Estimating expected first passage times using multilevel Monte Carlo algorithm

Thesis

In this paper we devise a method of numerically estimating the expected first passage times of stochastic processes. We use Monte Carlo path simulations with Milstein discretisation scheme to approximate the solutions of scalar stochastic differential equations. To further reduce the variance of the estimated expected stopping time and improve computational efficiency, we use the multi-level Monte Carlo algorithm, recently developed by Giles (2008a), and other variance-reduction techniques. Our numerical results show significant improvements over conventional Monte Carlo techniques.

Authors: Primozic, T

• Publisher: oxford university;mathematical institute
• Publication date: 2011-06-24
• Type of award: DPhil
• Level of award: Doctoral