{"title":"Non-Uniqueness and BV Blowup for Vacuously Liu-Admissible Solutions to P-System Via Computer-Assisted Proof","authors":"Sam G. Krupa","doi":"10.1007/s00021-026-01018-5","DOIUrl":null,"url":null,"abstract":"<div><p>In this paper, we consider non-uniqueness and finite time blowup of the BV-norm for exact solutions to genuinely nonlinear hyperbolic systems in one space dimension, in particular the <i>p</i>-system. The recent Bressan-De Lellis result [<i>Arch. Ration. Mech. Anal.</i>, 247(6):Paper No. 106, 12, 2023] shows that whenever a BV solution exists, with finite (but possibly very large total variation), it is unique if each shock verifies the Liu <i>E</i>-condition. We show non-uniqueness of solutions by convex integration. The solutions we construct are Liu-admissible for a trivial reason: there are no shocks, so the Liu <i>E</i>-condition is vacuously satisfied. But our construction shows exactly that there is an issue, a qualitative difference between small data solutions (for example in the BV class) and large data solutions, where this exact type of phenomenon might occur. Our result can be interpreted as a cautionary example to show that the Liu <i>E</i>-condition is satisfactory for small oscillations but <i>not for large oscillations</i> where typically these type of constructions appear. In particular, we present Riemann initial data which admits infinitely many bounded solutions, each of which experience, not just finite time, but in fact <i>instantaneous</i> blowup of the BV norm. The Riemann initial data is allowed to come from an open set in state space. Our method provably does not admit the natural strictly convex entropy. The proof of our theorem is computer-assisted. Our code is available on the GitHub.</p></div>","PeriodicalId":649,"journal":{"name":"Journal of Mathematical Fluid Mechanics","volume":"28 2","pages":""},"PeriodicalIF":1.5000,"publicationDate":"2026-04-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Mathematical Fluid Mechanics","FirstCategoryId":"100","ListUrlMain":"https://link.springer.com/article/10.1007/s00021-026-01018-5","RegionNum":3,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"MATHEMATICS, APPLIED","Score":null,"Total":0}
引用次数: 0
Abstract
In this paper, we consider non-uniqueness and finite time blowup of the BV-norm for exact solutions to genuinely nonlinear hyperbolic systems in one space dimension, in particular the p-system. The recent Bressan-De Lellis result [Arch. Ration. Mech. Anal., 247(6):Paper No. 106, 12, 2023] shows that whenever a BV solution exists, with finite (but possibly very large total variation), it is unique if each shock verifies the Liu E-condition. We show non-uniqueness of solutions by convex integration. The solutions we construct are Liu-admissible for a trivial reason: there are no shocks, so the Liu E-condition is vacuously satisfied. But our construction shows exactly that there is an issue, a qualitative difference between small data solutions (for example in the BV class) and large data solutions, where this exact type of phenomenon might occur. Our result can be interpreted as a cautionary example to show that the Liu E-condition is satisfactory for small oscillations but not for large oscillations where typically these type of constructions appear. In particular, we present Riemann initial data which admits infinitely many bounded solutions, each of which experience, not just finite time, but in fact instantaneous blowup of the BV norm. The Riemann initial data is allowed to come from an open set in state space. Our method provably does not admit the natural strictly convex entropy. The proof of our theorem is computer-assisted. Our code is available on the GitHub.
期刊介绍:
The Journal of Mathematical Fluid Mechanics (JMFM)is a forum for the publication of high-quality peer-reviewed papers on the mathematical theory of fluid mechanics, with special regards to the Navier-Stokes equations. As an important part of that, the journal encourages papers dealing with mathematical aspects of computational theory, as well as with applications in science and engineering. The journal also publishes in related areas of mathematics that have a direct bearing on the mathematical theory of fluid mechanics. All papers will be characterized by originality and mathematical rigor. For a paper to be accepted, it is not enough that it contains original results. In fact, results should be highly relevant to the mathematical theory of fluid mechanics, and meet a wide readership.