Weighted, Anisotropic & Hybrid Index Sets

• Weighted, anisotropic, and hybrid index sets are frameworks that characterize function space regularity near singularities to optimize PDE discretizations.
• They establish rigorous weighted Sobolev and analytic spaces by incorporating local geometric singularities, directional derivatives, and anisotropic scaling.
• These index sets guide optimal mesh design and error estimation in finite element and hp-methods, ensuring enhanced convergence rates on complex domains.

Weighted, anisotropic, and hybrid-regularity index sets constitute a foundational framework for the analysis of singularities and optimal discretizations in partial differential equations (PDEs) on non-smooth domains. These structures rigorously encode the interplay between localized geometric singularities and the directional regularity inherited by solutions. They underpin the design of optimal finite element and $hp$-methods, precise regularity results in function spaces, and the analysis of elliptic and parabolic equations with mixed or anisotropic scaling.

1. Weighted Sobolev and Analytic Spaces: Foundations

Let $\Omega\subset\mathbb{R}^3$ be a bounded polyhedral domain with sets of vertices ($V$), edges ($E$), and faces ($F$). The canonical distances to singular strata are

$$r_V(x) = \mathrm{dist}(x, \{\textrm{vertices}\}),\quad r_E(x) = \mathrm{dist}(x, \{\textrm{open edges}\}),\quad r_F(x) = \mathrm{dist}(x, \{\textrm{faces}\}).$$

Let $\beta = (\beta_V, \beta_E, \beta_F)$ be multi-exponents, typically constrained by the local geometry; e.g., for each vertex $V$ with interior angle $\omega_V$, $0\leq \beta_V<\pi/\omega_V$; analogously for each edge $\Omega\subset\mathbb{R}^3$0, $\Omega\subset\mathbb{R}^3$1 for dihedral angle $\Omega\subset\mathbb{R}^3$2.

The isotropically weighted Sobolev space is

$\Omega\subset\mathbb{R}^3$3

equipped with the norm

$\Omega\subset\mathbb{R}^3$4

This weighted structure naturally defines the domain of the Laplace operator on non-smooth domains and precisely describes regularity loss at geometric singularities (Bacuta et al., 2012, Costabel et al., 2010).

In analytic settings, analogous factorially-weighted seminorms and norms are constructed, e.g.,

$\Omega\subset\mathbb{R}^3$5

encoding analytic regularity relative to edges and corners (Costabel et al., 2010).

2. Anisotropic Regularity and Multi-index Sets

Weighted regularity can be further refined by exploiting directional regularity. Distinguish mutually orthogonal differential operators

$\Omega\subset\mathbb{R}^3$6

spanning, respectively, the edge direction, the face-transverse (but in-plane), and the fully transverse directions. A multi-index $\Omega\subset\mathbb{R}^3$7 allows arbitrary derivatives to be indexed by the number taken in each direction.

Anisotropic regularity results state: for $\Omega\subset\mathbb{R}^3$8 and homogeneous boundary data, the solution $\Omega\subset\mathbb{R}^3$9 to $V$0 on $V$1 admits

$V$2

where the index set $V$3 encodes directional regularity constraints, e.g.,

$V$4

with the parameters determined by the domain geometry (Bacuta et al., 2012). The effect is that higher-order derivatives along edges may be controlled even if global Sobolev regularity is scarce.

In analytic regularity, this is mirrored by introducing weights (e.g., $V$5) and splitting multi-indices into edge-parallel and edge-transverse parts (Costabel et al., 2010).

In frequency space, as for vortex sheet problems, weighted anisotropic Sobolev spaces such as $V$6 are defined with weights vanishing in singular directions, thus providing one-derivative gain away from critical rays and only tangential regularity otherwise. The index set in this setting is determined by the vanishing properties of the weight function in frequency variables, corresponding to anisotropic regularity dictated by the PDE symbol (Secchi, 2020).

3. Hybrid Index Sets and Flagged Spaces

Hybrid index sets unite isotropic and anisotropic viewpoints. The "hybrid" index set

$V$7

enables the enforcement of global $V$8 control while exploiting additional regularity along principal directions (Bacuta et al., 2012).

In analytic settings, hybrid or "flagged" spaces $V$9 require derivatives to satisfy different weights depending on the membership of a geometric singularity in subsets $E$0 (special Dirichlet corners) and $E$1 (edges with reduced weight). This construction interpolates between fully anisotropic and isotropic regularity regimes, matching the specific boundary condition and variational structure present in the problem (Costabel et al., 2010).

Hybrid index sets also appear in harmonic analysis and maximal regularity theory, as in weighted anisotropic mixed-norm function spaces. For example, in Lizorkin-Triebel spaces $E$2, intersection representations across variable blocks partition the anisotropic scaling and integrability requirements, generalizing the Fubini property and enabling sharp boundary data descriptions in maximal regularity problems (Lindemulder, 2019).

4. Mesh Design, Discretization, and hp-FEM

The structure of weighted and anisotropic index sets directly informs optimal mesh refinement strategies for finite element and $E$3 methods on polyhedral and polygonal domains. For near-singularity mesh refinement:

• Near an edge $E$4, set the mesh size $E$5, where $E$6 is the highest controlled directional derivative in the edge direction.
• Near a face $E$7, analogous relations $E$8 are used, with elements thin in the transverse direction.

This targeted grading ensures quasi-optimal finite element convergence rates: $E$9 or equivalently in terms of degrees of freedom $F$0, $F$1 for polynomial degree $F$2 (Bacuta et al., 2012).

In the $F$3-FEM context, the analytic regularity classes $F$4 are linked, via the index set correspondence, with exponential convergence rates provided that the polynomial degree $F$5 is chosen to match the local anisotropic weight $F$6, and mesh grading $F$7 compensates for singularity strength precisely (Costabel et al., 2010).

5. Index Set Structures in Function Space Theory

Weighted, anisotropic, and hybrid index sets have a parallel in harmonic analysis, especially in the context of weighted mixed-norm Besov and Lizorkin–Triebel spaces: $F$8 where $F$9 encodes the anisotropic dilation, $$r_V(x) = \mathrm{dist}(x, \{\textrm{vertices}\}),\quad r_E(x) = \mathrm{dist}(x, \{\textrm{open edges}\}),\quad r_F(x) = \mathrm{dist}(x, \{\textrm{faces}\}).$$0 is a product weight across variable blocks, and $$r_V(x) = \mathrm{dist}(x, \{\textrm{vertices}\}),\quad r_E(x) = \mathrm{dist}(x, \{\textrm{open edges}\}),\quad r_F(x) = \mathrm{dist}(x, \{\textrm{faces}\}).$$1 specifies mixed $$r_V(x) = \mathrm{dist}(x, \{\textrm{vertices}\}),\quad r_E(x) = \mathrm{dist}(x, \{\textrm{open edges}\}),\quad r_F(x) = \mathrm{dist}(x, \{\textrm{faces}\}).$$2-integrability. These spaces admit intersection representations, e.g.,

$$r_V(x) = \mathrm{dist}(x, \{\textrm{vertices}\}),\quad r_E(x) = \mathrm{dist}(x, \{\textrm{open edges}\}),\quad r_F(x) = \mathrm{dist}(x, \{\textrm{faces}\}).$$3

for any partition $$r_V(x) = \mathrm{dist}(x, \{\textrm{vertices}\}),\quad r_E(x) = \mathrm{dist}(x, \{\textrm{open edges}\}),\quad r_F(x) = \mathrm{dist}(x, \{\textrm{faces}\}).$$4 of variables, with full equivalence of norms (Lindemulder, 2019). This result refines the classical Fubini property, showing that the intersection holds even for distinct integrability exponents and general anisotropic scaling.

This structure is crucial in the analysis of parabolic and boundary value problems with mixed space-time regularity, where sharp data spaces are characterized as intersections of anisotropic function spaces tailored by the underlying index sets.

6. Synthesis and Applications

The combined apparatus of weighted, anisotropic, and hybrid index sets achieves:

• Rigorous description and control of singular behaviors near geometric singularities in domains.
• Precise regularity estimates (both Sobolev and analytic) necessary for optimal discretization and error estimation in finite element and $$r_V(x) = \mathrm{dist}(x, \{\textrm{vertices}\}),\quad r_E(x) = \mathrm{dist}(x, \{\textrm{open edges}\}),\quad r_F(x) = \mathrm{dist}(x, \{\textrm{faces}\}).$$5 schemes.
• Flexible and sharp modeling of regularity in anisotropic settings, aligning higher regularity in favorable directions with weaker control in singular directions.
• Direct application to advanced harmonic analysis, intersection theorems, and boundary data spaces for PDEs on non-smooth domains.
• Optimal mesh design and degree distribution in $$r_V(x) = \mathrm{dist}(x, \{\textrm{vertices}\}),\quad r_E(x) = \mathrm{dist}(x, \{\textrm{open edges}\}),\quad r_F(x) = \mathrm{dist}(x, \{\textrm{faces}\}).$$6-FEM, ensuring exponential or quasi-optimal convergence rates matched to solution regularity via the index set formalism.

The articulation of these index sets provides a unified framework for understanding, quantifying, and exploiting the complexity of regularity in PDEs and the function spaces governing their solutions (Bacuta et al., 2012, Costabel et al., 2010, Lindemulder, 2019, Secchi, 2020).