An efficient computational method for statistical moments of Burgers equation with random initial conditions.pdfVIP

AN EFFICIENT COMPUTATIONAL METHOD FOR STATISTICAL MOMENTS OF BURGER’S EQUATION WITH RANDOM INITIAL CONDITIONS HONGJOONG KIM Received 15 November 2005; Revised 4 July 2006; Accepted 12 September 2006 The paper is concerned with efficient computation of numerical solutions to Burger’s equation with random initial conditions. When the Lax-Wendroff scheme (LW) is ex- panded using the Wiener chaos expansion (WCE), random and deterministic effects can be separated and we obtain a system of deterministic equations with respect to Hermite- Fourier coefficients. One important property of the system is that all the statistical mo- ments of the solution to the Burger’s equation can be computed using the solution of the system only. Thus LW with WCE presents an alternative to computing moments by the Monte Carlo method (MC). It has been numerically demonstrated that LW with WCE approach is equally accurate but substantially faster than MC at least for certain classes of initial conditions. Copyright © 2006 Hongjoong Kim. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 1. Introduction Uncertainty is observed in many and various phenomena in engineering, physics, biology, or finance. For example, small-scale effects in multiphase flow [13, 14] or Navier-Stokes equations [18, 26] may not be completely known, but subject to some random envi- ronmental effects. Or observational errors from inaccurate measurements may not be removed. Long waves of relatively shallow depth driven by white noise can be described approximately by Korteweg-de Vries equation with stochastic perturbations [8, 30]. Op- tion pricing in financial markets may also introduce randomness [4, 10]. These uncer- tainties are represented mathematically by (functions of) random variables or stochastic processes. Then, the governi