A Meshless Method Based on the Dual Reciprocity Method for One-Dimensional Stochastic Partial Differential Equations.

This article describes a new meshless method based on the dual reciprocity method (DRM) for the numerical solution of one-dimensional stochastic heat and advection-diffusion equations. First, the time derivative is approximated by the time-stepping method to transforming the original stochastic partial differential equations (SPDEs) into elliptic SPDEs. The resulting elliptic SPDEs have been approximated with the new method, which is a combination of radial basis functions (RBFs) method and the DRM method. We have used inverse multiquadrics (IMQ) and generalized IMQ (GIMQ) RBFs, to approximate functions in the presented method. The noise term has been approximated at the source points, at each time step. The developed formulation is verified in two test problems with investigating the convergence and accuracy of numerical results. [ABSTRACT FROM AUTHOR]