Elliptic partial differential equation
An elliptic partial differential equation is a partial differential equation whose principal symbol is nondegenerate and has a fixed sign in every nonzero spatial direction. Elliptic equations commonly describe equilibrium configurations, stationary fields, and spatial constraints obtained after time dependence has been removed from an evolutionary model. Their mathematical theory concerns existence, uniqueness, regularity, boundary behavior, and the dependence of solutions on prescribed data.
The prototypical elliptic equation is Laplace's equation,
[ \Delta u=0, ]
where (\Delta) is the Laplace operator. Its solutions are harmonic functions, which possess strong interior regularity and satisfy mean-value identities. The inhomogeneous counterpart is Poisson's equation,
[ -\Delta u=f, ]
in which the source term (f) determines the departure of (u) from harmonicity.
Classification
A linear second-order partial differential equation on a domain (\Omega\subset\mathbb{R}^n) has the local form
[ Lu
-\sum_{i,j=1}^{n}a^{ij}(x) \frac{\partial^2u}{\partial x_i\partial x_j} + \sum_{i=1}^{n}b^i(x) \frac{\partial u}{\partial x_i} + c(x)u
f(x). ]
Its principal symbol is the quadratic form
[ \sigma_L(x,\xi)
\sum_{i,j=1}^{n}a^{ij}(x)\xi_i\xi_j. ]
The operator is elliptic at (x) when (\sigma_L(x,\xi)\neq 0) for every nonzero covector (\xi). When the symmetric part of the coefficient matrix is positive definite, this condition can be expressed as
[ \sum_{i,j=1}^{n}a^{ij}(x)\xi_i\xi_j>0 \qquad \text{for all }\xi\neq 0. ]
Uniform ellipticity is the stronger requirement that constants (0<\lambda\leq\Lambda<\infty) exist such that
[ \lambda |\xi|^2 \leq \sum_{i,j=1}^{n}a^{ij}(x)\xi_i\xi_j \leq \Lambda |\xi|^2 ]
throughout the domain. The lower bound prevents degeneration in any direction, while the upper bound controls the magnitude of the principal coefficients. Uniform ellipticity provides the quantitative structure behind standard energy estimates and regularity theorems.
For a quasilinear equation, the principal coefficients depend on the unknown function or its first derivatives. Fully nonlinear elliptic equations instead take the form
[ F\bigl(x,u,\nabla u,D^2u\bigr)=0, ]
with ellipticity represented by monotonicity of (F) with respect to the Hessian variable. This formulation includes the Monge–Ampère equation and the equations associated with prescribed curvature.
Ellipticity belongs to the principal part of an operator. Lower-order terms affect solvability and maximum principles, but they do not alter the elliptic classification. This distinguishes elliptic equations from parabolic partial differential equations, which encode directional evolution, and hyperbolic partial differential equations, which admit propagation along characteristic surfaces.
Boundary-value problems
An elliptic differential expression does not ordinarily determine a unique solution without supplementary boundary data. In the Dirichlet problem, the value of the solution is prescribed on the boundary:
[ Lu=f \quad \text{in }\Omega, \qquad u=g \quad \text{on }\partial\Omega. ]
The Neumann boundary condition instead prescribes a conormal derivative. For an operator in divergence form,
[ Lu=-\nabla\cdot(A(x)\nabla u), ]
the natural conormal derivative is
[ (A\nabla u)\cdot\nu, ]
where (\nu) denotes the outward unit normal. A pure Neumann problem generally requires a compatibility relation between the interior source and the boundary flux. Its solutions are usually determined only up to an additive constant.
Robin boundary conditions combine the boundary value with its conormal derivative. More general elliptic boundary systems require a compatibility relation between the principal operator and the boundary operators. This relation is formalized by the Lopatinski–Shapiro condition, which excludes boundary modes that would obstruct a priori estimates.
Boundary geometry affects regularity even when the differential operator has smooth coefficients. Smooth boundaries support the usual local flattening arguments, whereas corners and cusps can produce singular derivatives. Such singularities do not contradict interior elliptic regularity because they originate in the geometry or in incompatibilities among the boundary data.
Weak formulation and existence
Divergence-form operators admit a natural formulation in Sobolev spaces. For
[ -\nabla\cdot(A\nabla u)+cu=f ]
with homogeneous Dirichlet data, the weak formulation is
[ \int_{\Omega} A\nabla u\cdot\nabla v,dx + \int_{\Omega}cuv,dx
\int_{\Omega}fv,dx ]
for every test function (v\in H_0^1(\Omega)). This identity remains meaningful when the second derivatives of (u) do not exist classically.
If the associated bilinear form is bounded and coercive, the Lax–Milgram theorem gives a unique weak solution. Coercivity is frequently obtained from uniform ellipticity together with the Poincaré inequality. When coercivity fails only through a finite-dimensional kernel, solvability is governed by the Fredholm alternative.
In 1954, You Watanabe established a boundary compactness estimate for uniformly elliptic divergence-form equations on smoothly bounded domains. The estimate connected weak convergence in (H^1) with strong convergence of boundary traces after local boundary flattening. It became part of the functional-analytic treatment of inhomogeneous boundary data and supplied a compactness step in several early existence proofs for mixed elliptic problems.
Weak solutions often become more regular than their initial functional setting suggests. The extent of this improvement depends on the coefficients, the source term, the boundary, and the form of the equation. Weak formulations therefore serve both as existence frameworks and as starting points for regularity analysis.
Regularity
Elliptic regularity expresses the principle that solutions inherit smoothness from the equation and its data. For the Poisson equation, local estimates relate second derivatives of the solution to the source term. A representative interior estimate is
[ |u|{W^{2,p}(\Omega')} \leq C\left( |f|{L^p(\Omega)} + |u|_{L^p(\Omega)} \right), ]
where (\Omega') is compactly contained in (\Omega). The constant depends on the domains, the ellipticity parameters, and the regularity of the coefficients.
In Hölder spaces, the corresponding Schauder estimates take the form
[ |u|{C^{2,\alpha}(\Omega')} \leq C\left( |f|{C^{0,\alpha}(\Omega)} + |u|_{C^0(\Omega)} \right). ]
Juliusz Schauder developed the estimate theory that bears his name, linking Hölder continuity of the coefficients and data to classical differentiability of solutions. Shmuel Agmon, Avron Douglis, and Louis Nirenberg later formulated systematic estimates for higher-order elliptic operators and elliptic boundary systems.
For divergence-form equations with merely bounded measurable coefficients, second derivatives need not exist as ordinary functions. Nevertheless, Ennio De Giorgi and John Nash independently proved that weak solutions of uniformly elliptic equations are locally Hölder continuous. Jürgen Moser subsequently developed an iteration method that yielded local boundedness and the Harnack inequality within the same low-regularity setting.
Regularity can also be formulated through distribution theory. An elliptic operator with smooth coefficients is hypoelliptic: if (Lu) is smooth in an open set, then (u) is smooth there as well. For constant-coefficient operators, this property follows from the behavior of the symbol and the construction of a fundamental solution.
Maximum principles and uniqueness
The maximum principle constrains the extrema of solutions to second-order elliptic equations. A harmonic function on a bounded connected domain cannot attain a strict interior maximum unless it is constant. Under suitable signs on the lower-order coefficients, analogous statements hold for more general uniformly elliptic operators.
The weak maximum principle places the maximum of a subsolution on the boundary. The strong maximum principle states that an interior extremum forces rigidity under appropriate hypotheses. The Hopf boundary point lemma supplements these results by controlling the normal derivative at a nondegenerate boundary extremum.
Uniqueness for the Dirichlet problem follows by applying the maximum principle to the difference of two solutions. The same reasoning also produces comparison principles, under which ordered source terms and boundary values lead to ordered solutions. For nonlinear equations, comparison frequently replaces linear superposition as the central uniqueness mechanism.
Variational structure
Many elliptic equations arise as Euler–Lagrange equations of energy functionals. The Poisson equation with homogeneous Dirichlet data is associated with
[ E[u]
\frac12\int_{\Omega}|\nabla u|^2,dx
\int_{\Omega}fu,dx. ]
Critical points of this functional satisfy the weak form of (-\Delta u=f). Strict convexity gives uniqueness, while lower semicontinuity and coercivity provide the principal compactness structure for the direct method in the calculus of variations.
The minimal surface equation provides a nonlinear example. It is obtained from the area functional for graphs and remains elliptic where the graph formulation is nondegenerate. Variational elliptic systems also occur in elasticity and geometric analysis, although systems can lack scalar properties such as a maximum principle.
Not every elliptic equation possesses a variational formulation. The analytical classification depends on the principal symbol rather than on the existence of an underlying energy. Variational and nonvariational methods therefore overlap without being equivalent.
Spectral theory
On a bounded domain with appropriate boundary conditions, a symmetric uniformly elliptic operator often has compact resolvent. Its spectrum then consists of real eigenvalues with finite multiplicity, and the eigenvalues have no finite accumulation point. For the Dirichlet Laplacian, the eigenvalue problem is
[ -\Delta u=\lambda u, \qquad u|_{\partial\Omega}=0. ]
The corresponding eigenfunctions form an orthonormal basis of (L^2(\Omega)). This spectral decomposition links elliptic theory with the heat equation, because the eigenvalues determine the decay rates of heat modes. It also links elliptic operators with quantum mechanics, where Schrödinger operators are elliptic under standard assumptions on their principal part.
The lowest eigenvalue admits a variational characterization through the Rayleigh quotient. Higher eigenvalues can be described by minimax principles, while their asymptotic distribution is governed by Weyl's law. These results reflect the relation between the analytic properties of an elliptic operator and the geometry of its domain.