Global Existence for a Class of Keyfitz–Kranzer Systems with Application to Thin-Film Flows
Rahul Barthwal, Philipp Öffner, Christian Rohde · 2026
Read Global Existence for a Class of Keyfitz–Kranzer Systems with Application to Thin-Film Flows by Rahul Barthwal, Philipp Öffner, Christian Rohde (2026) 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-2604-16000.txt.
Introduction
Barthwal, Öffner, and Rohde prove global weak entropy solutions exist for a class of non-symmetric Keyfitz-Kranzer systems, including lubrication models for thin-film flow. They identify entropy and entropy-flux pairs admissible for a tailored second-order approximate system, motivated by higher-order dissipation. Finding an invariant region in state space yields a-priori bounds for solutions of the approximate system. Using the parabolic and transport structure tied to Riemann invariants, they justify the vanishing-diffusion limit and establish weak entropy solutions for the Cauchy problem.
Catalog record
| CREATOR | Rahul Barthwal, Philipp Öffner, Christian Rohde |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 61969 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.16000
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Global Existence for a Class of Keyfitz–Kranzer Systems with Application to Thin-Film Flows Rahul Barthwal, Philipp Öffner, Christian Rohde Abstract We prove the existence of global weak entropy solutions for a class of non-symmetric Keyfitz– Kranzer type systems that includes lubrication models for thin-film flow. We identify a family of entropy/entropy-flux pairs for these first-order systems, which is, in particular, admissible for a tailored second-order approximate system. The latter is motivated by higher-order dissipation operators in thin-film flow models. By identifying an invariant region in the state space, it is possible to derive a-priori L ∞ \displaystyle L^{\infty} -bounds for the sequence of solutions to the approximate system. Exploiting the parabolic and transport structure of the equations associated with the Riemann invariants, we then rigorously justify the vanishing-diffusion limit and establish the existence of weak entropy solutions for the Cauchy problem for the first-order systems. Key words. Keyfitz–Kranzer systems, thin-film flow, vanishing diffusion limit, entropy solutions MSC codes. 35L40 35L45 35L65 35A01 35D30 1 Introduction Let \displaystyle T>0 , \displaystyle n\in\mathbb{N} and consider the Cauchy problem for \displaystyle\mathbf{U}:\Omega_{T}:=\mathbb{R}\times(0,T)\to\mathcal{U}\subset\mathbb{R}^{n} given by \begin{array}[]{rcll}\mathbf{U}_{t}+{\mathbf{F}(\mathbf{U})}_{x}&=&0&\text{ in }\Omega_{T},\\ \mathbf{U}(\cdot,0)&=&\mathbf{U}_{0}&\text{ in }\mathbb{R}.\end{array} (1.1) In ( 1.1 ), \displaystyle\mathbf{F}:\mathcal{U}\to\mathbb{R}^{n} denotes the flux function and \displaystyle\mathbf{U}_{0}:\mathbb{R}\to\mathcal{U} is the initial datum. We assume that ( 1.1 ) is hyperbolic, i.e., the Jacobian \displaystyle\mathbf{D}\mathbf{F} of \displaystyle\mathbf{F} is diagonalizable in the state space \displaystyle\mathcal{U} and has real eigenvalues. Depending on \displaystyle\mathbf{F} , an appropriate weak solution concept is required for ( 1.1 ). The notion of weak entropy solutions refers to the existence of an entropy/entropy-flux pair \displaystyle(\mathcal{E},\mathcal{Q}) for ( 1.1 ), i.e., the function \displaystyle{\mathcal{E}}\in C^{2}({\mathcal{U}},\mathbb{R}) is supposed to be strictly convex and \displaystyle{\mathcal{Q}}\in C^{1}({\mathcal{U}},\mathbb{R}) has to be chosen such that the compatibility condition (\nabla_{{\mathbf{U}}}{\mathcal{E}})^{\top}{\mathbf{D}}{\mathbf{F}}=(\nabla_{\mathbf{U}}{\mathcal{Q}})^{\top} (1.2) holds in \displaystyle\mathcal{U} . A weak entropy solution is a weak solution of ( 1.1 ) that dissipates the entropy \displaystyle\mathcal{E} , see Definitions 2.3 and 2.4 for precise formulations. A standard approach to construct global weak entropy solutions is to introduce a second-order approximation of the quasilinear evolution equations in ( 1.1 ), namely, \displaystyle\displaystyle\mathbf{U}^{\epsilon}_{t}+{\mathbf{F}(\mathbf{U}^{\epsilon})}_{x}=\epsilon\big(\mathbf{B}(\mathbf{U}^{\epsilon})\mathbf{U}^{\epsilon}_{x}\big)_{x}\qquad\text{ in }\Omega_{T}. (1.3) Here, \displaystyle\epsilon>0 is a small parameter and \displaystyle\mathbf{B}:\mathcal{U}\to\mathbb{R}^{n\times n} is a suitably chosen diffusion matrix. The existence of a weak entropy solution for ( 1.1 ) is then established by deriving suitable estimates for the viscous approximations ( 1.3 ) and passing to the limit as \displaystyle\epsilon\to 0 . The system ( 1.3 ) is regarded as an admissible approximation of ( 1.1 ) if it fits to the entropy structure of the underlying hyperbolic system ( 1.1 ), in the sense that the entropy/entropy-flux pairs associated with ( 1.1 ) satisfy appropriate dissipation relations for the approximation system. Indeed, this admissibility is essential to derive the compactness properties required to pass to the limit \displaystyle\epsilon\to 0 in ( 1.3 ). For many hyperbolic systems, choosing the diffusion matrix as \displaystyle\mathbf{B}(\mathbf{U})=\mathbf{I} , \displaystyle\mathbf{I} being the unit matrix, has proven to be sufficient to produce an admissible approximation; see e.g. [ 7 , 12 , 16 , 21 ] . This provides sufficient a-priori estimates for the approximate system and allows to pass to the limit \displaystyle\epsilon\rightarrow 0 to yield global weak entropy solutions of the system ( 1.1 ). However, this is not the case for general systems. In particular, for systems with more delicate entropy structure, the specific form of the diffusion matrix \displaystyle\mathbf{B} plays a crucial role in the construction of an admissible regularization. In such situations, one often needs a diffusion matrix specifically tailored to the structure of the system rather than the standard unit matrix. An example is the work of Chen&Perepelitsa [ 8 ] on the compressible Euler equations with spherical symmetry, where the vanishing viscosity approximation is built with designed viscosity terms, in order to establish convergence to global weak entropy solutions. To the best of the authors’ knowledge, general existence results based on genuinely nonlinear or system-adapted diffusion matrices \displaystyle\mathbf{B} remain relatively limited in comparison with the classical theory for \displaystyle\mathbf{B(U)}=\mathbf{I} . In this article, we study the Cauchy problem for ( 1.1 ) with \displaystyle n=2 for a class of first-order conservation laws of the form \begin{aligned} u_{t}+{\big(u\phi(r(u,v))\big)}_{x}&=0,\vskip 5.69046pt\\[5.16663pt] v_{t}+{\big(v\phi(r(u,v))\big)}_{x}&=0,\end{aligned}\hskip 28.45274pt (1.4) with initial data given by \displaystyle\displaystyle(u,v)(\cdot,0)=(u_{0},v_{0})\in\left(L^{\infty}(\mathbb{R})\right)^{2}. (1.5) In ( 1.4 ) the coupling is realized in the product function r(u,v)=uv. (1.6) Moreover, \displaystyle\phi:\mathbb{R}\mapsto\mathbb{R} is a smooth function of \displaystyle r satisfying the following assumption. Assumption 1.1 . \displaystyle\phi\in C^{\infty}(\mathbb{R}) with \displaystyle\phi^{\prime}(r)>0 and \displaystyle\text{{\rm{meas}}}\{r:\,3\phi^{\prime}(r)+2r\phi^{\prime\prime}(r)=0\}=0 . Furthermore, the state space for the two-dimensional unknown \displaystyle{\mathbf{U}}=(u,v)^{\top} is defined for some constant \displaystyle m>0 as \displaystyle\displaystyle\mathcal{U}=\mathcal{U}_{m}=\{(u,v)^{\top}\in\mathbb{R}^{2}|~~u,v\geq m\}. (1.7) The system ( 1.4 ) turns out to be hyperbolic in \displaystyle{\mathcal{U}}_{m} . The results of this paper stem from the discovery that we can provide explicit Riemann invariants and a class of entropy/entropy-flux pairs for the system ( 1.4 ) in \displaystyle{\mathcal{U}}_{m} , see Section 2 . Systems of the form ( 1.4 ) belong to the Keyfitz–Kranzer class of hyperbolic systems [ 18 ] . General \displaystyle(n\times n) -Keyfitz–Kranzer systems provide a widely studied class of hyperbolic systems that are given by \displaystyle\displaystyle u_{it}+{(u_{i}(\Phi(u_{1},u_{2},\ldots,u_{n})))}_{x}=0,~i=1,2,\ldots,n. (1.8) The systems of the form ( 1.8 ) have been analyzed in diverse settings for various choices of the function \displaystyle\Phi . Some of the particular choices of \displaystyle\Phi include \displaystyle\Phi({\mathbf{U}})=\phi(|{\mathbf{U}}|) [ 14 , 15 ] for \displaystyle(n\times n )-systems and \displaystyle\Phi(u_{1},u_{2})=\Phi(\alpha u_{1}+\beta u_{2}) , \displaystyle\Phi(u_{1},u_{2})=\Phi(u_{1}/u_{2}) [ 27 , 28 ] , or \displaystyle\Phi(u_{1},u_{2})=\Phi(u_{1}u_{2}) [ 25 ] for ( \displaystyle 2\times 2 )-systems. It is interesting to see that the different choices of \displaystyle\Phi describe quite distinct physical phenomena. For instance, the choice of \displaystyle\Phi({\mathbf{U}})=\phi(|{\mathbf{U}}|) describes a simplified magnetohydrodynamics model [ 11 , 14 ] or the elastic string model [ 18 ] , while \displaystyle\Phi({\mathbf{U}})={k_{i}u_{i}}/{(1+u_{1}+\cdots+u_{n}}) describes \displaystyle n -component chromatography equations [ 17 ] . For the ( \displaystyle 2\times 2 )-system with \displaystyle\Phi(u_{1},u_{2})=u_{1}u_{2}/2 , the system ( 1.8 ) or equivalently the system ( 1.4 ) reduces to the system derived in [ 2 , 9 ] , which governs the first-order dynamics of a thin film flow under the influence of a perfectly soluble solute; see also [ 3 , 23 ] . It is remarkable to note that many of the Keyfitz–Kranzer type systems actually belong to the Temple systems [ 26 ] , where the shock and rarefaction curves coincide and form a straight line in the phase space. For Temple class systems, the total variation of Riemann invariants can be shown to be decreasing over time, and thus, the global existence of weak solutions with any initial data with bounded variation is possible using the classical \displaystyle L^{1} -theory. In particular, Serre investigated Temple systems in this regard in [ 24 ] . He also established the existence of a global unique weak solution for the case when the initial data is of bounded variation [ 24 ] . The condition of bounded variation initial data was further relaxed to a broader \displaystyle L^{\infty} -space for ( \displaystyle n\times n )-Temple systems by Heibig [ 16 ] . By employing the compensated compactness framework, he was able to prove the existence of global weak solutions for \displaystyle L^{\infty} -initial data. Moreover, by constructing Oleinik-type estimates, the uniqueness of