High order numerical schemes for second-order FBSDEs with applications to stochastic optimal control
School of Mathematics & Finance Institute, Shandong University(山东大学数学与金融学院);LSEC, Institute of Computational Mathematics, Academy of Mathematics and Systems Science, Chinese Academy of Sciences(中国科学院数学与系统科学研究院计算数学与科学工程计算研究所)
本文是关于多步方案求解向前向后随机微分方程(FBSDEs)及相关问题的一系列论文之一。我们扩展了[Zhao, Fu, Zhou, SIAM J. Sci. Comput., 36 (2014), pp. A1731-A1751.]中的多步方案,用于求解二阶FBSDEs(2FBSDEs)。多步方案的关键特征是使用欧拉法离散正向SDE,从而显著降低整个计算复杂度。此外,证明了通常感兴趣的量(如2FBSDEs中的解元组(Y_t, Z_t, A_t, Γ_t))仍具有高阶精度。给出了几个数值示例以展示所提数值方案的有效性。还展示了所提数值方案在随机优化控制问题中的应用。
英文摘要:
This is one of our series papers on multistep schemes for solving forward backward stochastic differential equations (FBSDEs) and related problems. Here we extend (with non-trivial updates) our multistep schemes in [W. Zhao, Y. Fu and T. Zhou, SIAM J. Sci. Comput., 36 (2014), pp. A1731-A1751.] to solve the second order FBSDEs (2FBSDEs). The key feature of the multistep schemes is that the Euler method is used to discrete the forward SDE, which dramatically reduces the entire computational complexity. Moreover, it is shown that the usual quantities of interest (e.g., the solution tuple $(Y_t, Z_t, A_t, Γ_t)$ in the 2FBSDEs) are still of high order accuracy. Several numerical examples are given to show the effective of the proposed numerical schemes. Applications of our numerical schemes for stochastic optimal control problems are also presented.
Convergence of particles and tree based scheme for singular FBSDEs
我们研究了[Chassagneux and Yang,2022]中引入的理论分裂方案的实现,用于奇异FBSDES [Carmona and Delarue 2013]及其相关的准线性退化PDE。完全可实现的算法是基于粒子近似的运输运营商和树状近似的扩散算子出现在理论分裂。在适当的条件下,我们证明了我们的数值方法对系数函数的收敛速度。这验证了[Chassagneux和Yang,2022]中获得的一些数值结果。最后,我们提出了各种实现的算法,并讨论其在实践中的效率的数值部分的文件。
Deep Truncated FBSDE Method for High-Dimensional Nonlinear PDEs and Fully Coupled FBSDEs
Research Institute of Intelligent Complex Systems, Fudan University(复旦大学智能复杂系统研究院);Department of Mathematics, Friedrich-Alexander-Universität Erlangen–Nürnberg(埃尔朗根-纽伦堡大学数学系)
A New Kind of High-Order Multi-step Schemes for Forward Backward Stochastic Differential Equations
School of Mathematics & Finance Institute, Shandong University(山东大学数学与金融学院);LSEC, Institute of Computational Mathematics, Academy of Mathematics and Systems Science, Chinese Academy of Sciences(中国科学院数学与系统科学研究院计算数学研究所)