A closed form approximation and error quantification for the response transition probability density function of a class of stochastic differential equations

摘要

A closed-form analytical approximation is derived for the response transition probability density function (PDF) of a certain class of stochastic differential equations with constant drift and nonlinear diffusion coefficients. This is done by resorting to a recently developed Wiener path integral based technique (WPI) in conjunction with a Cauchy-Schwarz inequality treatment of the problem. The derived approximation can be used, due to its analytical nature, as a direct SDE response PDF estimate that requires zero computational effort for its determination. Further, it facilitates an error quantification analysis, which yields an a priori estimate of the anticipated accuracy obtained by applying the approximate methodology. The reliability of the approximation is demonstrated via several engineering mechanics/dynamics related numerical examples pertaining to the stochastic beam bending problem, as well as to the response determination of stochastically excited nonlinear oscillators. (C) 2017 Elsevier Ltd. All rights reserved.