Direct Method of Moving Planes

• Direct Method of Moving Planes is an integral-based reflection technique that establishes symmetry, monotonicity, rigidity, and uniqueness for nonlinear elliptic PDE solutions.
• The method leverages maximum and narrow region principles alongside integral kernel estimates to analyze nonlocal operators like the fractional Laplacian and p-Laplacian.
• Its applications span proving geometric rigidity, quantitative symmetry, and uniqueness in diverse settings including nonlocal and variable-exponent problems.

The direct method of moving planes is an integral-based reflection technique for establishing symmetry, monotonicity, rigidity, and uniqueness properties of solutions to nonlinear elliptic PDEs and variational problems. Distinguished from extension-based approaches, the direct method acts entirely on the original (often nonlocal and/or nonlinear) operator, relying on maximum (and strong maximum) principles, integral kernel estimates, and reflection-induced difference constructions. It has proven especially effective for nonlocal operators such as the fractional Laplacian, uniformly elliptic nonlocal Bellman operators, variable-exponent fractional Laplacians, logarithmic Schrödinger operators, and quasilinear (p-Laplacian) equations, as well as in geometric rigidity settings such as the Alexandrov-Serrin theorem.

1. Integral Operator Setting and Prototypical Examples

The scope of the direct method encompasses a wide range of operators and equations. For nonlocal operators of the general form

$$\mathcal{I}[u](x) = \mathrm{P.V.}\int_{\mathbb{R}^n} G(u(x)-u(y))\,K(x,y)\,dy,$$

method applicability typically requires that the kernel $K(x,y)$ be singular of order $|x-y|^{-(n+2s)}$ (with $0 < s < 1$) and $G$ be Lipschitz near $0$ (Chen et al., 2014). Central instances:

• Fractional Laplacian: $$(-\Delta)^s u(x) = C_{n,s}\,\mathrm{P.V.}\int_{\mathbb{R}^{n}} [u(x) - u(y)]|x-y|^{-n-2s}\,dy$$ (Chen et al., 2014)
• Uniformly elliptic nonlocal Bellman operator: $$I_s[u](x) = \inf_{A} \mathrm{P.V.} \int_{\mathbb{R}^n} [u(x+y) + u(x-y) - 2u(x)]|A^{-1}y|^{-n-2s}\,dy$$, where $A$ ranges over symmetric positive definite matrices with ellipticity bounds (Dai et al., 2020)
• Fractional $p(x,y)$-Laplacian: $$\left(-\Delta\right)^s_{p(\cdot)} u(x) = 2\, \mathrm{P.V.} \int_{\mathbb{R}^n} |u(x) - u(y)|^{p(x,y)-2} (u(x)-u(y))\,|x-y|^{-n-s p(x,y)}\,dy$$ (Bahrouni et al., 2024)
• Logarithmic Schrödinger operator: $(I-\Delta)^{\log}u(x)$, with Fourier symbol $\log(1+|\xi|^2)$ and singular kernel involving a Bessel function (Zhang et al., 2022)
• Elliptic $p$-Laplacian: $$-\Delta_p u = -\mathrm{div}(|\nabla u|^{p-2}\nabla u)$$, relevant for quasilinear equations and quantitative symmetry (Gatti, 17 Feb 2025)

2. Maximum Principles and Key Comparison Results

The success of the direct moving planes method hinges on strong and localized maximum principles adapted to the operator's structure.

• Strong maximum principle (anti-symmetric version): For anti-symmetric $w$ (i.e., $w(x) = -w(x^T)$ where $T$ is a hyperplane), $w \geq 0$ in one side, and $I_s[w](x_0) \leq 0$ at a vanishing point $x_0$, then $w \equiv 0$ (Dai et al., 2020, Chen et al., 2014, Bahrouni et al., 2024).
• Narrow region principle: In a thin slab (width $d$), if $w$ solves $I_s[w] - c(x)w \leq 0$ and $d$ is small, negativity is precluded due to the singularity of the kernel (Dai et al., 2020, Chen et al., 2014).
• Boundary point (Hopf-type) lemma: Adapted using scaled radial barrier functions or, for nonlocal/variable-exponent cases, differences of singular integrals, ensuring the boundary precludes "stagnation" of reflection differences (Dai et al., 2020, Bahrouni et al., 2024).
• Decay-at-infinity principle: For unbounded domains, ensures a negative minimum cannot drift to infinity under mild growth assumptions on $c(x)$ (Dai et al., 2020).

For variable exponent and p-Laplacian settings, additional mean value and convexity arguments are incorporated to handle nonlinear powers and spatially variable exponents (Bahrouni et al., 2024, Gatti, 17 Feb 2025).

3. The Moving Planes and Sliding Procedures

The core method involves constructing a one-parameter family of reflected difference functions (commonly $w_\lambda(x) = u(x^\lambda) - u(x)$, with $x^\lambda$ the reflection of $x$ across the plane $T_\lambda$) (Dai et al., 2020, Chen et al., 2014). The main steps are:

1. Initialization ("start"): Place the reflection plane $T_\lambda$ in a region where the difference function $w_\lambda \geq 0$ (e.g., far to the left or at infinity, via decay of $u$).
2. Sliding ("move"): Increase or decrease $\lambda$, maintaining $w_\lambda \geq 0$ as permitted by the maximum principles and barrier inequalities.
3. Critical position ("stop"): Identify the extremal $\lambda_0$ where $w_{\lambda_0} \geq 0$ but $w_{\lambda_0 + \epsilon}$ would violate positivity. Use strong maximum or Hopf-type principles to conclude that $w_{\lambda_0} \equiv 0$ in the moving half-space.
4. Symmetry/monotonicity: Conclude that $u$ is symmetric with respect to $T_{\lambda_0}$ and, by iteration or rotation of coordinate directions, deduce full symmetry or monotonicity (Dai et al., 2020, Chen et al., 2014, Zhang et al., 2022, Bahrouni et al., 2024).

In unbounded or epigraph domains, a "sliding" method utilizing translations is used, with corresponding difference functions $w_t(x) = u(x + t e) - u(x)$ (Dai et al., 2020).

4. Technical Integral and Kernel Estimates

Maximum principles are enabled by technical estimates quantifying the nonlocal integral contributions at critical points:

• Local contribution: At a point $x_0$ realizing the minimum of $w$, $$I_s[w](x_0) \geq -C |w(x_0)|\,(\mathrm{dist}(x_0, T))^{-2s}$$, due to the kernel's singularity (Dai et al., 2020).
• Narrow region control: In a slab of width $d$, the nonlocal integral is of size $d^{-2s}$, so a sufficiently small $d$ rules out negative minima (Dai et al., 2020, Chen et al., 2014).
• Far-field decay: Ensures that away from the reflection plane, the influence of far field is controlled by spatial decay (Dai et al., 2020).

In the variable exponent case, the comparison of kernels under reflection is handled using monotonicity and mean value inequalities in $p(x,y)$ (Bahrouni et al., 2024).

5. Applications: Symmetry, Monotonicity, Uniqueness

The method is applied to a wide range of nonlinear and nonlocal problems:

• Symmetry and monotonicity of solutions: For equations $-I_s[u] = f(x, u, \nabla u)$ in convex domains or $\mathbb{R}^n$, symmetry with respect to hyperplanes or radial monotonicity is obtained under suitable structure conditions on $f$ (Dai et al., 2020, Chen et al., 2014, Bahrouni et al., 2024, Zhang et al., 2022).
• Geometric rigidity: Sharp quantitative versions of Alexandrov's "Soap-Bubble" theorem use the direct method to establish proximity to spherical symmetry as a function of mean curvature oscillation (Ciraolo et al., 2015).
• Uniqueness and classification: In the De Giorgi setting ($-I_s[u]+u = u^p$), one-dimensionality and strict monotonicity for entire solutions are established under mild growth conditions (Dai et al., 2020).
• Non-cooperative systems: In certain coupled systems without maximum principles, integral estimates and reflection differences recover monotonicity and 1D symmetry (Aftalion et al., 2019).
• Quantitative almost-symmetry: For perturbed $p$-Laplace equations, error estimates in symmetry/monotonicity proportional to (powers of) the perturbation parameter are obtained (Gatti, 17 Feb 2025).

6. Key Innovations and Methodological Distinctions

• The direct method works entirely with the original, possibly nonlocal, operator in a given dimension, avoiding extension or auxiliary variable approaches (cf. the Caffarelli-Silvestre extension for the fractional Laplacian) (Chen et al., 2014, Dai et al., 2020).
• For variable exponent and nonlinear operators, the method relies on convexity and mean value inequalities for the power nonlinearities (e.g., $|t|^{p-2} t$), as well as monotonicity of the kernel (Bahrouni et al., 2024).
• In the absence of classical maximum principles (e.g., non-cooperative systems), reflection-induced a priori bounds and sliding energy lemmas serve as surrogates (Aftalion et al., 2019).
• The method achieves both qualitative (exact symmetry, nonexistence) and quantitative (stability/approximate symmetry with explicit rates) conclusions (Ciraolo et al., 2015, Gatti, 17 Feb 2025).

7. Extensions and General Scope

The direct method is robust under generalizations to:

• Operators with kernels comparable to $|x-y|^{-n-2s}$; nonlinearities of Lipschitz type; lower-order and spatially dependent coefficients (Chen et al., 2014, Dai et al., 2020).
• Nonlocal geometric flows (e.g., fractional mean curvature estimates in nonlocal Alexandrov theorems) (Ciraolo et al., 2015).
• Domains with complex geometry: bounded, unbounded, epigraph, or stripes, with corresponding forms of the sliding method (Dai et al., 2020).
• Problems with critical nonlinearities, e.g., $u^p$ at the critical Sobolev exponent, including use of Kelvin transforms for subcritical/categorical cases (Chen et al., 2014).

This suggests that the method forms a unifying framework for symmetry and rigidity results across a spectrum of local and nonlocal, linear and nonlinear, scalar and system-level PDEs. The central requirement remains the validity of suitable maximum/narrow region/decay principles for reflection-induced difference functions tailored to the operator and domain under consideration.