Control variates for variance-reduced ratio of means estimators

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Bocquet-Nouaille, Louison, Morio, Jérôme, Bobbia, Benjamin
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909891349708800
author Bocquet-Nouaille, Louison
Morio, Jérôme
Bobbia, Benjamin
author_facet Bocquet-Nouaille, Louison
Morio, Jérôme
Bobbia, Benjamin
contents The control variates method is a classical variance reduction technique for Monte Carlo estimators that exploits correlated auxiliary variables without introducing bias. In many applications, the quantity of interest can be expressed as a ratio of expectations. We propose a variance-reduced estimator for such ratios, which applies control variates to both the numerator and the denominator. The control variate coefficients are optimized jointly to minimize the variance of the resulting estimator. This approach theoretically guarantees variance reduction and naturally extends to approximate control variates. Simulation studies show significant variance reduction, particularly when correlations between variables and control variates are strong. The practical value of the method is illustrated on multi-fidelity applications: estimating a proportion in an aircraft design use case and a conditional value-at-risk in an electromagnetic dataset.
format Preprint
id arxiv_https___arxiv_org_abs_2510_13504
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Control variates for variance-reduced ratio of means estimators
Bocquet-Nouaille, Louison
Morio, Jérôme
Bobbia, Benjamin
Statistics Theory
The control variates method is a classical variance reduction technique for Monte Carlo estimators that exploits correlated auxiliary variables without introducing bias. In many applications, the quantity of interest can be expressed as a ratio of expectations. We propose a variance-reduced estimator for such ratios, which applies control variates to both the numerator and the denominator. The control variate coefficients are optimized jointly to minimize the variance of the resulting estimator. This approach theoretically guarantees variance reduction and naturally extends to approximate control variates. Simulation studies show significant variance reduction, particularly when correlations between variables and control variates are strong. The practical value of the method is illustrated on multi-fidelity applications: estimating a proportion in an aircraft design use case and a conditional value-at-risk in an electromagnetic dataset.
title Control variates for variance-reduced ratio of means estimators
topic Statistics Theory
url https://arxiv.org/abs/2510.13504