Solving high-dimensional Hamilton–Jacobi–Bellman PDEs using neural networks: perspectives from the theory of controlled diffusions and measures on path space
Partial Differential Equations and Applications · 2021-08 · 2 (4)
DOI: 10.1007/s42985-021-00102-x ↗Kurz gefasst
Fachartikel in der Zeitschrift Partial Differential Equations and Applications, online erschienen am 21.06.2021. Verfasst von Nikolas Nüsken und 1 weiteren Autor:innen. Die Arbeit wurde bisher 53-mal zitiert (Crossref) und nennt 128 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Deutsche Forschungsgemeinschaft.
Abstract (Originalsprache)
Optimal control of diffusion processes is intimately connected to the problem of solving certain Hamilton–Jacobi–Bellman equations. Building on recent machine learning inspired approaches towards high-dimensional PDEs, we investigate the potential of iterative diffusion optimisation techniques, in particular considering applications in importance sampling and rare event simulation, and focusing on problems without diffusion control, with linearly controlled drift and running costs that depend quadratically on the control. More generally, our methods apply to nonlinear parabolic PDEs with a certain shift invariance. The choice of an appropriate loss function being a central element in the algorithmic design, we develop a principled framework based on divergences between path measures, encompassing various existing methods. Motivated by connections to forward-backward SDEs, we propose and study the novel log-variance divergence, showing favourable properties of corresponding Monte Carlo estimators. The promise of the developed approach is exemplified by a range of high-dimensional and metastable numerical examples.
Förderung
- Deutsche Forschungsgemeinschaft
Zitieren
APA
Nüsken, N., & Richter, L. (2021). Solving high-dimensional Hamilton–Jacobi–Bellman PDEs using neural networks: perspectives from the theory of controlled diffusions and measures on path space. Partial Differential Equations and Applications, 2(4). https://doi.org/10.1007/s42985-021-00102-x
Vancouver
Nüsken N, Richter L. Solving high-dimensional Hamilton–Jacobi–Bellman PDEs using neural networks: perspectives from the theory of controlled diffusions and measures on path space. Partial Differential Equations and Applications. 2021;2(4). doi:10.1007/s42985-021-00102-x
BibTeX
@article{Nsken2021,
author = {Nüsken, Nikolas and Richter, Lorenz},
title = {Solving high-dimensional Hamilton–Jacobi–Bellman PDEs using neural networks: perspectives from the theory of controlled diffusions and measures on path space},
journal = {Partial Differential Equations and Applications},
year = {2021},
volume = {2},
number = {4},
doi = {10.1007/s42985-021-00102-x}
}
EUVIMED