AI & ML
Unconditionally $L^2$-stable explicit--implicit--null filtered Runge--Kutta methods for general nonlinear degenerate parabolic equations
arXiv cs.NA0 views
arXiv:2611.02282v2 Announce Type: replace Abstract: We develop unconditionally $L^2$-stable explicit--implicit--null (EIN) filtered Runge--Kutta (RK) methods for nonlinear degenerate parabolic equations with space-, time-, state-, and gradient-dependent diffusion coefficients. The methods are linearly implicit in the sense that they evaluate nonlinear terms explicitly and require only linear solves. Unlike classical EIN methods that are direct applications of implicit--explicit (IMEX) RKs, our construction is built upon explicit RKs. The key ingredient is two-sided symmetric filtering applied at each RK stage: a high-order filter is applied both before and after evaluating the original nonlinear residual. It leads to a dissipation estimate independent of the time step and retains the underlying RK order formally. Assuming time-dependent splitting operators with stagewise pointwise coefficient bounds, we rigorously prove non-increasing $L^2$ norm for arbitrary time steps. A quadratic matrix criterion covers explicit RK-based methods with positive weights, while strong stability preserving (SSP) compositions yield exact energy identities and weighted dissipation estimates. We present SSP methods through fourth order and a method based on the classical RK4. The framework accommodates spatial discretizations with discrete gradient structures including continuous Galerkin, Bassi--Rebay 1, the method of Brezzi et al., and local discontinuous Galerkin. Compatible summation-by-parts discretizations with simultaneous approximation terms are also covered in their discrete mass norm.
Source:https://arxiv.org/abs/2611.02282PreviousCS-PIKAN: Curvilinear Separable Physics-Informed Kolmogorov-Arnold Networks with Autograd-Based Metric Evaluation and Certified Error Bounds for PDEs on Complex DomainsNextFrom Consensus-Based Optimization to Evolution Strategies: Proof of Global Convergence
Related posts
Re-Testing LLM-Based C-to-Rust Translations Against Their Original C
0arXiv cs.SE
Reliability is a user-specific property of large language models
0arXiv cs.SE
Safe Personal RSI: Evaluating and Repairing Personal-Agent Updates
0arXiv cs.SE
Option-GRPO: Hierarchical RL Against AI Slop in Code Generation for Unseen SDKs
0arXiv cs.SE
Systematic Mapping Protocol. Generative AI-Augmented Requirements Engineering: A Systematic Mapping Study of Practices, Specifications, and Reuse in Software Development
0arXiv cs.SE
RootSink: Reproducible C/C++ Crashes with Tested Repair Sites
0arXiv cs.SE