Mentor

Tyler Brough

Document Type

Presentation

Publication Date

5-1-2014

First Page

1

Last Page

25

Abstract

We show various methods that increase the precision and convergence speed of simulated stochastic processes. We demonstrate our precision and speed by using an example from the finance world, namely that of an Asian Option. An Asian Option is a path dependent pricing mechanism that is normally priced using Monte Carlo methods, it can be thought of as a path dependent diffusion equation. We show that the precision of the simulations is increased by 70% using Control Variates (derived by approximating the true mean from an analytic closed form solution). Using sequential Monte Carlo and parallel computing across a GPU (CUDA) increases convergence speed.

Download

Additional Files

ControlVariateAsianOption2.py (2 kB)
Python Code Simple MC

bridge.r (1 kB)
R Code Brownian Bridge

Included in

Physics Commons

Share

COinS