Reaction-diffusion problems with random parameters using the generalized stochastic finite difference method

Abstract

The main idea here is to demonstrate the new stochastic discrete computational approach consisting of the generalized stochastic perturbation technique based on the Taylor expansions of the random variables and, at the same time, classical Finite Difference Method on the regular grids. As it is documented by the computational illustrations, it is possible to determine using this approach also higher probabilistic moments and to provide full hybrid analytical-discrete analysis for any random dispersion of input variables unlike in the second order second moment technique worked out before. A numerical algorithm is implemented here using the straightforward partial differentiation of the reaction-diffusion equation with respect to the random input quantity; all symbolic computations of probabilistic moments and characteristics are completed by the computer algebra system MAPLE.