Efficient solution for a class of Markov chain models of telecommunication systems

A reduction algorithm based on a modification of the classical Gaussian elimination procedure is presented for the computation of the steady-state solution of large Markov chains. The proposed method is shown to be very efficient for a class of interesting models in the domain of queueing and high-speed telecommunication systems. The complexity of the approach is assessed in terms of storage requirements and number of elementary operations, and the advantages and disadvantages of the proposed method are highlighted.