Randomized quasi-Monte Carlo methods in global sensitivity analysis

Giray Ökten, Yaning Liu
April 2021
Cite DOI
Reliability Engineering and System Safety

Abstract

Randomized quasi-Monte Carlo methods have enjoyed increasing popularity in applications due to their faster convergence rate than Monte Carlo, and the existence of simple statistical tools to analyze the error of their estimates similar to Monte Carlo. In this paper we give a survey of randomized quasi-Monte Carlo methods, transformation methods for low-discrepancy sequences, and provide some examples.

Type
Journal article
Publication
Reliability Engineering and System Safety (210)
  • A survey of randomized quasi-Monte Carlo methods is given.
  • Randomized quasi-Monte Carlo methods are used in estimating Sobol’ sensitivity indices.
  • Transformation methods for low-discrepancy sequences are discussed.
  • Applications to crack propagation and Rothermel’s wildland fire spread model are considered.
Source Themes
Giray Ökten
Giray Ökten
Co-president

My research interests include computational finance and (quasi-) Monte Carlo methods

Related