Deep neural networks algorithms for stochastic control problems on finite horizon, Part 2: numerical applications

Abstract : This paper presents several numerical applications of deep learning-based algorithms that have been analyzed in [11]. Numerical and comparative tests using TensorFlow illustrate the performance of our different algorithms, namely control learning by performance iteration (algorithms NNcontPI and ClassifPI), control learning by hybrid iteration (algorithms Hybrid-Now and Hybrid-LaterQ), on the 100-dimensional nonlinear PDEs examples from [6] and on quadratic Backward Stochastic Differential equations as in [5]. We also provide numerical results for an option hedging problem in finance, and energy storage problems arising in the valuation of gas storage and in microgrid management.
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01949221
Contributeur : Huyen Pham <>
Soumis le : mercredi 12 décembre 2018 - 13:59:50
Dernière modification le : samedi 16 mars 2019 - 01:59:01
Document(s) archivé(s) le : mercredi 13 mars 2019 - 12:44:38

Fichiers

Deepconsto-Partie2_Final.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01949221, version 1
  • ARXIV : 1812.05916

Citation

Achref Bachouch, Côme Huré, Nicolas Langrené, Huyen Pham. Deep neural networks algorithms for stochastic control problems on finite horizon, Part 2: numerical applications. 27 pages, 11 figures. 2018. 〈hal-01949221〉

Partager

Métriques

Consultations de la notice

132

Téléchargements de fichiers

130