Deep Learning for Continuous-Time Stochastic Control with Jumps

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Cheridito, Patrick, Dupret, Jean-Loup, Hainaut, Donatien
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866914256455204864
author Cheridito, Patrick
Dupret, Jean-Loup
Hainaut, Donatien
author_facet Cheridito, Patrick
Dupret, Jean-Loup
Hainaut, Donatien
contents In this paper, we introduce a model-based deep-learning approach to solve finite-horizon continuous-time stochastic control problems with jumps. We iteratively train two neural networks: one to represent the optimal policy and the other to approximate the value function. Leveraging a continuous-time version of the dynamic programming principle, we derive two different training objectives based on the Hamilton-Jacobi-Bellman equation, ensuring that the networks capture the underlying stochastic dynamics. Empirical evaluations on different problems illustrate the accuracy and scalability of our approach, demonstrating its effectiveness in solving complex high-dimensional stochastic control tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2505_15602
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Deep Learning for Continuous-Time Stochastic Control with Jumps
Cheridito, Patrick
Dupret, Jean-Loup
Hainaut, Donatien
Machine Learning
Systems and Control
Optimization and Control
Portfolio Management
93E20, 68T07, 65C30
I.2.8; I.2.6
In this paper, we introduce a model-based deep-learning approach to solve finite-horizon continuous-time stochastic control problems with jumps. We iteratively train two neural networks: one to represent the optimal policy and the other to approximate the value function. Leveraging a continuous-time version of the dynamic programming principle, we derive two different training objectives based on the Hamilton-Jacobi-Bellman equation, ensuring that the networks capture the underlying stochastic dynamics. Empirical evaluations on different problems illustrate the accuracy and scalability of our approach, demonstrating its effectiveness in solving complex high-dimensional stochastic control tasks.
title Deep Learning for Continuous-Time Stochastic Control with Jumps
topic Machine Learning
Systems and Control
Optimization and Control
Portfolio Management
93E20, 68T07, 65C30
I.2.8; I.2.6
url https://arxiv.org/abs/2505.15602