AI & ML
Conservative Finite Element Methods for Kirchhoff-Type Equations
arXiv cs.NA0 views
arXiv:2610.31487v1 Announce Type: new Abstract: We construct and analyse a fully discrete scheme for a Kirchhoff-type nonlinear integro-differential equation in two spatial dimensions with time-dependent coefficients. The scheme combines a symmetric three-layer time discretization of Crank--Nicolson type with a conforming finite element method of arbitrary polynomial degree $p$ and a Picard iteration for the nonlinearity. We prove that the Picard iteration is well defined and, that the iterates converge to the Ritz projection of the exact solution measured in fractional-power norms of the discrete Laplacian. Running the same energy argument in the unshifted norms, and under the same hypotheses, we then obtain the optimal order $\mathcal{O} ( h^{p+1} + \tau^2 )$ in $L^2$ and $\mathcal{O} ( h^{p} + \tau^2 )$ in $H^1$ for the full error. For the homogeneous problem with constant coefficients, the exact solution of the scheme moreover preserves a discrete energy identically. Numerical experiments with manufactured solutions reproduce the predicted second order in time and the full spatial order $h^{p+1}$ in the $L^2$-norm.
Source:https://arxiv.org/abs/2610.31487PreviousA Fast Algorithm for Quaternion Tensor Singular Value Decomposition under the T-productNextMixed-Precision Coalesced Matrix-Free Multigrid with Vertex-Patch Smoothing
Related posts

Behavioral Design Unpacked: Nudges, Funnels & Persuasion | UX Unpacked #3
0dev.to UX
National Grid buys Teesside land for UK largest substation to power AI data center
0dev.to News

Leonardo AI Alternative in 2026
0dev.to Discuss

This is my first HacktoberFest, and I don't know what to feel.
0dev.to Discuss

12% of the Economy: The Benchmark That Grades AI Agents Like Employees
0dev.to Machine Learning
A headline said the CMB killed huge black holes. I read the two papers.
0dev.to Career