Abreu Equation: Geometry & Optimization

• Abreu Equation is a fourth-order, fully nonlinear elliptic PDE that governs scalar curvature in toric Kähler geometry and underpins variational problems.
• It generalizes classical scalar curvature equations with weighted formulations and strict boundary conditions defined by Guillemin potentials.
• Advanced analytical methods, including Legendre transforms and determinant bounds, ensure existence, regularity, and stability in both smooth and singular settings.

The Abreu equation is a fourth-order, fully nonlinear elliptic PDE, centrally important in the study of symplectic and Kähler geometry, convex variational calculus with convexity constraints, and affine differential geometry. Its structure generalizes the scalar curvature equation on toric Kähler manifolds to a broader array of geometric and analytic contexts, including both smooth and singular settings. The equation also provides a rigorous PDE framework for the approximation of minimizers in constrained convex optimization problems arising, for example, in economics and elasticity.

1. Classical and Generalized Forms

In its canonical form, the Abreu equation for a convex function $u: \Omega \subset \mathbb{R}^n \to \mathbb{R}$ reads: $$-\sum_{i,j=1}^n \frac{\partial^2 u^{ij}}{\partial x_i \partial x_j} = f(x)$$ where $(u^{ij}) := (\partial_{x_i x_j}^2 u)^{-1}$ is the matrix inverse of the Hessian of $u$. Equivalently, introduce the cofactor matrix

$U^{ij} = (\det D^2 u)\, (D^2 u)^{-1}_{ij}$

and set $w = (\det D^2 u)^{-1}$. The equation is then expressed as $U^{ij} w_{ij} = f(x)$, which is the standard Abreu operator form encountered in Kähler geometry and the theory of extremal metrics on toric manifolds (Chen et al., 2010, Chen et al., 2010).

In the context of homogeneous toric bundles or fibrations, the generalized Abreu equation incorporates a positive weight function $D(\xi)$ and becomes: $$\frac{1}{D(\xi)} \sum_{i,j=1}^n \frac{\partial^2}{\partial \xi_i \partial \xi_j} \bigl( D(\xi) u^{ij}(\xi) \bigr) = -A(\xi)$$ with boundary conditions determined by the Guillemin potential, encoding the polytope geometry and singularities at the facets (Chen et al., 2016, Li et al., 2016).

2. Geometric and Variational Origins

Toric Kähler Geometry

On compact toric Kähler manifolds, the Abreu equation governs the scalar curvature $S(u)$ of torus-invariant Kähler metrics, expressed entirely in terms of the convex symplectic potential $$-\sum_{i,j=1}^n \frac{\partial^2 u^{ij}}{\partial x_i \partial x_j} = f(x)$$0 on the moment polytope $$-\sum_{i,j=1}^n \frac{\partial^2 u^{ij}}{\partial x_i \partial x_j} = f(x)$$1: $$-\sum_{i,j=1}^n \frac{\partial^2 u^{ij}}{\partial x_i \partial x_j} = f(x)$$2 This scalar curvature prescription connects the existence and regularity of extremal and constant scalar curvature Kähler (cscK) metrics to nonlinear PDE analysis on convex polytopes (Chen et al., 2010, Chen et al., 2010, Chen et al., 2010). The boundary behavior is governed by the Guillemin function, enforcing logarithmic singularities at each facet.

Convex Variational Problems

The Abreu equation, and more generally its singular variants, emerge as Euler–Lagrange equations for constrained variational problems: $$-\sum_{i,j=1}^n \frac{\partial^2 u^{ij}}{\partial x_i \partial x_j} = f(x)$$3 where $$-\sum_{i,j=1}^n \frac{\partial^2 u^{ij}}{\partial x_i \partial x_j} = f(x)$$4 is convex and satisfies Dirichlet-type conditions. To enforce strict convexity, a penalization such as $$-\sum_{i,j=1}^n \frac{\partial^2 u^{ij}}{\partial x_i \partial x_j} = f(x)$$5 is introduced. As $$-\sum_{i,j=1}^n \frac{\partial^2 u^{ij}}{\partial x_i \partial x_j} = f(x)$$6, the solutions to the associated penalized Abreu-type equations converge to minimizers of the original convexity-constrained functional. This strategy underpins recent analysis in high-dimensional convex variational calculus, optimal transport, and applications such as the two-dimensional Rochet–Choné model in economics (Le, 2018, Le, 2022, Kim, 10 Apr 2025).

3. Singular Abreu Equations and Second Boundary Value Problems

Singular Abreu equations are those for which the right-hand side is a highly nonlinear or measure-valued operator in $$-\sum_{i,j=1}^n \frac{\partial^2 u^{ij}}{\partial x_i \partial x_j} = f(x)$$7 or its derivatives, exemplifying cases where gradient or Hessian terms possess singular or quasilinear structures. The standard form is (Le et al., 2020, Le, 2018, Kim et al., 2022): $$-\sum_{i,j=1}^n \frac{\partial^2 u^{ij}}{\partial x_i \partial x_j} = f(x)$$8 where $$-\sum_{i,j=1}^n \frac{\partial^2 u^{ij}}{\partial x_i \partial x_j} = f(x)$$9, $(u^{ij}) := (\partial_{x_i x_j}^2 u)^{-1}$0 is a regularization, and the system is closed by setting $(u^{ij}) := (\partial_{x_i x_j}^2 u)^{-1}$1.

The second boundary value problem for the Abreu equation prescribes both the Dirichlet data for $(u^{ij}) := (\partial_{x_i x_j}^2 u)^{-1}$2 and boundary values for $(u^{ij}) := (\partial_{x_i x_j}^2 u)^{-1}$3 (i.e., the Monge–Ampère measure) on $(u^{ij}) := (\partial_{x_i x_j}^2 u)^{-1}$4. This is essential for variational approximations and regularity theory, as standard Dirichlet boundary conditions are insufficient to control ellipticity at the boundary in fourth-order, non-divergence elliptic PDEs (Chen et al., 2010, Ma et al., 2024, Kim et al., 2022).

4. Analytical Methods and Regularity Theory

The advanced analysis of Abreu-type equations couples convex-analytic techniques with potential theory and nonlinear elliptic regularity. Key steps include:

• Legendre Transform Techniques: Switching between $(u^{ij}) := (\partial_{x_i x_j}^2 u)^{-1}$5-coordinates (variable of $(u^{ij}) := (\partial_{x_i x_j}^2 u)^{-1}$6) and $(u^{ij}) := (\partial_{x_i x_j}^2 u)^{-1}$7 variables to exploit dual PDEs that may admit superior maximum principles or determinant bounds (Le et al., 2020, Li et al., 2016).
• Determinant Bounds: Establishing lower and upper control for $(u^{ij}) := (\partial_{x_i x_j}^2 u)^{-1}$8 via barrier functions and duality arguments is foundational for ensuring strict ellipticity. The presence of weights or singularities requires careful adaptation of these constructions (Li et al., 2016, Le et al., 2020).
• Interior Hölder/Regularity Estimates: Once determinant bounds are available, one applies Caffarelli–Gutiérrez theory for the linearized (Monge–Ampère) operators and Schauder estimates to bootstrap $(u^{ij}) := (\partial_{x_i x_j}^2 u)^{-1}$9 to full interior smoothness (Chen et al., 2010, Le et al., 2020).
• Green’s Function and Lorentz Space Methods: Recent progress leverages critical Lorentz-space bounds for the Green’s function of the linearized Monge–Ampère operator, enabling fine regularity results even for PDEs with right-hand sides given by singular measures or divergence of bounded vector fields (Gu et al., 19 Nov 2025, Cui et al., 5 Nov 2025).
• Twisted Harnack Inequalities: For singular equations with nontrivial drift or subdomain-located singularities, twisted Harnack inequalities—adapting classical Moser–De Giorgi iteration to measure-valued, nonuniformly elliptic settings—are used to obtain local Harnack and Hölder continuity results (Le, 2022).

5. Existence, Stability, and Approximation Schemes

Existence and regularity for the Abreu equation and its generalizations are established under geometric and analytic stability conditions.

• Uniform K-stability: The Mabuchi and Futaki functionals define algebraic stability notions (inspired by geometric invariant theory) that are necessary and sufficient for the existence of solutions in the toric and homogeneous cases. Uniform K-stability yields uniform $u$0 bounds, crucial for the continuity method and compactness (Chen et al., 2016, Li et al., 2016).
• Continuity Method and Degree Theory: Solutions are constructed by deformation from known model metrics (e.g., the Guillemin potential) using a parameterized path and bootstrapping a priori estimates at each step (Chen et al., 2010, Chen et al., 2016).
• Approximation by Singular Abreu Equations: For convex variational problems, the minimizer under a convexity constraint is recovered as the uniform limit of penalized (regularized) singular Abreu equations as the penalization parameter tends to zero. This approach is rigorously developed even in the absence of uniform convexity of cost functions (Le, 2018, Le, 2019, Le, 2022, Kim, 10 Apr 2025).

6. Extensions, Explicit Solutions, and Applications

The analytic framework for the Abreu equation has been extended to multiple domains:

• Affine Differential Geometry: The Abreu equation (with various choices of powers of the determinant) describes geometric flows and curvature prescription problems for affine hypersurfaces, such as affine maximal and constant curvature graphs (Sun et al., 15 Jan 2025).
• Noncompact and Complete Structures: Explicit complete, non-quadratic solutions in affine geometry have been characterized and classified, and their geometric implications for completeness and curvature negativity detailed (Sun et al., 15 Jan 2025).
• Low Dimensions and Special Cases: In one dimension, singular Abreu equations reduce to second-order ODEs with distinctly different solvability and regularity phenomena compared to higher dimensions (Kim, 2024).
• Economic Models: The Rochet–Choné mechanism design problem and related optimal transport screening models use singular Abreu-type equations to approximate profit-maximizers under convexity constraints (Le et al., 2020, Le, 2022).

7. Open Problems and Ongoing Directions

Despite substantial advances, several challenging questions persist:

• Global Boundary Regularity: General global boundary regularity for the (generalized or singular) Abreu equation remains open—partial results exist only in special geometric settings or under further structural assumptions on the data (Chen et al., 2010, Li et al., 2016).
• Higher-Dimensional Singular Problems: For singular Abreu equations in dimensions $u$1, solvability requires either smallness, symmetry, or additional regularity hypotheses on source terms or cost functionals. Removing these restrictions is an active topic (Le, 2019, Kim et al., 2022).
• Uniqueness and Flow Problems: Issues of uniqueness for strongly singular or degenerate equations, and the dynamics of Abreu-type flows with geometric or analytic singularities, remain largely unsettled (Li et al., 2016, Gu et al., 19 Nov 2025).
• Approximation Frameworks for General Lagrangians: New schemes allow for general superquadratic Lagrangians and weak lower-order structure, but optimal penalization and passage to the limit for nonsmooth data pose subtle questions (Kim, 10 Apr 2025).

The Abreu equation underpins a rich interplay between nonlinear PDE, convex geometry, algebraic stability, and applied convex optimization. Ongoing work continues to expand its analytic and geometric horizons, increasingly extending from smooth to singular environments and from classical geometric analysis to practical economic and physical modeling.