Deep learning numerical methods for high-dimensional fully nonlinear PIDEs and coupled FBSDEs with jumps

This repository contains the code for the numerical experiments of the paper "Deep learning numerical methods for high-dimensional fully nonlinear PIDEs and coupled FBSDEs with jumps". We provide a minimal implementation of the algorithm to solve both low and high dimensional problems. With reference to existing notebooks, one can easily generalize to other issues.

Dependencies

Before running the notebook, ensure that PyTorch is installed and the appropriate environment is activated, as the code is written in Jupyter Notebook using the PyTorch backend.

Reference

@misc{wang2023deeplearningnumericalmethods,
      title={Deep learning numerical methods for high-dimensional fully nonlinear PIDEs and coupled FBSDEs with jumps}, 
      author={Wansheng Wang and Jie Wang and Jinping Li and Feifei Gao and Yi Fu},
      year={2023},
      eprint={2301.12895},
      archivePrefix={arXiv},
      primaryClass={math.NA},
      url={https://arxiv.org/abs/2301.12895}, 
}