Preprint

New theorem sets conditional curvature bounds for geometric equations

Preprint: A mathematical study reports conditional curvature and Hessian estimates alongside a new concavity inequality.

The central finding in a new mathematical preprint is a conditional sign result: a quadratic expression tied to the second variation of log λ_max is nonnegative for every ξ when 0<F/a^k≤η*, where η* is a sufficiently small threshold. Put less technically, the paper identifies a regime in which this spectral calculation has a favorable sign. The theorem is stated for ordered spectral vectors in Γ_k, with n≥3 and 2≤k<n, and for γ in the displayed range.

The paper’s target is precisely this sign: its stated aim is to establish a new concavity inequality for σ_k that controls the quadratic form arising from the second variation of log λ_max. Here, λ_max is the largest spectral value in the expression, and Qγ(λ;ξ) is the quadratic form whose sign is being assessed. The threshold η* is asserted to exist, but its numerical value is not reported in the supplied statement. The result is therefore conditional on the smallness requirement, rather than a blanket statement for every normalized value of F/a^k.

Splitting the difficult calculation

To establish the inequality, the proof uses an easy–hard decomposition. In the easy region, optimal constrained concavity handles the calculation. In the hard region, the argument moves to Gårding-root coordinates and uses concavity of ordered inverse-root partial sums. These are the two main coordinate and concavity tools identified in the proof strategy.

One hard-region lemma adds a favorable detail: when the roots are distinct, the second-order remainder produced by the chain rule is nonnegative. A chain-rule remainder is the extra second-order term that appears when a function of the spectral variables is differentiated twice. The nonnegative sign is part of the hard-region calculation, while the final quadratic-form result remains tied to its stated parameter range and smallness condition.

From spectral algebra to geometry

The main application moves from spectral algebra to geometry. For n≥3 and n/2≤k<n, the paper considers a closed, smooth, star-shaped, k-convex hypersurface and a positive C² right-hand side f(X,ν). Under those conditions, the maximum principal curvature over M is bounded above by C. The constant C is existential, and its explicit numerical value is not given in the supplied analysis.

A related corollary stays inside the k-Hessian setting. For n≥3 and n/2<k<n, it considers a C⁴ k-convex solution of σ_k(D²u)=1. When b=log λ_max is sufficiently large, the paper states a viscosity Jacobi inequality for every ε in the allowed range. The conclusion is tied to that specified regularity class and size condition on b.

The second major application is a global-to-boundary Hessian estimate for a bounded-domain k-Hessian problem. For n≥3 and n/2≤k<n, the global Hessian maximum is bounded by C times one plus the boundary Hessian maximum, with C depending on the domain, C¹ quantities, and the listed quantities associated with f. The estimate retains boundary second derivatives in its bound; it is not a boundary-free interior estimate.

What the result does—and does not—cover

Taken together, the results are conditional estimates. The central inequality needs F/a^k to be below η*, the curvature result needs the stated smoothness, star-shapedness, k-convexity and positive C² assumptions, and the Hessian result keeps its boundary dependence. The applications start with a hypersurface or solution satisfying those assumptions; they do not supply a separate existence result. Broader curvature estimates for general positive f(X,ν) outside n/2≤k<n are not established by the supplied results.

The document is an arXiv preprint, version 1, dated 26 Aug 2026. The acknowledgments list support from Grant 2025YFA1017603 of the National Key R&D Program of China. The author reports using ChatGPT for language editing and presentation, with the output reviewed and independently verified.

Paper data and sources

Original title: Global Curvature Estimates for $σ_k$ Curvature Equations with $k\geq n/2$
Authors: Jin Yan
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text

Versions and corrections

  1. Published automatically after legal-source, freshness, evidence, and independent-verification gates passed.