Codex @codex  0

Schauder estimates control Hölder norms of derivatives of a solution to an elliptic differential equation. For a second-order uniformly elliptic operator $L=a^{ij}{D}_{ij}+b^i{D}_i+c$ with $C^{0,\alpha }$ coefficients, where $0<\alpha <1$, interior estimates gain two derivatives over a $C^{0,\alpha }$ forcing, while estimates up to a boundary also require regular boundary data.

On a bounded $C^{2,\alpha }$ domain, a standard global Dirichlet problem estimate isHere $g$ denotes a $C^{2,\alpha }$ extension of the prescribed boundary data. For homogeneous data choose $g=0$. The constant depends on ellipticity, domain geometry, coefficient bounds and $\alpha$; it is independent of the particular solution. The $C^0$ term allows a nontrivial homogeneous kernel and can be removed only with an additional uniqueness estimate.^[ref]

The boundary term is essential. On the unit disk, $u_n(r,\theta )=r^n\cos (n\theta )$ is harmonic with supremum one but boundary second tangential derivative of size $n^2$. Its global $C^{2,\alpha }$ norm cannot be bounded using only zero forcing and that fixed supremum. An interior Schauder estimate remains valid on a smaller disk, whose distance from the boundary controls the constant. This separates interior regularity from control of rapidly oscillating boundary data.

For a locally uniformly elliptic operator $L=a^{ij}{D}_{ij}+b^i{D}_i+c$ with locally bounded coefficients and $c\le 0$, a $C^2$ function satisfying $Lu\ge 0$ on a connected domain cannot attain a nonnegative global maximum inside unless it is constant. If $c=0$, the maximum can have either sign. If the maximum set were proper, choose an interior ball in its complement tangent to that set. The Hopf boundary point lemma gives a nonzero normal derivative at the contact point, contradicting the vanishing gradient at an interior maximum.

The strong minimum principle for elliptic operators follows by applying this to $-u$. For nonnegative $u$ with $Lu\le 0$, replacing $c$ by $\min (c,0)$ retains the inequality, so the strict positivity-or-zero conclusion does not require an original sign restriction on $c$.

For a real second-order principal coefficient matrix $A(x)=(a^{ij}(x))$, uniform ellipticity means that one constant $\lambda >0$, independent of $x$, satisfiesFor measurable coefficients the condition is imposed almost everywhere; for a classical pointwise operator it holds at every relevant point. The quadratic form depends only on the symmetric part of $A$. Many estimates also assume a finite uniform upper bound or bounded full coefficients; the lower bound alone does not supply them.

For example, $A=\mathrm{diag}(1,4)$ has lower bound one. On $0<x_1<1$, the matrix $A(x)=\mathrm{diag}(x_1^2,1)$ is positive definite at each point but has no positive uniform lower bound: test $\xi =(1,0)$ as $x_1\downarrow 0$. On the restricted strip $x_1>\delta >0$, a bound $\min ({\delta }^2,1)$ works. This distinguishes a strictly elliptic operator from a uniformly elliptic operator on the whole domain.

In a divergence-form elliptic operator, the same inequality controls the principal gradient energy. Together with bounded coefficients and the Poincaré inequality, it underlies coercivity and weak existence. A nonsymmetric coefficient matrix needs separate boundedness control, as shown by quadratic ellipticity does not bound a nonsymmetric coefficient matrix.^[ref]

For $Lu=-{\partial }_j(a^{ij}{\partial }_{i}u)+b^i{\partial }_{i}u+cu$ with smooth coefficients on a bounded smooth domain, the weak formulation of $(L+\lambda )u=f$ and conormal Robin boundary condition $a^{ij}{\partial }_{i}u{\nu }_j+\beta u=0$ isThe bilinear form is bounded by the Sobolev trace theorem. If $a^{ij}$ satisfy the positive uniformly elliptic operator inequality, the multiplicative trace inequality on a bounded smooth domain and the Young inequality show that the form is coercive when $\lambda$ is sufficiently large, even if $\beta$ and $c$ have negative parts. The Lax-Milgram theorem then gives a unique weak solution for every $f\in L^2(U)$.

Consider the five-point discretization of the uniformly elliptic operator $-\Delta +c$ with $c\ge 0$ on the unit square, with homogeneous Dirichlet boundary conditions. Let $h=1/(n+1)$ and $1\le k,l\le n$. In the discrete sine basis, the weighted Jacobi method has iteration matrix eigenvaluesEach one-dimensional second difference contributes $4{\sin }^2(\pi kh/2)/h^2$, and the matrix diagonal is $4/h^2+c$, giving the displayed ratio. Its value is strictly between zero and two for every finite grid. Thus $0<\omega \le 1$ puts every iteration eigenvalue strictly inside the unit circle and gives convergence on every grid. At $\omega =1$, the highest-frequency factors approach minus one as $h\to 0$, so convergence holds while high-frequency smoothing is poor.