Adomian decomposition: a tool for solving a system of fractional differential equations

Adormian decomposition method has been employed to obtain solutions of a system of fractional differential equations. Convergence of the method has been discussed with some illustrative examples. In particular, for the initial value problem: [D(alpha1)y(1),..., D(alphan)y(n)](t) = A (y1, ... , y(n))(t), yi (0) = c(i), i = 1,..., n, where A = [a(ij)] is a real square matrix, the solution turns out (y) over bar (x) E-(alpha1,E-...,E-alphan),1(x(alpha1) A(1),..., x(alphan) A(n))(y) over bar (0), whereE((alphaa,...,alphan),1) denotes multivariate Mittag-Leffler function defined for matrix arguments and A(i) is the matrix having ith row as [a(i1) ... a(in)], and all other entries are zero. Fractional oscillation and Bagley-Torvik equations are solved as illustrative examples. (C) 2004 Elsevier Inc. All rights reserved.