Stabilized multilevel Monte Carlo method for stiff stochastic differential equations

Assyr Abdulle, Adrian Blumenthal
Academic Press Inc Elsevier Science, 2013
Journal paper

A multilevel Monte Carlo (MLMC) method for mean square stable stochastic differential equations with multiple scales is proposed. For such problems, that we call stiff, the performance of MLMC methods based on classical explicit methods deteriorates because of the time step restriction to resolve the fastest scales that prevents to exploit all the levels of the MLMC approach. We show that by switching to explicit stabilized stochastic methods and balancing the stabilization procedure simultaneously with the hierarchical sampling strategy of MLMC methods, the computational cost for stiff systems is significantly reduced, while keeping the computational algorithm fully explicit and easy to implement. Numerical experiments on linear and nonlinear stochastic differential equations and on a stochastic partial differential equation illustrate the performance of the stabilized MLMC method and corroborate our theoretical findings. (C) 2013 Elsevier Inc. All rights reserved.

About this result

This page is automatically generated and may contain information that is not correct, complete, up-to-date, or relevant to your search query. The same applies to every other page on this website. Please make sure to verify the information with EPFL's official sources.

Related concepts

Related publications

Assyr Abdulle, Adrian Blumenthal
Academic Press Inc Elsevier Science, 2013
Journal paper

Abstract

Official source

https://infoscience.epfl.ch/record/183502?ln=en

About this result

Related concepts (15)

Related concepts

Related publications (26)

Related publications

Stochastic differential equation

A stochastic differential equation (SDE) is a differential equation in which one or more of the terms is a stochastic process, resulting in a solution which is also a stochastic process. SDEs have m

Monte Carlo method

Monte Carlo methods, or Monte Carlo experiments, are a broad class of computational algorithms that rely on repeated random sampling to obtain numerical results. The underlying concept is to use rando

Partial differential equation

In mathematics, a partial differential equation (PDE) is an equation which computes a function between various partial derivatives of a multivariable function. The function is often thought of as

MATHICSE Technical Report: A discontinuous Galerkin reduced basis numerical homogenization method for fluid flow in porous media

Assyr Abdulle, Ondrej Budác

We present a new conservative multiscale method for Stokes flow in heterogeneous porous media. The method couples a discontinuous Galerkin finite element method(DG-FEM) at the macroscopic scale for the solution of an effective Darcy equation with a Stokes solver at the pore scale to recover effective permeabilities at macroscopic quadrature points. To avoid costly computation of numerous Stokes problems through-out the macroscopic computational domain, the pore geometry is parametrized and a model order reduction algorithm is used to select representative microscopic geometries. Accurate Stokes solutions and related permeabilities are obtained for these representative geometries in an offline stage. In an online stage, the DG-FEM is computed with permeabilities recovered at the desired macroscopic quadrature points from the precomputed Stokes solutions. The multiscale method is shown to be mass conservative at the macro scale and the computational cost for the online stage is similar to the cost of solving a single scale Darcy problem. Numerical experiments for two and three dimensional problems illustrate the efficiency and the performance of the proposed method.

MATHICSE2015

Accurate and efficient evaluation of failure probability for partial different equations with random input data

Peng Chen, Alfio Quarteroni

Several computational challenges arise when evaluating the failure probability of a given system in the context of risk prediction or reliability analysis. When the dimension of the uncertainties becomes high, well established direct numerical methods can not be employed because of the "curse-of-dimensionality". Many surrogate models have been proposed with the aim of reducing computational effort. However, most of them fail in computing an accurate failure probability when the limit state surface defined by the failure event in the probability space lacks smoothness. In addition, for a stochastic system modeled by partial differential equations (PDEs) with random input, only a limited number of the underlying PDEs (order of a few tens) are affordable to solve in practice due to the considerable computational effort, therefore preventing the application of many numerical methods especially for high dimensional random inputs. In this work we develop hybrid and goal-oriented adaptive reduced basis methods to tackle these challenges by accurately and efficiently computing the failure probability of a stochastic PDE. The curse-of-dimensionality is significantly alleviated by reduced basis approximation whose bases are constructed by goal-oriented adaptation. Moreover, an accurate evaluation of the failure probability for PDE system with solution of low regularity in probability space is guaranteed by the certified a posteriori error bound for the output approximation error. At the end of this paper we suitably extend our proposed method to deal with more general PDE models. Finally we perform several numerical experiments to illustrate its computational accuracy and efficiency. (C) 2013 Elsevier B.V. All rights reserved.

Elsevier Science Sa2013

, ,

We consider the numerical approximation of a risk-averse optimal control problem for an elliptic partial differential equation (PDE) with random coefficients. Specifically, the control function is a deterministic, dis- tributed forcing term that minimizes the expected mean squared distance between the state (i.e. solution to the PDE) and a target function, subject to a regularization for well posedness. For the numerical treatment of this risk-averse optimal control problem, we consider a Finite Element discretization of the underlying PDEs, a Monte Carlo sampling method, and gradient type iterations to obtain the approximate optimal control. We provide full error and complexity analysis of the proposed numerical schemes. In particular we compare the complexity of a fixed Monte Carlo gradient method, in which the Finite Element discretization and Monte Carlo sample are chosen initially and kept fixed over the gradient iterations, with a Stochastic Gradient method in which the expectation in the computation of the steepest descent direction is approximated by independent Monte Carlo estimators with small sample sizes and possibly varying Finite Element mesh sizes across iterations. We show in particular that the second strategy results in an improved computational complexity. The theoretical error estimates and complexity results are confirmed by our numerical experiments.

MATHICSE2018