Rare Event Simulation of Small Noise Diffusions

Eric Vanden Eijnden, Jonathan Weare

Research output: Contribution to journalArticle

Abstract

An important sampling method for certain rare event problems involving small noise diffusions is proposed. Standard Monte Carlo schemes for these problems behave exponentially poorly in the small noise limit. Previous work in rare event simulation has focused on developing estimators with optimal exponential variance decay rates. This criterion still allows for exponential growth of the statistical relative error. We show that an estimator related to a deterministic control problem not only has an optimal variance decay rate but can have vanishingly small relative statistical error in the small noise limit. The sampling method based on this estimator can be seen as the limit of the zero variance importance sampling scheme, which uses the solution of the second-order partial differential equation (PDE) associated with the diffusion. In the scheme proposed here this PDE is replaced by a Hamilton-Jacobi equation whose solution is computed pointwise on the fly from its variational formulation, an operation that remains practical even in high-dimensional problems. We test the scheme on several simple illustrative examples as well as a stochastic PDE, the noisy Allen-Cahn equation.

Original languageEnglish (US)
Pages (from-to)1770-1803
Number of pages34
JournalCommunications on Pure and Applied Mathematics
Volume65
Issue number12
DOIs
StatePublished - Dec 2012

Fingerprint

Rare Event Simulation
Partial differential equations
Sampling Methods
Estimator
Decay Rate
Sampling
Importance sampling
Partial differential equation
Allen-Cahn Equation
Rare Events
Stochastic Partial Differential Equations
Importance Sampling
Hamilton-Jacobi Equation
Exponential Growth
Variational Formulation
Relative Error
Second order differential equation
Control Problem
High-dimensional
Zero

ASJC Scopus subject areas

  • Mathematics(all)
  • Applied Mathematics

Cite this

Rare Event Simulation of Small Noise Diffusions. / Vanden Eijnden, Eric; Weare, Jonathan.

In: Communications on Pure and Applied Mathematics, Vol. 65, No. 12, 12.2012, p. 1770-1803.

Research output: Contribution to journalArticle

@article{ecc5beb6a0bc470a9a08c53e789277a8,
title = "Rare Event Simulation of Small Noise Diffusions",
abstract = "An important sampling method for certain rare event problems involving small noise diffusions is proposed. Standard Monte Carlo schemes for these problems behave exponentially poorly in the small noise limit. Previous work in rare event simulation has focused on developing estimators with optimal exponential variance decay rates. This criterion still allows for exponential growth of the statistical relative error. We show that an estimator related to a deterministic control problem not only has an optimal variance decay rate but can have vanishingly small relative statistical error in the small noise limit. The sampling method based on this estimator can be seen as the limit of the zero variance importance sampling scheme, which uses the solution of the second-order partial differential equation (PDE) associated with the diffusion. In the scheme proposed here this PDE is replaced by a Hamilton-Jacobi equation whose solution is computed pointwise on the fly from its variational formulation, an operation that remains practical even in high-dimensional problems. We test the scheme on several simple illustrative examples as well as a stochastic PDE, the noisy Allen-Cahn equation.",
author = "{Vanden Eijnden}, Eric and Jonathan Weare",
year = "2012",
month = "12",
doi = "10.1002/cpa.21428",
language = "English (US)",
volume = "65",
pages = "1770--1803",
journal = "Communications on Pure and Applied Mathematics",
issn = "0010-3640",
publisher = "Wiley-Liss Inc.",
number = "12",

}

TY - JOUR

T1 - Rare Event Simulation of Small Noise Diffusions

AU - Vanden Eijnden, Eric

AU - Weare, Jonathan

PY - 2012/12

Y1 - 2012/12

N2 - An important sampling method for certain rare event problems involving small noise diffusions is proposed. Standard Monte Carlo schemes for these problems behave exponentially poorly in the small noise limit. Previous work in rare event simulation has focused on developing estimators with optimal exponential variance decay rates. This criterion still allows for exponential growth of the statistical relative error. We show that an estimator related to a deterministic control problem not only has an optimal variance decay rate but can have vanishingly small relative statistical error in the small noise limit. The sampling method based on this estimator can be seen as the limit of the zero variance importance sampling scheme, which uses the solution of the second-order partial differential equation (PDE) associated with the diffusion. In the scheme proposed here this PDE is replaced by a Hamilton-Jacobi equation whose solution is computed pointwise on the fly from its variational formulation, an operation that remains practical even in high-dimensional problems. We test the scheme on several simple illustrative examples as well as a stochastic PDE, the noisy Allen-Cahn equation.

AB - An important sampling method for certain rare event problems involving small noise diffusions is proposed. Standard Monte Carlo schemes for these problems behave exponentially poorly in the small noise limit. Previous work in rare event simulation has focused on developing estimators with optimal exponential variance decay rates. This criterion still allows for exponential growth of the statistical relative error. We show that an estimator related to a deterministic control problem not only has an optimal variance decay rate but can have vanishingly small relative statistical error in the small noise limit. The sampling method based on this estimator can be seen as the limit of the zero variance importance sampling scheme, which uses the solution of the second-order partial differential equation (PDE) associated with the diffusion. In the scheme proposed here this PDE is replaced by a Hamilton-Jacobi equation whose solution is computed pointwise on the fly from its variational formulation, an operation that remains practical even in high-dimensional problems. We test the scheme on several simple illustrative examples as well as a stochastic PDE, the noisy Allen-Cahn equation.

UR - http://www.scopus.com/inward/record.url?scp=84867596369&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84867596369&partnerID=8YFLogxK

U2 - 10.1002/cpa.21428

DO - 10.1002/cpa.21428

M3 - Article

AN - SCOPUS:84867596369

VL - 65

SP - 1770

EP - 1803

JO - Communications on Pure and Applied Mathematics

JF - Communications on Pure and Applied Mathematics

SN - 0010-3640

IS - 12

ER -