CONVERGENCE ANALYSIS OF AN ITERATIVE AGGREGATION DISAGGREGATION METHOD FOR COMPUTING STATIONARY PROBABILITY VECTORS OF STOCHASTIC MATRICES/

An aggregation/disaggregation iterative algorithm for computing statio nary probability vectors of stochastic matrices is analysed. Two conve rgence results are presented. First, it is shown that fast, global con vergence can be achieved provided that a sufficiently high number of r elaxations is performed on the fine level. Second, local convergence i s shown to take place with just one relaxation performed on the fine l evel. The convergence proofs are general and require no assumptions on the magnitude of off-diagonal elements (blocks). Furthermore, a relat ionship between the errors on the fine and on the coarse level is desc ribed. To illustrate the theory, the results of some numerical experim ents are presented. (C) 1998 John Wiley & Sons, Ltd.