LisMath Seminar   RSS

22/04/2020, 17:00 — 18:00 — Online
Afonso Moreira, Instituto Superior Técnico, Universidade de Lisboa

Machine Learning Driven Optimal Stopping

Optimal stopping problems constitute a subset of stochastic control problems in which one is interest in finding the best time to take a given action. This framework has relevant contributions extending across different fields, namely finance, game theory and statistics. Recently the literature on machine learning has grown at a very large pace, specially in what concerns the usage of its techniques in other fields beyond computer science, in the hope that those might shed some light in long persisting problems such as, for instance, the well known curse of dimensionality. In light with this trend the literature on both stochastic control and optimal stopping has presented several contributions by either incorporating reinforcement learning techniques (Machine Learning Control) or by making use of neural networks to estimate the optimal stopping time of a given problem.


[1] G. Peskir and A. Shiryaev, Optimal stopping and free-boundary problems, 2006, Springer.

[2] W. H. Fleming and H. M. Soner, Controlled Markov processes and viscosity solutions vol. 25, 2006, Springer Science & Business Media.

[3] S. Becker, P. Cheridito and A. Jentzen, Deep optimal stopping, Journal of Machine Learning Research, vol. 20, 2019.

[4] S. Becker, P. Cheridito, A. Jentzen and T. Welti, Solving high-dimensional optimal stopping problems using deep learning, arXiv preprint arXiv:1908.01602.

[5] H.J. Kappen, An introduction to stochastic control theory, path integrals and reinforcement learning, AIP conference proceedings, vol. 887, nr. 1, pp. 149–181, 2007.

[6] E. Theodorou, J. Buchli and S. Schaal, A generalized path integral control approach to reinforcement learning, Journal of Machine Learning Research, vol. 11, Nov, pp. 3137–3181, 2010.

See also



Universidade de Lisboa FCUL