Abstract

We develop methods to solve general optimal stopping problems with opportunities to stop that arrive randomly. Such problems occur naturally in several econometric applications. Pivotal to our approach is that our methods operate on random rather than deterministic time scales. More specifically, we convert the original problem, which is essentially defined on a finite time horizon, into an equivalent discrete-time optimal stopping problem with N0-valued stopping times and an infinite horizon. To numerically solve this problem, we revisit the dual martingale approach, consider a random times least squares Monte Carlo method and, in particular,  analyze an iterative stochastic policy improvement procedure in an infinite horizon setting. The efficiency of our methods is demonstrated at some numerical case studies. (Joint with Josha Dekker, Roger Laeven, Michel Vellekoop.)

Speaker

John Schoenmakers (Weierstrass Institute)