On the number of solutions to the planar dual Minkowski problem

Yannan Liu, Jiangbo LuScience 2026 · DOI

TL;DR

This paper investigates the number of solutions to the planar dual Minkowski problem, which is equivalent to solving a specific second-order ordinary differential equation. The authors provide a complete answer for the number of solutions for different ranges of a parameter $q$, revealing a rich structure of constant and non-constant solutions. This work clarifies a long-standing question in the field and has implications for understanding the geometry of convex bodies.

What Problem Does This Paper Address?

The dual Minkowski problem, first proposed in 2016, asks for the conditions under which a given measure on a sphere can be interpreted as the dual curvature measure of a convex body. For the specific case of a planar convex body and a constant dual curvature, this problem reduces to solving a Monge-Ampère type equation. When the dual curvature is a constant, this equation simplifies to a second-order ordinary differential equation: $h(|h|^2 + h'^2)^{\frac{q}{2}} (h'' + h) = 1$ on $S^1$. A crucial aspect of this problem is determining the number of positive solutions, which correspond to valid convex bodies. While solutions have been studied for various ranges of the parameter $q$, the case $0 < q \le 4$ (excluding $q=2$) remained largely open, with no complete answer regarding the uniqueness or nonuniqueness of solutions. This paper aims to fill this gap by providing a complete characterization of the number of solutions for $0 < q \le 4$ and also extends results for $q > 4$.

Methods

The core of the paper involves analyzing the behavior of solutions to the ordinary differential equation (1.2). The authors employ a combination of theoretical analysis and numerical estimation.

  • Integral Representation: The problem of finding the period of solutions is transformed into analyzing a specific integral, denoted as $\Theta_q(r)$. The number of solutions is then directly related to the number of times this integral equals $\frac{2\pi}{k}$ for some positive integer $k$.
  • Monotonicity Analysis: For the case $0 < q \le 2$, the authors prove that $\Theta_q(r)$ is strictly decreasing with respect to $r$. This monotonicity is crucial for determining the number of solutions. They leverage a monotonicity result from a previous work by Andrews.
  • Duality: The paper utilizes the duality between solutions of the dual Minkowski problem and the $L_p$-Alexandrov problem. This allows results for one problem to be translated to the other.
  • Numerical Estimation and Theoretical Bounds: For cases where monotonicity is lost (e.g., $q > 3$), the authors resort to a combination of theoretical bounds and numerical estimations to analyze the behavior of $\Theta_q(r)$. This involves developing sophisticated techniques to bound the integral and using computational methods to verify these bounds.
  • Taylor Expansions: For specific cases, particularly for $q=4$, detailed Taylor expansions of relevant functions are used to derive precise inequalities and bounds.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • Complete Characterization for $0 < q \le 4$: The paper provides a definitive answer to the number of solutions for the planar dual Minkowski problem with constant dual curvature for $0 < q \le 4$. This resolves a significant open problem.
  • Rich Solution Structure: The number of solutions is not simply one or none. For $0 < q < 1$, there are exactly two solutions (a constant and a non-constant one). For $1 \le q \le 4$, there is only one solution (the constant one). This reveals a non-trivial dependence on the parameter $q$.
  • Nonuniqueness for $q > 4$: The paper establishes that for $q > 4$, there are at least $[\sqrt{q}] - 2$ non-constant positive smooth solutions, indicating that the number of solutions grows with $q$.
  • Duality as a Powerful Tool: The paper effectively demonstrates the power of duality in relating different geometric problems and simplifying analysis.
  • Bridging Theory and Computation: The work highlights the necessity of combining rigorous theoretical analysis with numerical estimation to tackle problems where simple monotonicity arguments fail.

Key Results

  • Theorem 1.1:
    • For $0 < q < 1$, there are exactly two solutions (up to rotation): one constant and one positive smooth non-constant solution.
  • For $1 \le q \le 4$, there is only one solution, the constant solution.
  • Theorem 1.2: For $q > 4$, there are at least $[\sqrt{q}] - 2$ non-constant positive smooth solutions (up to rotation).
  • Theorem 1.3: Provides a corresponding classification of solutions for the $L_p$-Alexandrov problem for $p < 0$, based on the duality with the dual Minkowski problem.

Limitations and Future Work

  • Numerical Estimation for $q > 4$: While the paper proves the existence of at least $[\sqrt{q}] - 2$ solutions for $q > 4$, the exact number of solutions is not determined. The authors suggest that the numerical estimation techniques used might be further developed to solve remaining unsolved cases for other ranges of $p$ and $q$ in related problems.
  • Complexity for $q > 4$: The paper notes that for $q > 3$, the function $\Theta_q(r)$ is no longer monotonic, making the analysis significantly more challenging. The methods developed for $q \le 4$ might not directly extend to all $q > 4$.
  • Generalization to Higher Dimensions: The current work focuses on the planar case ($n=2$). Extending these results to higher dimensions would be a natural next step.
  • Non-constant $f$: The paper primarily deals with the case where the function $f$ is constant (specifically $f=1$). Investigating the problem for non-constant $f$ would be a significant extension.

Summary updated

BibTeX
@inproceedings{liu2026number,
  author = {Yannan Liu and Jiangbo Lu},
  title = {{On the number of solutions to the planar dual Minkowski problem}},
  booktitle = {Science},
  year = {2026},
  doi = {10.1007/s11425-025-2492-8},
}
Figure