Uniqueness Domains for 𝐋 ∞ {\bf L}^{\infty} Solutions of 2 × 2 2\times 2 Hyperbolic Conservation Laws
Alberto Bressan, Elio Marconi, Ganesh Vaidya · 2025
Read Uniqueness Domains for 𝐋 ∞ {\bf L}^{\infty} Solutions of 2 × 2 2\times 2 Hyperbolic Conservation Laws by Alberto Bressan, Elio Marconi, Ganesh Vaidya (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-2505-00420.txt.
Introduction
Shock waves in fluids obey conservation laws whose rough solutions may not be unique, a scandal the theory has lived with for decades. Alberto Bressan and co-authors carve out a uniqueness island in a 2025 paper for two-by-two systems. Their solutions start small and, crucially, calm down fast: their total variation fades faster than the classic rate.
Within that fast-calming class, solutions depend on the starting data in a Hölder-continuous way, with the exponent pushed close to one by shrinking the data. They also exhibit starting profiles that achieve the fast calming. The result sharpens what Glimm and Lax began in 1970.
Catalog record
| CREATOR | Alberto Bressan, Elio Marconi, Ganesh Vaidya |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 90 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-00420
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Uniqueness Domains for 𝐋 ∞ {\bf L}^{\infty} Solutions of 2 × 2 2\times 2 Hyperbolic Conservation Laws Alberto Bressan, Elio Marconi, Ganesh Vaidya Abstract For a genuinely nonlinear 2 × 2 2\times 2 hyperbolic system of conservation laws, assuming that the initial data have small 𝐋 ∞ {\bf L}^{\infty} norm but possibly unbounded total variation, the existence of global solutions was proved in a classical paper by Glimm and Lax (1970). In general, the total variation of these solutions decays like t − 1 t^{-1} . Motivated by the theory of fractional domains for linear analytic semigroups, we consider here solutions with faster decay rate: Tot.Var. { u ( t , ⋅ ) } ≤ C t α − 1 \hbox{Tot.Var.}\bigl\{u(t,\cdot)\bigr\}\leq Ct^{\alpha-1} . For these solutions, a uniqueness theorem is proved. Indeed, as the initial data range over a domain of functions with ‖ u ¯ ‖ 𝐋 ∞ ≤ ε 1 \|\bar{u}\|_{{\bf L}^{\infty}}\leq\varepsilon_{1} small enough, solutions with fast decay yield a Hölder continuous semigroup. The Hölder exponent can be taken arbitrarily close to 1 by further shrinking the value of ε 1 > 0 \varepsilon_{1}>0 . An auxiliary result identifies a class of initial data whose solutions have rapidly decaying total variation. 1 Introduction Consider a strictly hyperbolic system of conservation laws in one space dimension u_{t}+f(u)_{x}~=~0, (1.1) with initial data u(0,x)~=~\overline{u}(x). (1.2) It is well known that ( 1.1 ) generates a Lipschitz continuous semigroup of Liu-admissible weak solutions [ 4 , 8 , 9 , 10 , 11 , 13 , 18 , 21 , 27 ] , on a domain {\cal D}\subset{\bf L}^{1}(\mathbb{R};\,\mathbb{R}^{n}) of functions with suitably small total variation. In the papers [ 12 , 31 , 32 , 33 ] a semigroup was constructed on a domain of functions with large but finite total variation. In view of the recent uniqueness theorems proved in [ 14 , 16 ] , these results show that, for a wide class of systems, the Cauchy problem ( 1.1 )-( 1.2 ) has a unique solution which depends continuously on the initial data, as long as the total variation remains bounded. A major open problem is to understand whether the semigroup generated by ( 1.1 ) can be extended to a domain of {\bf L}^{\infty} functions, possibly with unbounded total variation. At present, this is known in the scalar case [ 20 , 28 ] , and for some special systems of Temple class [ 15 ] . In addition, unique solutions with {\bf L}^{\infty} initial data were constructed in [ 17 ] as limit of vanishing viscosity approximations, for a class of 2\times 2 systems of conservation laws in triangular form, not necessarily strictly hyperbolic. Even for 2\times 2 strictly hyperbolic, genuinely nonlinear systems, where existence of {\bf L}^{\infty} solutions can be proved by a compensated compactness argument [ 22 ] , it is not known whether the vanishing viscosity limit yields unique solutions. In particular, the uniqueness of the weak solutions constructed by Glimm and Lax in [ 24 ] has remained an elusive open problem for over 50 years. Aim of this paper is to provide some results in this direction. Our approach is motivated by the theory of parabolic equations and fractional domains for analytic semigroups [ 26 , 35 ] , which we briefly review. Consider a bounded open set \Omega\subset\mathbb{R}^{n} with smooth boundary. Call u(t,\cdot)=e^{t\Delta}\overline{u} the solution to Cauchy problem for the linear heat equation \left\{\begin{array}[]{rll}u_{t}-\Delta u&=~0&\hbox{on}~\Omega,\\[5.69054pt] u&=~0&\hbox{on}~\partial\Omega,\end{array}\right.\qquad\qquad u(0,x)\,=\,\overline{u}(x). (1.3) For any initial data \overline{u}\in{\bf L}^{2}(\Omega) , it is well known that this solution satisfies the decay property \bigl\|\Delta u(t,\cdot)\bigr\|_{{\bf L}^{2}}~\leq~{C\over t}\,\,\|\overline{u}\|_{{\bf L}^{2}}\,. (1.4) Moreover, if the initial data lies in the domain of a fractional power of the Laplace operator, say \bar{u}\in X^{\alpha}\doteq Dom(-\Delta)^{\alpha} for some 0<\alpha<1 , then a faster decay rate is achieved: \bigl\|\Delta u(t,\cdot)\bigr\|_{{\bf L}^{2}(\Omega)}~\leq~{C\over t^{1-\alpha}}\,\|\overline{u}\|_{X^{\alpha}}\,. (1.5) Relying on ( 1.5 ) one can show that the semilinear parabolic Cauchy problem \left\{\begin{array}[]{rll}u_{t}-\Delta u&=~F(x,u,\nabla u)&\hbox{on}~\Omega,\\[5.69054pt] u&=~0&\hbox{on}~\partial\Omega,\end{array}\right.\qquad\qquad u(0,x)\,=\,\overline{u}(x), (1.6) with F Lipschitz continuous, is well posed for initial data \overline{u}\in X^{\alpha} , for a suitable 0<\alpha<1 depending on the dimension n of the space. For all details we refer to [ 26 ] . We shall pursue a similar approach in connection with a 2\times 2 genuinely nonlinear, strictly hyperbolic system of conservation laws, as considered in the classical paper [ 24 ] . We recall that, for any bounded measurable initial data \overline{u}:\mathbb{R}\mapsto\mathbb{R}^{2} , if the norm \|\overline{u}\|_{{\bf L}^{\infty}}~\leq~\varepsilon_{1} (1.7) is sufficiently small, then the Cauchy problem ( 1.1 )-( 1.2 ) has a weak solution, global in time. Its {\bf L}^{\infty} norm and its total variation satisfy bounds of the form \bigl\|u(t,\cdot)\bigr\|_{{\bf L}^{\infty}}~\leq~M_{0}\,\sqrt{\|\bar{u}\|_{{\bf L}^{\infty}}}\,, (1.8) \hbox{Tot.Var.}\bigl\{u(t,\cdot)\,;~[a,b]\bigr\}~\leq~C_{0}\cdot\left(\sqrt{\|\bar{u}\|_{{\bf L}^{\infty}}}+{b-a\over t}\right)\,, (1.9) for some constants M_{0},C_{0} , any interval [a,b] and every time t\in\,]0,1] . See also [ 5 ] for a simpler proof, based on front tracking approximations. In analogy with ( 1.5 ), we introduce the domain \begin{array}[]{rl}\widetilde{\cal D}_{\alpha}&\displaystyle\doteq~\Bigg\{\overline{u}\in{\bf L}^{\infty}(\mathbb{R};\,\mathbb{R}^{2})\,;~\hbox{ the Cauchy problem (\ref{1})-(\ref{2}) has a weak solution}\\[11.38109pt] &\qquad\qquad\qquad\displaystyle\hbox{such that}\quad\sup_{0<t<1}t^{1-\alpha}\cdot\hbox{Tot.Var.}\bigl\{u(t,\cdot)\bigr\}\,<\,+\infty\bigg\}\,.\end{array} (1.10) We observe that \bar{u}\in\widetilde{\cal D}_{\alpha} implies \bigl\|u(\tau)-\overline{u}\|_{{\bf L}^{1}}~=~{\cal O}(1)\cdot\tau^{\alpha}\qquad\qquad\tau\in[0,1]. (1.11) Indeed, by Theorem 4.3.1 in [ 21 ] , one has \limsup_{h\to 0+}~{\bigl\|u(t+h)-u(t)\bigr\|_{{\bf L}^{1}}\over h}~=~{\cal O}(1)\cdot\hbox{Tot.Var.}\bigl\{u(t)\bigr\}\,. (1.12) For \bar{u}\in\widetilde{\cal D}_{\alpha} , the estimate ( 1.11 ) is an easy consequence of ( 1.12 ). Here and in the sequel, the Landau notation {\cal O}(1) denotes a uniformly bounded quantity. Our main result shows that, for every initial data \overline{u}\in\widetilde{\cal D}_{\alpha} with \|\bar{u}\|_{{\bf L}^{\infty}} sufficiently small, the entropy weak solution which satisfies ( 1.11 ) is unique. Moreover, such solutions form a Hölder continuous semigroup w.r.t. the {\bf L}^{1} distance on the initial data. By restricting the domain of the semigroup to functions with even smaller {\bf L}^{\infty} norm, the Hölder exponent of the semigroup can be rendered arbitrarily close to 1. The main idea is illustrated in Fig. 1 . For any \tau>0 , consider the positively invariant domain \begin{array}[]{l}{\cal D}^{(\tau)}~=~\Big\{u(t,\cdot)\,;~~t\geq\tau,~~u~\hbox{is an entropy weak solution of (\ref{1})-(\ref{2}) satisfying (\ref{TVdec}),}\\[8.53581pt] \qquad\qquad\qquad\qquad\hbox{with}~\bar{u}\in{\bf L}^{1}(\mathbb{R};\,\mathbb{R}^{2}),~\|\bar{u}\|_{{\bf L}^{\infty}}\leq\varepsilon_{1}\Big\}.\end{array} (1.13) Here \varepsilon_{1}>0 is a fixed constant, small enough so that the Glimm-Lax decay estimates ( 1.9 ) on the total variation are valid. Observing that these functions have uniformly bounded variation on bounded intervals, our first goal is to show that ( 1.1 ) generates a Lipschitz semigroup S on the domain {\cal D}^{(\tau)} . More precisely \bigl\|S_{t}\bar{u}-S_{t}\bar{v}\bigr\|_{{\bf L}^{1}}~\leq~L{(\tau)}\cdot\|\bar{u}-\bar{v}\|_{{\bf L}^{1}}\,,\qquad\qquad\hbox{for all}~\bar{u},\bar{v}\in{\cal D}^{(\tau)},~~t\in[0,1], (1.14) for some Lipschitz constant L(\tau) . Of course, since the total variation of functions \bar{u}\in{\cal D}^{(\tau)} becomes arbitrarily large as \tau\to 0 , we expect that \lim_{\tau\to 0}L(\tau)=+\infty . Next, consider two solutions u_{1},u_{2} of the hyperbolic system ( 1.1 ) with the same initial data \bar{u}\in\widetilde{D}_{\alpha} . For any 0<\tau<t\leq 1 , by ( 1.14 ) and ( 1.11 ), their distance satisfies \begin{array}[]{l}\bigl\|u_{1}(t)-u_{2}(t)\|_{{\bf L}^{1}}~\leq~L(\tau)\cdot\bigl\|u_{1}(\tau)-u_{2}(\tau)\|_{{\bf L}^{1}}\\[5.69054pt] \qquad\leq~L(\tau)\cdot\Big(\bigl\|u_{1}(\tau)-\bar{u}\|_{{\bf L}^{1}}+\bigl\|u_{2}(\tau)-\bar{u}\|_{{\bf L}^{1}}\Big)~=~L(\tau)\cdot{\cal O}(1)\cdot\tau^{\alpha}.\end{array} (1.15) Keeping t>0 fixed and letting \tau\to 0 , the condition \lim_{\tau\to 0}~\tau^{\alpha}\,L(\tau)~=~0 (1.16) implies u_{1}(t)=u_{2}(t) for all t\in[0,1] , proving uniqueness. Figure 1: Sketch of the estimates ( 1.14 )–( 1.15 ). To carry out this program, two main ingredients are needed: (1) Estimate the Lipschitz constant L(\tau) . In particular, give an asymptotic estimate on how it grows as \tau\to 0 . (2) Identify a class