Linearize or Discretize First? Error Estimates under Mesh-Uniform Regularity and Exponential Stability for the KdV Equation with Boundary Dissipation
Kangsheng Liu
math.AP
Oct 29, 2026 · v2
math.NA math.OC
TL;DR
Theorems 4.1–4.4 on the ordering error were checked in Lean 4 under certain premises, using AI tools following an author-designed procedure.
Abstract
We study the Korteweg-de Vries equation on a bounded interval with boundary dissipation, and its spatial semidiscretization by a weighted cubic Hermite Galerkin method. A small nonlinear wave can be approximated by a linear discrete simulation in two orders: linearize first and then discretize, or discretize first and then linearize. Both orders lead to the same total error. They divide this error differently between linearization and discretization. The difference of the two divisions is what we call the ordering error. We show that it is quadratic in the size of the initial datum and carries a positive power of the mesh size, for all time. It can be used to assess the effect of a coarse-mesh correction for nonlinear problems. For the continuous problem, the solution map is analytic, and the semigroup of the linearized equation has Crandall-Pazy regularity with the exact index 3/2. For the weighted scheme, the discrete semigroups of the linearized equation inherit exponential stability, Kato smoothing and Crandall-Pazy regularity, with bounds uniform in the mesh size. The nonlinear solution maps are analytic, also uniformly in the mesh size. These properties give error estimates in which the error and the datum are measured in the same norm: in the energy norm, in $H_0^1$ and in $H^2$. On every interval $[t_0,\infty)$, $t_0>0$, same-norm convergence holds uniformly on balls at algebraic rates. The ordering error is of order $h^{1/6}$ times the square of the energy norm of the datum, uniformly for all $t\ge0$.
Problem
A small nonlinear KdV wave on a bounded interval with boundary dissipation can be approximated by a linear discrete model, either by linearizing first or by discretizing first. The question is how the 'ordering error', the difference between the two error splittings, behaves in the datum size and the mesh size, and over what time range.
Approach
The KdV equation is semidiscretized in space by a weighted cubic Hermite Galerkin method in the weighted space L^2((0,L),(L+x)dx). The author proves exponential stability, Kato smoothing and Crandall–Pazy regularity (exact index 3/2) for the continuous and discrete linearized semigroups, with bounds uniform in the mesh size. The author also proves that the nonlinear solution maps are analytic uniformly in the mesh size. These properties give same-norm error estimates in the energy norm, H_0^1 and H^2.
Results
The ordering error is bounded by C h^{1/6} times the squared energy norm of the datum, uniformly for all t≥0 and with exponential decay; at positive times it is of order h^{4/15}ε^2 in H^2. Same-norm convergence holds uniformly on balls at algebraic rates on every interval [t_0,∞). The main Theorems 4.1–4.4 were checked in Lean 4 under certain premises.