Particle systems for mean reflected BSDEs with jumps

Yiqing Lin,Kun Xu
2024-06-18
Abstract:In this paper, we study the mean reflected backward stochastic differential equations with jump (BSDEJs). We extend the work of Briand and Hibon on the propagation of chaos for mean reflected BSDEs \cite{briand2021particles} to the jump framework. Besides, we study the reflections for the particle system and obtain the rate of of convergence of the particle system towards the deterministic flat solution to the mean reflected BSDEJ.
Probability
What problem does this paper attempt to address?
The paper primarily studies a class of mean reflected backward stochastic differential equations with jumps (BSDEJs). Specifically, the objectives of the paper can be summarized as follows: 1. **Extension of Existing Work**: The paper extends the work of Briand and Hibon on mean reflected BSDEs to include jumps. That is, it considers a new class of mean reflected BSDEJs with jump processes. 2. **Theoretical Framework**: The paper first establishes the basic theoretical framework for this class of BSDEJs, including definitions, assumptions, etc., and proves the existence and uniqueness of solutions under specific conditions. 3. **Study of Particle Systems**: The paper further investigates the particle systems related to these BSDEJs and explores the propagation of chaos properties of these particle systems. In particular, it proves that the solutions of the particle systems converge to a deterministic flat solution. 4. **Convergence Rate Analysis**: The paper also provides estimates for the convergence rate of the particle system solutions to the deterministic flat solution. In summary, the main purpose of this paper is to study the theoretical properties of mean reflected BSDEJs with jump processes and to approximate their solutions using particle systems, while also providing estimates for the convergence rate. This is of significant importance for understanding and solving such stochastic differential equations and their applications.