Abstract
In this paper we propose a new algorithm for the numerical approximation of solutions to the multidimensional Kolmogorov equation, based on the averaging of Feynman-Chernoff iterations for random operator-valued functions. In the case when the values of operator-valued functions belong to the representation of some finite-dimensional Lie group, the proposed algorithm has a lower computational complexity compared to the standard Monte Carlo algorithm that uses the Feynman-Kac formula. In particular, we study the case of a group of affine transformations of a Euclidean space. For the considered algorithms we also present the results of numerical calculations.
Publisher
Keldysh Institute of Applied Mathematics
Reference20 articles.
1. Graham C., Talay D. Stochastic Simulation and Monte Carlo Methods// Springer Berlin, Heidelberg (2013).
2. Graham C., Kurtz T.G., Méléard S., Protter P.E., Pulvirenti M., Talay D. Probabilistic Models for Nonlinear Partial Differential Equations// Springer-Verlag Berlin, Heidelberg (1996).
3. Sherman A.S., Peskin C.S. A Monte Carlo method for scalar reaction diffusion equations// SIAM J. Sci. Stat. Comput., 7, 1360–1372 (1986).
4. Richter L., Berner J., Robust SDE-Based Variational Formulations for Solving Linear PDEs via Deep Learning// Proceedings of the 39th International Conference on Machine Learning, PMLR, 162, 18649-18666 (2022).
5. Kovachki N., Li Z., Liu B., Azizzadenesheli K., Bhattacharya K., Stuart A., Anandkumar A. Neural Operator: Learning Maps Between Function Spaces// arXiv:2108.08481 (2021).