setting. To overcome this, the paper introduces a descent strategy:

Descent to a Countable Subfield: All data associated with $(X, \Delta, T)$ and $\nu_0$ are descended to a countable algebraically closed subfield $k_0 \subseteq k$ .
Construction of Generic Limits: Using a diagonal subsequence argument and properties of vector bundles over Grassmannians (Lemmas 3.4–3.6), the author constructs a "generic limit" filtration. This involves analyzing the behavior of valuative ideals $a_p(\nu_i)$ modulo $a_N(\nu_0)$ as the sequence of valuations $\nu_i$ approaches the infimum.
Uniform Fujita Approximation: The paper extends the uniform approximation of the $S$ -invariant (Theorem 3.2) to the local setting, establishing bounds on the difference between the continuous $S$ -invariant and its discrete approximations $S_m$ .
Quasi-Monomiality: A key step involves proving that the constructed limit filtration is computed by a $T$ -invariant quasi-monomial valuation. This relies on Lemma 2.7, which asserts that for any $T$ -invariant graded sequence of ideals, there exists a $T$ -invariant quasi-monomial valuation computing the log canonical threshold.

Key Contributions and Results
The main result is Theorem 1.1, which states:

Let $x \in (X, \Delta; T)$ be an $n$ -dimensional log Fano cone singularity over an uncountable algebraically closed field $k$ . Let $\nu_0 \in \text{Val}^*_{T, X, x}$ be a $T$ -invariant valuation with $A_{X,\Delta}(\nu_0) < \infty$ . Then, there exists a $T$ -invariant quasi-monomial valuation $\nu' \in \text{Val}_{T, X, \ni x}$ that minimizes $\delta(X, \Delta; \nu_0)$ .

Key technical contributions include:

Extension of Definitions: The paper defines $\delta(X, \Delta; \nu_0)$ using the set of valuations $\text{Val}^*_{T, X, \ni x}$ (valuations whose center specializes to $x$ ) rather than strictly $\text{Val}^*_{T, X, x}$ (valuations centered exactly at $x$ ). Corollary 2.8 demonstrates that these definitions yield the same infimum, though a minimizer may only exist in the former set (Remark 2.9).
Generic Limit Adaptation: The paper successfully adapts the generic limit argument to the local cone setting by proving an analog of [BJ20, Proposition 6.7] via Lemma 3.5, circumventing the missing Proposition 6.3.
Counterexample Construction: In Section 4, the paper provides Example 4.2, a counterexample showing that the existence of a minimizer fails if the log Fano cone structure is removed, even for a klt pair on a regular surface. This highlights the necessity of the cone structure for the generic limit argument to descend the formal filtration to the ring $R$ .

Significance and Claims
The paper claims that the existence of a minimizer is a crucial component of the stable degeneration theorem in K-stability theory. By establishing the existence of a minimizer for the local $\delta$ -invariant in the log Fano cone setting, the work extends the framework of local K-stability (cf. [Li17], [LX18], [Zhu24a]) to cases where the reference valuation $\nu_0$ is an arbitrary $T$ -invariant valuation, not just a Reeb vector weight.

The author explicitly notes that the result generalizes the case where $\nu_0 = \text{wt}_\xi$ for a Reeb vector $\xi$ , which was previously proven in [Hua22, Theorem 5.0.4]. The paper does not claim to solve the minimizer problem for general local singularities without the cone structure; indeed, it posits Question 4.1, asking whether a minimizer exists for general klt pairs, and answers it negatively via the provided counterexample. Furthermore, the paper raises Question 4.4 regarding whether the Rees algebra of the minimizing valuation is finitely generated, noting this is known for Reeb vectors and normalized volume minimizers but remains open for general $\delta$ -minimizers.

The work relies on the assumption that the base field is uncountable and algebraically closed, and the proof is conducted within the framework of characteristic 0. The author acknowledges the use of AI tools for literature search and argument refinement but asserts full responsibility for the mathematical correctness of the main theorem and the independently verified counterexample.

More like this