Rellich–Kondrachov-like Compact Embeddings

• Rellich–Kondrachov-like compact embedding theorem is a framework extending classical Sobolev compactness to settings with generalized derivatives, non-Euclidean geometries, and variable critical exponents.
• It employs techniques such as local Poincaré inequalities, Rothschild–Stein lifting, and spectral analysis to address subelliptic, fractal, and nonlocal contexts.
• These compactness results are crucial for advancing geometric analysis, variational problems, and finite element methods by ensuring strong convergence in function spaces.

A Rellich–Kondrachov-like compact embedding theorem refers to a family of results asserting that, for certain function spaces equipped with a generalized derivative (or structure), the inclusion into an $L^q$ space is not only continuous but compact under appropriate conditions on the underlying geometry, topology, or operator. The “like” modifier signals that the setting differs from the classical Euclidean Sobolev space scenario, and may include subelliptic operators, symmetry constraints, structured singularities, or generalizations to metric, group, or nonlocal contexts.

1. Classical and Abstract Framework

The Rellich–Kondrachov theorem asserts that for a bounded Lipschitz domain $\Omega\subset\mathbb{R}^n$ and $1 \leq p < n$, the Sobolev embedding

$$W_0^{1,p}(\Omega) \hookrightarrow L^q(\Omega), \quad 1 \leq q < p^* := \frac{np}{n-p}$$

is compact. Chua, Rodney, and Wheeden established an abstract generalization based on measure theory and local Poincaré-type inequalities, applying to degenerate and weighted Sobolev spaces, and even to quasimetric spaces lacking a gradient (Chua et al., 2011). In this formulation, compact embedding is a consequence of local oscillation and Poincaré estimates, rather than the explicit metric or differential structure.

2. Subelliptic, Degenerate, and Non-Euclidean Contexts

Major advances concern Sobolev spaces associated with vector fields satisfying Hörmander’s bracket-generating (subelliptic) condition. For a smooth vector-field system $X = (X_1,\dots,X_m)$ on $U\subset\mathbb{R}^n$, and with the generalized Métivier index $\nu$ (supremum of pointwise homogeneous dimensions), the critical Sobolev exponent adapts as $p^* := p \nu/(\nu - k p)$. A sharp result is:

• Rellich–Kondrachov Compact Embedding for Hörmander Systems: For $k\in\mathbb{N}$, $p \geq 1$, and $k p < \nu$, the embedding

$W_{X,0}^{k,p}(\Omega) \hookrightarrow L^q(\Omega)$

is compact for all $1 \leq q < p^*$, continuous for $q = p^*$, independent of the smoothness of $\partial\Omega$ (Chen et al., 2024). The proof leverages Rothschild–Stein lifting, saturation methods, and fine representation formulas exploiting the Carnot–Carathéodory geometry.

Chen–Chen–Yuan (Chen et al., 2022) prove analogous results for subelliptic Dirichlet problems, showing that the embedding, controlled by the Métivier index $v$, is critical for the analysis of minimax variational principles in degenerate, non-Métivier settings.

3. Group, Symmetry, and Nonlocal Extensions

On compact locally compact abelian groups (LCA groups), the spectral formulation allows one to define Sobolev spaces $H^s(G)$ via the Fourier transform, with weight functions on the dual group $G^\wedge$. The compactness of the embedding $H^s(G) \hookrightarrow L^2(G)$ is equivalent to the finiteness of sets $\{\chi \in G^\wedge: y(\chi) \leq R\}$ for all $R>0$, tying together harmonic analysis and compactness through spectral decay (Ccajma et al., 2022, Górka et al., 2012).

In Riemannian geometry, symmetry-reduced subspaces (such as $G$-invariant or foliation-invariant functions) admit compact embeddings with larger critical exponents due to the lower effective dimension. For a singular Riemannian foliation $\mathcal{F}$ on a compact manifold $M$, the $\mathcal{F}$-invariant Sobolev space $W^{1,p}_\mathcal{F}(M)$ embeds compactly into $L^q(M)$ for the sharp range $1 \leq q < \frac{n p}{n-p}$, and, when $p \geq n-d_*$ (with $d_*$ minimal leaf dimension), into $L^q$ for all $q<\infty$ (Alexandrino et al., 2023).

4. Metric, Fractal, and Piecewise Settings

Analysis extends to metric spaces via Reshetnyak–Sobolev maps and ultra-completion methods. For any bounded sequence in $W^{1,p}(\Omega,X)$ for a complete metric space $X$, there exists a limit in a possibly larger metric space $Z$ with convergence in $L^q$ for all $q < p^*$, a construction invaluable in geometric analysis and modern geometric measure theory (Guo et al., 2017).

In domains with fractal or irregular boundaries, compact embedding is governed by $d$-set and Ahlfors regularity, and is valid for any Sobolev admissible domain with the associated $D_s$–$L_d$ measure-doubling and lower-mass conditions. The trace operator is also compact in this framework (Rozanova-Pierrat, 2020).

For piecewise $H^1$ spaces over shape-regular triangulations (no quasi-uniformity), the embedding

$$H^1(\Omega,\mathcal{T}_h) \hookrightarrow L^2(\Omega)$$

is compact, crucial for finite element analysis and the study of discontinuous Galerkin methods (Zhang, 2013).

5. Fractional and Nonlocal Cases

Fractional analogues are constructed through spectral theory or Caffarelli–Silvestre-type extension techniques. When $X\hookrightarrow Y$ are Hilbert spaces with a compact dense embedding and $A:X \to X'$ is elliptic and symmetric, then for $0 < s < 1$, the fractional domain $H^s = D(A^{s/2})$ is compactly embedded in $Y$, with sharp trace representations in weighted Dirichlet spaces (Bueno et al., 14 Apr 2025). On hyperbolic and non-compact symmetric spaces, compactness for radial (or symmetry-reduced) fractional Sobolev spaces persists and underpins the theory for nonlinear elliptic and parabolic equations (Bruno et al., 15 Sep 2025).

6. Nonlinear and Constrained Embeddings

In Sobolev-type spaces with nonlinear constraints (e.g., vector fields $A\in L^3(\mathbb{R}^3;\mathbb{R}^3)$ with $\nabla\wedge A \in L^{3/2}$ and $\mathrm{div}(|A|A)=0$), “nonlinear Rellich–Kondrachov” compactness holds: weak convergence of $\nabla\wedge A_n$ plus the constraint yields strong convergence in $L^q$ for $q<3$ (Frank et al., 2021). This mechanism is vital for the existence of extremal vector fields and optimal solutions in magnetic Dirac and related variational problems, generalizing beyond the classical linear regime.

7. Unified Perspective and Criticality

Across geometric, analytic, and algebraic generalizations, the unifying insight is that compactness of Sobolev-type embeddings hinges on effective dimension, symmetry reduction, local geometry (e.g., Carnot–Carathéodory, metric, foliation), and analytic capacity to preclude “mass escape” at infinity. The critical exponents in each context are dictated by spectral or geometric invariants (Métivier index, homogeneous dimension, quotient dimension). In degenerate or symmetry-reduced settings, the critical exponents increase, reflecting enhanced compactness due to geometric or analytic constraints, a phenomenon also captured by the rise in the Strauss exponent for radial embeddings and group-invariant counterparts (Skrzypczak et al., 2020). This sharpness is typically optimal and cannot be improved generically, as demonstrated by spectral and cocompactness criteria across the literature.