Skip to Main content Skip to Navigation
Theses

Numerical approximation of hyperbolic stochastic scalar conservation laws

Abstract : In this thesis, we study a scalar hyperbolic conservation law of order one, with stochastic source term and non-linear flux. The source term can be seen as the superposition of an infinity of Gaussian noises depending on the conserved quantity. We give a definition of solution of this stochastic partial differential equation (SPDE) with an intermediate point of view between that of the analyst (nonregular solution in space, introduction of an additional kinetic variable) and that of the probabilist (right continuous with left limits in time stochastic process solution). Uniqueness of the solution is proved thanks to a doubling of variables à la Kruzkov. We study the stability of the conservation law, in order to give a general theorem where the conditions of existence of a solution and conditions of convergence of a sequence of approximate solutions towards the solution of the conservation law are given. This study is done thanks to probabilistic tools : representation of martingales in the form of stochastic integrals, existence of a probability space on which the convergence of probability measures is equivalent to the almost sure convergence of random variables. To finish the study, we prove the existence of a solution thanks to the properties of the approximation of the SPDE given by an explicit in time Finite Volumes numerical scheme, then the convergence of this approximation towards the solution of the SPDE. The tools used are those of the numerical analysis, especially those of the Finite Volume Method, and those of the stochastic calculus (probabilistic tools).
Complete list of metadatas

Cited literature [87 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-01661124
Contributor : Sylvain Dotti <>
Submitted on : Friday, May 18, 2018 - 2:19:15 PM
Last modification on : Friday, June 1, 2018 - 1:14:36 AM

File

TheseSylvainDotti.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-01661124, version 4

Collections

Citation

Sylvain Dotti. Numerical approximation of hyperbolic stochastic scalar conservation laws. Analysis of PDEs [math.AP]. Aix-Marseille Université (AMU), 2017. English. ⟨tel-01661124v4⟩

Share

Metrics

Record views

183

Files downloads

159