Convergence of hyperbolic approximations to higher-order PDEs for smooth solutions
Jan Giesselmann, Hendrik Ranocha · 2025
Read Convergence of hyperbolic approximations to higher-order PDEs for smooth solutions by Jan Giesselmann, Hendrik Ranocha (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-04112.txt.
Introduction
Jan Giesselmann and Hendrik Ranocha published a 2025 paper in math that puts a shortcut on firm ground. Some laws of physics, for water waves and pattern growth, involve high-order terms that make sums awkward. A standard fix rewrites them as sets of simpler hyperbolic equations with a small parameter. Workers used the trick for years; proof that it converges was missing.
The authors supply the proof for five equations, including the Korteweg-de Vries and Kuramoto-Sivashinsky equations. Their demands are gentle: the solution must be smooth, while the rough versions need only meet the weak entropy rules that computers produce. Tests with numbers back the theory. The result matters because such models guide science. Knowing the shortcut truly nears the target equation, as the parameter shrinks, turns a hopeful trick into a sound tool for wave modelers.
Catalog record
| CREATOR | Jan Giesselmann, Hendrik Ranocha |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 69 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-04112
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Convergence of hyperbolic approximations to higher-order PDEs for smooth solutions Jan Giesselmann, Hendrik Ranocha Abstract We prove the convergence of hyperbolic approximations for several classes of higher-order PDEs, including the Benjamin-Bona-Mahony, Korteweg-de Vries, Gardner, Kawahara, and Kuramoto-Sivashinsky equations, provided a smooth solution of the limiting problem exists. We only require weak (entropy) solutions of the hyperbolic approximations. Thereby, we provide a solid foundation for these approximations, which have been used in the literature without rigorous convergence analysis. We also present numerical results that support our theoretical findings. keywords hyperbolization, hyperbolic relaxation, relative energy, relative entropy, structure-preserving methods, energy-conserving methods, summation-by-parts operators, Benjamin-Bona-Mahony equation, Korteweg-de Vries equation, Gardner equation, Kawahara equation, Kuramoto-Sivashinsky equation AMS subject classification. 35L03, 35M11, 65M12, 65M06, 65M20 1 Introduction Hyperbolic approximations, also known as hyperbolic relaxations or hyperbolizations, of higher-order partial differential equations (PDEs) have been studied for a long time, such as the hyperbolic heat equation discussed by Cattaneo and Vernotte [ 16 , 70 ] . These hyperbolic approximations have been applied in various contexts, including elliptic problems [ 52 , 63 , 64 , 54 ] , dispersive wave equations [ 35 , 7 , 13 , 29 , 61 ] , and other higher-order PDEs [ 21 , 34 , 45 ] . Typically, such hyperbolizations are developed for specific PDEs using tailored techniques with the goal of maintaining essential structures of the original PDE. Notably, an augmented Lagrangian approach [ 29 ] has proven successful. Ketcheson and Biswas [ 45 ] introduced a broad framework to create stable hyperbolizations of a class of linear PDEs and presented numerical results suggesting that this method may also be effective for nonlinear and more complex PDEs. They additionally offer a review of several hyperbolizations along with motivations for developing them. As far as we know, hyperbolizations have been used quite extensively, but usually without a rigorous convergence analysis to support their approximation properties. Some notable exceptions include the hyperbolic heat equation [ 51 ] , and the Serre-Green-Naghdi system at the analytical level [ 23 ] , as well as the Korteweg-de Vries and Benjamin-Bona-Mahony equations at the numerical level [ 9 , 10 ] . In this paper, we provide a rigorous convergence analysis of a broad class of hyperbolic approximations to higher-order PDEs using the relative energy/entropy method. We start with a structure similar to the Benjamin-Bona-Mahony equation that involves mixed space-time derivatives in Section 2 . We introduce the relative energy method in this context and explain some choices we make in the convergence analysis. Section 3 is dedicated to more general classes of PDEs with higher-order spatial derivatives. We perform the convergence analysis for each class and present numerical results that support our findings. Finally, we summarize our results in Section 4 and discuss potential extensions of our work. The relative entropy (sometimes also called relative energy) method has a long history for both hyperbolic and parabolic problems. It relies on leveraging an energy or entropy structure of the considered system in order to control the distance between two solutions. Depending on the context, the functionals of interest are either called energy or entropy. In this paper, we will use the term relative energy consistently (although we could call it relative entropy equally well). For consistency with the literature, we will still use the common notion of weak entropy solutions (and will not call them weak energy solutions). For the use of the relative energy method in parabolic models, in particular with respect to studying the convergence to steady states, we refer to [ 42 ] and references therein. For hyperbolic models, it dates back to Dafermos [ 18 ] and di Perna [ 22 ] , where it was, in particular, used to establish weak-strong uniqueness. In addition to being a well-established tool for investigating the stability of solutions to systems of hyperbolic conservation laws, it also has a long history as a tool for describing the relationship between solutions of different models, where usually one can be understood as the relaxation limit of the other. It has been particularly successful in identifying low Mach and large friction limits of hyperbolic systems, see [ 69 , 48 , 36 , 30 , 27 , 32 ] , but also in comparing different moment approximations of kinetic equations [ 2 ] . For an extension to hyperbolic-parabolic systems, see [ 17 ] . A general property in this analysis is that the solution to the limiting system needs to be a strong solution whereas the solutions to the approximating system can be allowed to be (weak) entropy solutions. In this paper, we leverage the fact that many hyperbolic approximations of higher-order equations, as derived in [ 45 ] , are equipped with a (rather simple) energy that allows for relative energy estimates. One feature that creates challenges not encountered in previous works is that the energy degenerates, i.e., the modulus of convexity approaches zero in many directions when the limit problem is approached. We explain how we handle this in Section 2.1 . Let us also notice that several hyperbolic approximations suggested in the literature [ 45 ] do not admit a (simple) energy. In these cases, we suggest alternative hyperbolic approximations that do admit a (simple) energy. Incidentally, these new approximations are easier to handle numerically in the sense that we can construct energy-preserving discretizations. To demonstrate the analytical convergence results, we implement several model problems numerically using (upwind) summation-by-parts (SBP) operators in space and (additive) Runge-Kutta methods in time. More details on upwind SBP operators can be found in [ 50 , 59 ] ; applications to hyperbolic approximations of the KdV and BBM equations are discussed in [ 9 , 10 ] . A more general discussion of the background of SBP operators is contained in the review articles [ 67 , 31 ] . Although we mainly focus on finite differences [ 46 , 66 , 15 ] , the general framework of SBP operators also includes finite volumes [ 53 ] , finite elements [ 40 , 39 , 1 ] , discontinuous Galerkin (DG) methods [ 33 , 14 ] , flux reconstruction (FR) [ 41 , 71 , 60 ] , active flux methods [ 28 , 5 ] , as well as meshless schemes [ 38 ] . We use the time integration schemes of [ 4 ] and refer to [ 12 , 11 ] for background information on implicit-explicit (IMEX) and asymptotic-preserving methods. All source code and data required to reproduce the numerical results are available in our reproducibility repository [ 37 ] . We implemented all methods in Julia [ 8 ] , used SummationByPartsOperators.jl [ 56 ] for the spatial discretizations, and created the figures using Makie.jl [ 19 ] . 2 Dispersive equations with mixed space-time derivatives First, we concentrate on PDEs of the form \partial_{t}u(t,x)+\partial_{x}f(u(t,x))-\partial_{x}^{2}\partial_{t}u(t,x)=0 (2.1) with periodic boundary conditions and appropriate initial data. A classical example is the Benjamin-Bona-Mahony (BBM) equation [ 6 ] \partial_{t}u(t,x)+\partial_{x}\frac{u(t,x)^{2}}{2}-\partial_{x}^{2}\partial_{t}u(t,x)=0, (2.2) where f(u)=u^{2}/2 . Note that the typical form of the BBM equation includes an additional linear term +\partial_{x}u(t,x) that is not present in ( 2.2 ) since we follow the