We consider iterative methods for the minimal nonnegative solution of the m
atrix equation G = Sigma(i=0)(infinity)A(i)G(i), where the matrices A(i) ar
e nonnegative and Sigma(i=0)(infinity)A(i) is stochastic. Convergence theor
y for an inversion free algorithm is established. The convergence rate of t
his algorithm is shown to be comparable with that of the fastest iteration
among three fixed point iterations. (C) 1999 Elsevier Science Inc. All righ
ts reserved.