Sensitivity study of heston stochastic volatility model using GPGPU

Authors:
Emanouil I. Atanassov;Sofiya Ivanovska
Affiliations:
Institute of Information and Communication Technlogies, Bulgarian Academy of Sciences, Sofia, Bulgaria;Institute of Information and Communication Technlogies, Bulgarian Academy of Sciences, Sofia, Bulgaria
Venue:
LSSC'11 Proceedings of the 8th international conference on Large-Scale Scientific Computing
Year:
2011

Citing 6
Cited 0

Random number generation and quasi-Monte Carlo methods

Random number generation and quasi-Monte Carlo methods
Scrambling sobol' and niederreiter-xing points

Journal of Complexity
Global sensitivity indices for nonlinear mathematical models and their Monte Carlo estimates

Mathematics and Computers in Simulation - IMACS sponsored Special issue on the second IMACS seminar on Monte Carlo methods
A New Efficient Algorithm for Generating the Scrambled Sobol' Sequence

NMA '02 Revised Papers from the 5th International Conference on Numerical Methods and Applications
The Shape and Term Structure of the Index Option Smirk: Why Multifactor Stochastic Volatility Models Work So Well

Management Science
Tuning the generation of sobol sequence with owen scrambling

LSSC'09 Proceedings of the 7th international conference on Large-Scale Scientific Computing

Quantified Score

Hi-index	0.00

Visualization

Abstract

The focus of this paper is on effective parallel implementation of Heston Stochastic Volatility Model using GPGPU. This model is one of the most widely used stochastic volatility (SV) models. The method of Andersen provides efficient simulation of the stock price and variance under the Heston model. In our implementation of this method we tested the usage of both pseudo-random and quasi-random sequences in order to evaluate the performance and accuracy of the method. We used it for computing Sobol' sensitivity indices of the model with respect to input parameters. Since this method is computationally intensive, we implemented a parallel GPGPU-based version of the algorithm, which decreases substantially the computational time. In this paper we describe in detail our implementation and discuss numerical and timing results.