Bonjour,

Vous êtes tous cordialement invités à assister à la présentation du sujet de recherche de Nicolas Gagné (partie orale de l'examen prédoctoral).

Jury

Abstract

Optimal stopping is a mathematical problem that consists of having only one decision to make: when to stop. In order to make the subject more tangible, we will use the problem of pricing an American option as our running example. We will use Monte Carlo methods because they scale well with the dimension of the state space. The contribution is the introduction of random time coordinate transforms for solving optimal stopping problems. This extra degree of freedom in the choice of the coordinate system is interesting for many reasons, including that it allows to limit the number of possible stopping points to consider. Once in the reduced problem, one can leverage many of the existing methods. We show how our approach compares to other methods, list a few intriguing observations and the research directions in which they point, as well as the numerical experiments we would like to perform to empirically benchmark them.